September 6, 2023 • 45 mins
It's a remarkable fact that, excluding Google, Algolia Search surpasses the combined search requests of giants like Bing, Yahoo, and others.
In this enlightening episode of the Digital Discovery Podcast, host Nigel Price is joined by Search Expert Bruce Macdonald to dive deep into the fascinating world of Onsite Search. This is not merely a chat about technology, but a revelation about how search and digital discovery directly affects your business's bottom line and the overall customer experience.
Notice: Before making any decisions regarding your ecommerce strategy, we strongly recommend you conduct your own in-depth research and seek expert advice.
Related Links:
Listen
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Welcome to the show, Bruce. How are you going?
(00:14):
Mate Nigel, thank you so much for having me. Long time, this is our first time caller.
That's what they say, isn't it?
You're welcome, mate. I'm really excited to have you on the show today. And when I reached
out to you a few weeks ago, and I was thinking, you know, we were talking about personalization
a few weeks ago, and I was thinking to myself, search is next. And the search piece is such
(00:36):
an important piece of, you know, e-commerce and retail at the moment. So I was thinking,
I have to reach out to Bruce and we'll get him on the show. So thank you so much for
joining the show today.
No, no, I appreciate it. And again, you've hit the nail on the head with search becoming
more relevant. We've seen obviously an influx of inquiries, probably in the back of chat
and CPT and open R. That's been the people have now talked about the conversation versus,
(01:01):
you know, trying to serve them up what they have. And that's what's really exciting about
our markets. It's in this kind of flux state. And I think our goal is, you know, as a business,
I can explain what our goal is in a minute, but I think we're really well positioned to
help not only just retail, but help other call it verticals, digitize their practice
and potentially monetize it as well. So for us, we're in a really good spot at the moment
(01:24):
on the back of all this AI, Vector, ML capabilities that's actually been more accessible to everyday
consumers, which is basically driving us to drive our customers as well.
Yeah, yeah, no, absolutely. Yeah, it's interesting, isn't it? You're talking about digitizing
the businesses and digital transformation, right? I mean, it's really on the minds of
(01:47):
a lot of people have been talking to a few CEOs and things at the moment, CFOs and heads
of retail and so on and, and, and e-commerce. And yeah, I can see that there's, you know,
since 2020, 2021, there's more of a focus on their online experience and their, and
their websites and things. And as we know, personalization piece is a big piece, but
(02:08):
even bigger than that, I think is obviously search and discovery starts with either browsing
via the main navigation on a website by categorization, or it starts with search. And I'm finding
that, and I don't know what you're seeing, maybe you can shine some, some light on that,
but how search has really changed in the last say 10 years, when Magento first released,
(02:29):
obviously they had solo search, you would go to a website and all of the listeners would
know you go to site and you'd search for something obscure, let's just say, or not so obscure.
It didn't really matter. The results were always different. You know, you'd search for
the same thing even, you know, two times you would find different results. So we know search
is getting better and better. You know, one of the reasons I actually met or discovered
(02:53):
Algolia was in the startup area of a Magento conference. And at the time I met Gaitan and
we can go into that story in a minute, but I was looking for a solution for search and,
and nobody really had anything. I was looking for something that I could give to my developers
to say, Hey guys, why don't you check this out? Meanwhile, they're on the hamster wheel
trying to fix, you know, the search.
(03:16):
The airplane in the air with the engine. What was the catalyst for you to do that though?
Was it just simply that you weren't seeing, you knew that search was a problem because
customer feedback or, you know, was it that you recognize that if we're going to make
some quick wins, as you said, it's like, can I be able to browse in the search and they're
the kind of two vehicles of how they can engage with your platform. What was the kind of driver
(03:37):
there? Was it just simply that you're looking for?
Well, I think there's probably a handful of drives. I think the main one was the customer
experience. Obviously it's frustrating and that's the first thing you try to fix. You
keep your customers happy so that you always put them first. Then I think it was, you know,
just generally the business, you know, when you're talking to suppliers and they're searching
for their product or SKUs or part of a model number and it's not appearing in our search,
(04:00):
they think it's not on the website and then it creates this internal kind of turmoil and
it just chews up time and it's really bad for productivity. So you've got the customer
side, the retail side where they can't find your product reliably via search. And then
you've got the internal battle that's going on with trying to explain to people, unless
you say I won't name names, buyers, let's just say that no, the product is there, but
(04:26):
search is just rubbish, you know, how it was back in the day and, you know, and that we're
working on it. But obviously it's very hard to have that internal conversation about how
complicated search is. We try to explain latent semantic indexing and, you know, search algorithms
and things to someone who's just interested in buying and selling stuff and making money
in between, which is basically retail. It's quite a complex conversation to have and it's
(04:50):
very hard to sit there and justify why it keeps happening, you know, and it's just the
technology hadn't caught up to, you know, the fact that consumers were really now starting
to embrace online. So really it was, that was the key drivers, you know, consumer, so
customer experience and, you know, internal pressures from not being able to find product.
It's actually, you've kind of hit the nail on the head and I'll go back and talk you
(05:13):
through kind of the evolution of where search has come from, but those three users that
you talk about, you have the developer, they're kind of like, how do we get these guys to
build something that's going to service both our customers and our business users? That's
the, you know, the stakeholders and, you know, how do we merchandise the site? And I think
(05:34):
those three users, the customer, the business user, and really the merchandising team who
are responsible for making sure product comes up and is relevant and it's moving, right?
But I think search as it is today and where it's come from, you know, search has been
around for like the, like the 90s, I think that 70s and 80s in a decade kind of built
the first initial call it multifaceted search. But when you look at solar and elastic search,
(05:59):
the underlying premise of that is a thing called the CERN, it's Apache CERN technology.
It's basically a mathematical equation that looks for words or statements or phrases and
then puts a weight into it. And that's how you build the algorithm. The way that basically
the two founders of Algolia came up with our technology and that's where the name Algolia
(06:20):
comes from, it's algorithm. It's based on this tie breaking approach. So it's, so if
you look at search across all the technologies today and we'll get into vectors in a minute,
but it's under the premise of keywords. And obviously each SKU has a stock unit or each
product has attributes assigned to its size, whatever it is. And obviously we take those
(06:40):
attributes and make them relevant to a search inquiry. So small t-shirt, large t-shirt located
next to like close to me. But what we've noticed though, as the evolution has come through
is that A, that the quality of data is really important as in what data is being serviced
for these different products and SKUs. And then more importantly, how to understand the
customer's needs, what we're actually asking for. So keywords are great for the SKUs because
(07:03):
they're all embedded there, but trying to understand where the customers are coming
from. We used to build synonyms. So t-shirt, is it spelled T-E-E, is it spelled T--shirt,
is it spelled T-space shirt or all together? So it's like five different synonyms. Your
poor business users or the merchandisers sitting there, writing these things out and out and
(07:24):
out. And so I think where we've evolved from that era where it's been very much about weighting
words and terminology is actually helping with the automation side. And that's kind
of where that vector technology comes, which we'll go into in a minute, but I think that
would be really important to discuss because it really is an evolution that's been around
for a good 10 years now. That's only become more accessible. So now you don't have to
(07:47):
hire basically a data scientist. You don't have to host a data lake. Then you have to
host it like a DXP platform to stitch all together and then serve that. It's all kind
of been delivered as a service now. And that's really, I think where, I suppose where search
has exploded over the last probably two or three years now. COVID has demonstrated to
not only just the e-comm world, but to the business world. So the C levels of these large
(08:10):
institutions across all verticals, it doesn't matter if it's media, traditional FMC, geo
retail. Even the sports industry is now trying to monetize digital engagements. And I think
e-commerce is across the board. And I think that's where you'll find the evolutions coming
through for search. Go to the customers and they'll tell you what they want because they've
(08:31):
got Netflix, they've got Spotify, they've got Google, they've got all that rich, very
personalized content being served to them as they need it. Now we have to case you catch
up to that. And that's kind of where this platform of our goal really plays into quite
nicely.
On the point of knowing or understanding what your customers are asking for. And I think
that's a, that's a really good point because people think a lot of times businesses will
(08:56):
say to me, well, when I search for this, I'm expecting this. I'm saying, yeah, but hang
on a second. You're expecting it because you're, you're predicting what the future outcome
is going to be for that search term. But a customer who's in a discovery phase, they
might be searching around the topic. They're not actually, they haven't got to the point
of where they're searching for that particular term you're, you're trying to, you're searching
(09:17):
for now. So there is, and this is starts a conversation with how search actually works
and the intense signals and things that people actually show before they get to your site
and collecting that information, using it to understand. And I think for me, a good
example would be we looked at JB Hi-Fi. We did a whole bunch of, you know, our research
(09:39):
and I was leading that research. And one of the things we noticed was like, if you search
for radio head, right? So JB have media content, obviously they have albums and yep, media
and radio head would show up. But if you just quickly searching for radio, expecting to
see radio head, cause all you look at JB Hi-Fi is for example, their albums and so forth.
(09:59):
You're really not interested in digital radios, right? There's a lot of different use cases
we'd worked out that those intense signals aren't just the primer for the search path.
It's also what have they previously looked at in the past, wherever they come from on
Google search, for example, you know, all of these different data points to understand,
(10:20):
okay, when they do come, if they do search for something, it's more than likely going
to be weighted across the media section of the business, for example, albums and records
and DVDs or whatever. So I recommend you should go back to their website in about a month's
time staples.ca the staples can, uh, Canadian store. They've got this, it's kind of our
(10:41):
example side, but we've, we've just created a thing called query categorization. Um, and
so it's, it's in session experience. So the concept is, you know, if you're engaged with
JB and you've, uh, you're looking for albums or we see a lot of use case for coffee. Are
you looking for ground coffee or beans? It was the coffee machines, coffee pods, coffee
(11:03):
tables. And so the idea that you have all that product, how do you categorize it so
that the next engagement that you have with the platform is it's predicting that intent.
And that's kind of really what's the evolution is. So all those buying signals that you've
already captured, be it their net new customer, they're accustomed with a, like a very weak
(11:24):
profile or they're a well-known customer to you. How do you service those three customers?
And you know, they're all about different personas and buying habits. And that's what
we're trying to do now is, uh, I'm trying to be the, the end state, which you're your
full customer profile, because that, that runs the risk of us holding onto data that
we shouldn't do. But I think we can definitely supplement that with either third party data
(11:45):
or the stuff, the buying signals that we capture and which you would know about.
Yeah. All the tags and things. Yep. That, that idea of understanding what your, what
your intent is, um, be it cause you're a net new customer, there's no understanding
about who you are. That's a really important thing to, for a search provider to do because
it's surfacing relevant results. And that's the most important thing. That's what we designed
(12:06):
to do, right? Solve your problem, solve your need, help you be understood to this retailer
and then deliver, deliver a product service capability back then really quickly.
Yeah, that's right. I mean, if you go, just on that example, if you go search for coffee
and you've been looking at coffee beans, but you also sell machines and grinders and everything
else, like you, you, you, and you're constantly buying coffee off that particular retailer.
(12:29):
For example, the last thing you want to do is refactor those results again and put to
the top coffee machines when you know that all they, all they buy off you is coffee.
You'd want to put them further on down the list or highlight coffee machines and say,
by the way, did you know we've got an offer here? We can give you 20% off a coffee machine.
Is your coffee machine good enough? Maybe start them down that path. If you want to
(12:50):
do as a marketer, you could quite easily do that. But I think that's a really, really
good example. Yeah. Back to your point about Gaetan and I'll touch on that story. It's
quite interesting. So please tell us location. Tell us, okay, so Magento conference 2016.
I think it was 20. Magento. Were they owned by Adobe by then? No, no, no, no. This was
(13:13):
um, yeah. Yeah. Yeah. Yeah. Yeah. Yeah. So it was a great conference. You know, he had
this little tiny booth and Gaetan was there and I'd messaged him and I said, Hey, I don't
think it was at his booth. I think he was with a client. So, so we sat down and we had
a coffee and he was telling me about search. And when I was explaining to him all these
issues and like he'd heard it all before. So he knew exactly what to expect. He goes,
(13:34):
well, look, I can solve your problems, but I'm going to have to show you on my little
laptop and he had like a little laptop and he just had his phone and it was tethered
to like a, it was 3G I think connection. And it was, and I was having issues too. It was
three and 4G, right? So this is how long ago it was. There was no 5G. I'm thinking how
the hell is he going to tether his, his notebook to his phone and then run a search query and
(13:56):
get back anything in enough time to, to impress a customer. So I literally saw him, he dialed
in pulled up a query screen and you could see that it was live. There was no caching
or anything like that. And he goes to something he was searching over something. It's a huge
amount and he searched and I had to go as well, but he searched for a couple of search
terms and just, it was instant. And I looked at it. I looked at him and I was just, I was
(14:20):
just gobsmacked. I'm like, where's the spinny wheel thing? You know, the spinning wheel,
like Ajax rubber. I'm like, where's that? Where's that gone? And I'm thinking, what
this is blowing me away. And he's like, Oh no, no, you know, he started and look, I,
on purpose, I brought my propeller, my propeller hat. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah.
(14:41):
I should put my propeller hat on right now. So I can start spinning, but yeah, it brought
back all these results. And I searched a few things. I think it was a, a taxation library
of, you know, all this law or something. So it was all text-based. And it was just instant
and then all the refactoring there. And it had the most popular pages and it had, and
(15:01):
I'm thinking, and I looked at him and at the time I said, how much is it? And he told me
and I went, sign me up. That's it. I'm done. You're in. I mean, that was it. What was the
result though? Like, did you see a, like it was an increase in revenue? Was it a, was
it a conversions? Everything. I mean, it's conversions, revenue conversions, but average
order order amount. But what I was basing the success metric for me obviously was conversions,
(15:25):
but it was actually the number of products displayed per search term inside a category
that I knew was our core business. So what did I do? Because we didn't have any analysts
or anything where I was working at the time. So what I did was I took all of the keyword
search terms out of our Google account, out of our AdWords account. I then overlaid that
(15:46):
with the number of search or the number of products returning results on the site based
on that, on the ad spend the ad keywords that we had. I expected all brand terms, all the
store locator terms. It was just more product terms. So you actually waiting to terms as
well because of the cost of them. Yeah. So I aligned our margins per category. So the
(16:09):
category, so I went from the, the, I worked back from the margins and the popularity of
the products. So stock turns, then margin, margin, so availability, the margin, availability.
Yeah. Correct. So availability was, was third. All right. So because the thing he did, this
was a debate because you don't always want to show things that are in stock over things
(16:32):
that are out of stock if they're not what the intent of the search is. So for example,
if for example, Sony TV number one has the most stock turns and has the most margin,
but he's out of stock and I put that third or fourth, it's not likely to result in a
conversion, but if I put it first, they'll might sign up for when it comes back into
stock or go in store and place an order in store. Okay. And white. And that's a lot of,
(16:55):
a lot of retailers didn't realize that because a lot of people were just looking online and
going in store. So I would wait it based on that. And I could see what we were selling.
So I had all the reports so I could see the impact. It was very hard for me to show the
business what that impact was because Google would just have a row as multiplier. Yeah.
And then I worked back from that. And then I said, okay, well I've only got five products
(17:16):
showing there's 50 products in that particular category, but I'm search on that search term.
Our click through rate on our ads is really high. The quality score is really high on
our ads. And I can see, you know, we've got a dedicated landing page, but the search results
when they get there, they hit the landing page, but then they go to search. For example,
I start looking at the range and so I did it that way. So I worked back in, in that,
(17:37):
in that way to work out how to work out the ROI. Yeah. Yeah. I mean, that's, that's exactly
what I mean, that is literally the business model today. Tell us your business metrics.
Are you trying to move your stuff? You have built your own, basically you're your own
BA and your own almost yeah, like almost auto management systems to a certain degree, right?
(17:58):
Cause you're looking at inventory in real time and what the data, but what the data
piece is exactly the, that is exactly what we're about. We're the second largest search
vendor behind Google by a fair way. But if you add up all the other search vendors like
your Yahoo's and your bins times and by four, and that's kind of the size of our goal. Yeah.
Wow. That's huge. Yeah. We're seeing this wealth of information. What you just said
(18:21):
there about all these conversion metrics. If we capture that from all our customers,
we can predict what's happening globally. Now for us as a very, very PI and GDPR company,
we will never sell that data. That's part of our DPA. Google does though. And that's
how Google does their evolution. They take their analytic traffic and then they use that
(18:41):
to influence understanding. And that's what drives it. So it's a completely different
approach. Google's for the masses. We're kind of for you as the consumer. And that's
why we've kind of made our system quite closed. We don't, you can influence that by third
party buying signals and third party data, like a Google analytics or a GTM tagging and
things like that. And be extensible. But the reality is you have to understand your business
(19:05):
the best so that we can help make the relevancy change. If you need to move more product,
more margin, if you've got customer profiles, we'll help understand that for you and give
you lots of ways to do it. But you were bleeding edge man way back then doing that, looking
at the buying signals and then augmenting or merchandising the e-commerce experience
(19:25):
and the search experience and the results that are being rendered, which is amazing.
Yeah. Thanks man. Yeah. I mean, it's just retail in the day. But now, I mean, if you
look back over the past 10 years, Algolia has really been just on the point of your,
you know, the number of search queries. It's been basically part of the evolution of search
(19:45):
and it's changed the industry, the e-commerce industry globally. I think that's as profound
the impact that Algolia has actually had. And you see a lot of spin-offs now, different
products going, but they don't have the underlying purpose because it's purpose that, I don't
want to sound like a line from the matrix, but it's purpose that drives everybody, right?
You've got to have a purpose. Algolia was for the developers. I remember sitting at,
(20:10):
and I think it was a year, a couple of years after when I met Gaetan and it was already
customer and it had proven that it was working, you know, and there's always these little
teething things that we had, but that was sorted straight away pretty much. And that's
important, you know, as a product and as a retailer, you can slam your hand down on the
desk and say, oh, you know, this is not acceptable, but you've got to work with people as well.
(20:30):
You've got to work and you've got to be prepared to put the right, sell it internally and have
the right mindset of how search is. It's never perfect. It aims to be perfect. But as you
say, with the advent of AI and everything else, eventually it will be fully AI. It will
understand what you're looking for, what your preferences are in terms of brand. Like it
will, it will just know in order to just happen. And like, what do they say? What's that saying?
(20:54):
And technology can be indistinguishable from magic. And I think eventually you just go
to a site and expect it. And it'll just be, as you say, you've got all of this data. It'll
just work. It'll just have the similar search experience across most websites. I think that's
probably one of the most frustrating things. You go to a site and the search is not as
(21:15):
good as another site that you've been to. And it kind of, it sort of falls short in
your expectations that you want to shop there because it just becomes really frustrating.
Yeah. Thanks man. That's super interesting.
No, no. Even just unpacking that a little bit more. Like if you look at our customer
base and it's heavily weighed towards retail. So JB Hi-Fi, Bingley, Culture Kings, we've
(21:39):
got Woolworths as a customer. And then when I look at our in-store base, it's like 20,000
customers. I think we say 17,000, but some of these customers have multiple customers
and they've won. And so all that 60% of those retailers, they're helping us innovate. Our
customers send feature requests like it's no one's business because they do, they expect
(22:01):
that right? Their customers are asking them and they're asking us. And so that only makes
us fight more for that capability with a development base of like 350 odd team. We've got so many
developers and different capabilities trying to work on different strategies to help improve
that. One of those ones we'll talk about, which is SearchIO. We acquired them in Australian
(22:23):
business and I would highly recommend you get Hamish on. Hamish was one of the co-founders
of SearchIO. He's kind of like the godfather for us when it comes to vector search and
this idea of natural language, large language models. He is amazing in terms of what his
perception is. But when I say perception about meeting the customer where they are, I just
(22:44):
did a session in commerce tools, like an e-com for Shopify or Magento. And they were talking
about the advent of the mobile device being the full platform offering you through WeChat
in China. You buy your groceries, you pay homeless people on the street if you want
to tip them. It's all QR codes. People just want to be understood. And for us, you said
(23:10):
it before. When you think about their experiences, be it in-store experience where they're trying
to be understood searching for something, I need to find the toilet. I need to find
Nike's that fit a kid that's aged five, putting size in there. I'm looking for a lilac sweater
or a lavender sweater. Now, if you don't have purple or mauve or whatever the color
(23:31):
of shade it is, from a keyword, you're screwed. They're not going to come back with any relevancy.
It's going to give you a jumper, hoodie, jacket, pullover, whatever it is. And so I think that
understanding to whatever platform it is, be it a point of sale device, be it in-store
experience, be it digital apps, be it for business users as well, trying to do searches
(23:51):
on give us the latest inventory levels and what products are moving quickest like what
you did before. That's where we need to help the customer be better serviced.
The browse experience and content that's being served, I think that also is important because
those digital natives that are coming from Instagram, Facebook, and only these social
platforms, they want to feel that same brand and trust come through to whatever platform
(24:15):
or brand they're touching. And so I think getting that consistent message that you were
saying before, going to one site, getting let down and then going to another site and
being, wow, for the audience, look, search converts. There's no doubt about it. For every
person that engages with search and look at Google Analytics, jump on there and say, people
that are having search sessions, what's the conversion rate now compared to browse sessions?
(24:38):
It's that simple. If you're doing it right, it should be in the realm of like five to
six X. That's five to six times conversions through search than it is to browse. Because
customers are coming to the website either as a known quantity or an unknown quantity
and you're just trying to work out which bucket you're going to fit in and then help them
convert for you. One thing I would say though is when you think about those pillars, you
(25:00):
need to be AI movement and where you buy one platform or one capability and it solves all
your problems. I think you got to be aware that there's you as a consumer and you as
a vendor, as a brand, you know your customers better than anyone. You can't expect some
third party coming in and say, just turn on the tweaks of AI and it'll run itself. You
(25:23):
still need to have some level of input to help drive your signals, your buying signals.
Because we know everyone wants that Nike because on sale and there's none of it left. So they
want to keep coming back and asking for it. And to the idea of, I know you're looking
for the size 10 and I know you're coming from this location. So how do I help you find it
in store or how do I help? That may not serve us their needs, but at least it gives them
(25:46):
opportunities to help discover what else we have to offer for you and see what else you're
interested in. And then once you've got them in that buying mode and try and help them
discover what you have, they're like, oh wow, you've got heaps of stuff, right? This is
feels like an endless aisle experience kind of approach. When you think about Algolia's
technology, that's actually fundamentally our biggest offering in terms of capability.
When it comes to speed and the speed of those results, not only to index data really quickly,
(26:10):
but more importantly to surface that data back to mobile device. When we talk about
speed, we're talking about like sub 20 milliseconds. If you're the other spectrum, which is a hundred
milliseconds or even like the inability to load a page and a part of that page might be
simply recommendations and you're taking like 300 milliseconds to serve that up. That's
money walking out that door for every point one of a second. It's like, like it reduces
(26:33):
your percentage of conversion by 8% black Friday coming up. How do we get the season,
the right seasonal data into play? And I think that's really fundamentally important as well.
So I'm hoping on that a bit.
No, no, I mean, it's, it's absolutely spot on. One of the things I've told many people
who've asked me about search is that, you know, it's a bit like buying a Ferrari and
(26:54):
not taking it on the racetrack. If you don't put the right sort of structure around your
search, you know, the thing is it's just making sure that whatever team you put around your
search, you know, it's the right team. So you've got to, when you pull into the pit
lane, you want to be in and out really quickly. If something goes wrong, you know, and I'm
using that analogy to simple, simplify it down for the general person out there that's
(27:14):
listening who's a shopper and not really running an e-commerce website, for example, you know,
you've got to, you've got to put the right team around it to make it work for you because
it is such a powerful tool. But at the same time, if you want to be hands off, you can,
you know, engage with you guys and get it all set up and, and, and do that. And it's
a really good point about, you know, black Friday and a lot of the revenue, actually,
(27:40):
I think a lot of retailers are banking on having a big final quarter. And I think in
my mind, what I used to do, and it's a, I don't know if you're thinking about it or
not out there in listening land, but if you've got products, you know, that you've got coming
back this year, it's not a category you normally delve in, but you've got, you know, quantities
(28:01):
that you want to sell for black Friday. I would make sure that you check your search,
make sure that your search is picking up, you got your metadata structured correctly
and go back and have a look at last year's search data in Algolia, because there is so
many good reports and analytics. There's so much intelligence in the Algolia platform.
If you are already a customer, if you've got Google analytics and you're just using that,
(28:26):
that's fine too. You can just go in and have a look at your search reports for last year
and have a look at the queries, run those queries again, and make sure that something
is appearing in those search terms and you're not serving up white pages of, sorry, there's
no products in that category or we can't find what you're looking for. Yeah, no results
or whatever. And cause that's just a market is nightmare. And from a business owner's
point of view, if I was a business owner looking at that and not finding the products that
(28:49):
I know I've got sitting in the warehouse, for example, in search, cause someone along
the line hasn't put the right content in or, you know, it's not in the right category because
it always happens. It can go wrong, it will go wrong. I would say a tip from me would
be to make sure you have, you know, your Black Friday search data looked at and make sure
that it's, you know, you've got product appearing in those categories because it is super important.
(29:13):
You know, you get conversions cause yeah, a lot of retailers are banking on that big
quarter for the rest of the year, you know. So search IO, so in Australia, our goal is
47 staff now in total. Okay. And a chunk of those came from search. Yeah. Well, a large
portion as in data scientists, engineers all came from the search IO acquisition back in
(29:35):
September. I met with search IO, sorry, a while ago. I remember they, they had a cool
feature which was, I think, and correct me if I'm wrong, they had a, like a visual search.
You could search over your database and it would find the black t-shirt. It would find
into our product roadmap. So visual neural is what we refer to. And that may have to
(29:57):
like put something visual. What you just, you said that exactly what it is, right? It's
the idea that we are looking for opportunities to fill in the gaps of the data that's missing.
We're not going to be granular in a sense of, we know that there's a small shoe here.
We're going to surface a small shoe where you don't have in stock, but the reality is
through search IO plus our own engine, we're basically bringing those two worlds together.
(30:23):
And the real nut that search IO cracked that we, we as an organization, we're leaning towards
this is the idea of creating basically vector hashes or these neural hashes. And what they
are are pockets or packets of data specific to your individual use case that allow us
to basically provide the keyword and like a closest proximity to that and bind them
(30:48):
together and service it up. So when, and again, vector search and like the big retailers across
the globe have tried this for a long time. They recognize that running a data warehouse
is very expensive. And so the search IO team have been done an amazing job to build this
is neural hashing. What that means is it allows us to do stuff at scale. So we, when we build
(31:11):
our profiles for customers, it is a learning model. It's a large language model that we
use from the, you know, that everyone uses basically the large language model that is
sourced by the internet. But what we've done is almost applied our own logic of AI top
of it and said, okay, well for this retailer, we know they don't sell fishing rods. So why
(31:31):
is that part of a large language model, for example, or whatever it is, but combine those
keywords and this vector search and neural hashing allows us to provide a really efficient
fast. So we don't lose the speed. We improve the relevancy and to turn it on is literally
a toggle. So what you just said there about Black Friday, we've taken all our customers
now is like for all of those results that don't get, you know, like a competing result
(31:57):
of buying signal against it for no results pages for those category pages that are obscure
or the way that you're putting in your terminology, the statistics are in, you know, we've gone
from two years ago, it was 2.6 words. Now it's 8.9 or something like that words per query.
So really expressions are getting, well, that's, that's interesting. Yeah. Yeah. Super interesting.
(32:19):
Yeah. I didn't know that. So the idea that you're searching for, head to my sentences
now really it's a sentence. Yeah. It's a full sentence. Headphones to wear in the gym. I
sweat a lot small. When you think about that, this, that's a statement. Yep. There's no
keywords that the headphones is it, what else are you going to do? Like, you know, you could
be showing sweaters and headphones. Absolutely. I mean, you know, so what we do with that
(32:45):
and we call that is completely different. Yeah, exactly. And there are, if what you
did before, when you're saying you're looking back at your Google search results, there
are millions and billions of queries just like this and you're going on serve because
there's no keyword matching. And so what our goalie has done is not just address that,
but we've applied the same neural module to the front end to improve the query. So understanding
(33:09):
intent to help with the retrieval. So we understand that it can be served by keyword vector. Let's
apply another vector model on top of it to provide a better result from retrieval. If
you want to see in like page nation, you want to see it in different modeling. And then
of course, the most probably the most important part of the business is what do you want to
do? Do you want to sell more of them? Do you want to sell less of them? Do you want to
(33:30):
sell more margin? All that kind of business around that. So no more rules, the ability
for you to build custom rules and logic when you're surfacing. If someone puts in their
headphone or head, you don't want to be surfacing phones or head mask or whatever it is. You
want to understand the full content. So you get rid of rules, get rid of synonyms. You
get rid of the idea that keywords have to be improved or attributes have to be added
(33:53):
to support that. And there's no more kind of hacking of the platform. You don't have
to natural languages are already included with it. So you're basically bridging that
gap to helping a customer being understood. But that technology, it's just amazing. And
this is, this is, this is homegrown stuff, right? It's insane. Sorry. I geek out, but
it's a really amazing shift.
No, no, no. Look, that's why I've got my hat here, man. I was going to literally put
(34:15):
it on so the propeller could start spinning, but I love it. It gets my propeller spinning
too, mate. Don't worry about that. I love it. I think, and a lot of our audience out
there, you know, who are retailers, they want to understand more of this. It's a perfect
way for us to just give them a bit of a, a bit of insight into the importance of search
because consumers are always a moving target. And I speak about this quite a bit from a,
(34:40):
from a customer perspective and brands perspective. So there's two sides of it. You've got the
brand who's always trying to maintain relevance. I mean, there is no better way to maintain
the relevance with your audience and making sure your search results on your website match
what people are thinking about looking for and what they're expecting. If you're not,
(35:01):
you should be already in there and looking at it and keep it top of mind because it speaks
to future revenue. It speaks to future delegate with what you're saying. So you, you will
know that, right? If you're an executive and you're seeing your ad spend go up, that's
a leading indicator of one or two things. One is they don't trust the site. So they'll
(35:21):
go back to Google and search vendor this and they'll go, okay, Google has called my site
better and serves me better relevance. And you'll see that come through and they'll go.
And so that should be a big leading indicator that they're not searching on site because
they don't trust it or they can't find it. Right?
Well yeah, early days. I remember there were a lot of competitors to Algolia and their
(35:42):
search were actually better optimized than the websites were. Have you got any brands
or any examples of brands? Maybe one story of where they've introduced Algolia and seen
some really good, you don't have to speak to the exact results, but probably the best
one. The first one that comes to mind is probably culture king. So he's an Australian brand,
well-known Shopify account. You know, they got acquired by AKA brands two years ago.
(36:07):
So Simon's no longer, he's obviously still investing, but they still involved in business.
They launched in Vegas last year, Black Friday for them. They saw one like 5.2% increase
in conversions just through some baseline tweaks. And on this website here, it says
trending and shop by new. That's our recommend platform as well. So because we helped them
(36:32):
make money, they actually reinvested in us. So they said like, we actually were helping
them drive their business case for them to continue to invest. So when you add to cart,
the idea that you're getting frequently brought together, the idea that you can also now do
things where you can add capabilities inside and gauge your customer is really what we've
done for them. So the culture kings example is probably the one that comes first to mind
(36:54):
because they're Australian brand going international. And that excites me because they basically
drive innovation for us. When you think about metrics though, like what we care for them
is conversion is the most important through. So conversion lifts through searching conversions,
average order values, exit rate, time on site, time in app, how do we help them from a net
(37:15):
promoter score or minimizing churn? They all come part of the kind of complete package
of us. And the call out for this one is that the customer success manager, Captain Thorogood,
that basically works with the culture kings team. She is their SME when it comes to this
stuff, the subject matter expert. She helps them using other customer examples and shows
them what to do. And I think that's the real, the gel here is that we just want you to sell
(37:40):
your platforms and leave it all to yourself. And I think you probably have done it in the
past. The new way is don't do stuff that's going to take you months, get stuff that our
goal is going out of the box and it'll take you days, days to implement and maybe a week
to do regression testing and make sure it works and then put it out in the wild and
A B test and see the results. That's what you need to do. Iterate, iterate, iterate.
(38:01):
And I think that's the most important. That's exactly what culture kings have done with
us in their strategy. For those listening, I'm doing a search and I'm searching for balls
hat, which is the type of hat I'm wearing at the moment because I'm a huge Michael Jordan
fan. And it's, you can see it's literally immediate the moment you search. And if you
(38:21):
just do a search for balls, for example, I've just do a misspelling where I use two U's
straight away. You are exactly what I'm searching for. All that stuff should be table stakes
and we, I think we gloss over it, but auto suggestion, you'd have to be logged in to
know that you're an existing customer. So the idea that there's, there's personalization
(38:42):
out of the box. That stuff should be like 101. And I say that like, liberally that,
you know, when you, when you are looking at search, make sure they do the, the, the baseline
stuff good because if they haven't thought about that and they're trying to use AI to
fudge it, then you don't have any control over what's being produced. Culture kings
will drill us. They'll go, why is this coming up here? What's the merchandising strategy
(39:04):
here? And we'll help them guide them through what data they've given us. It's very much
a great job. I mean, this site looks really good, Bruce. And I think that the, the fact
that they've got our goal is search running on there just makes it so easy because you've
got, when you do a search for those listening, you know, you've got all the categorization
on the left. You've got genders, you've got brand size, color, discount price buckets,
(39:27):
size sliders. So you've got different ways to interact with the results as you would
expect. But the fact that it is so quick, you don't feel fatigued. You know, sometimes
you go to websites, you do some searching and then after you kind of want to go and
have a glass of wine and sit down, it's just like, Oh my God, I feel exhausted. Still search
for Jordan's off. I don't even misspell it, but bang straightaway. Jordan's bang. It's
got a landing page. Jordan's, if you're looking for Jordan says keyword data to re to reaffirm
(39:51):
that you're at the right place. And let's look at this website. I'm saying right now,
the reason why there's a shirt, two shirts come up in that first tile. It's been merchandise
that way, right? Yeah. You know, they're products that they want to move. I imagine I'm just
making a statement, but the reason why it's not coming up with all Jordan's is because
Jordan is a brand. It has, has clothing as well as exactly. Yeah. Yes. If I, if I, if
(40:12):
I search for Jordan shoes, well it has on the left footwear and they have 184. So I'm
medium. Well, that's, I'm impressed. That's pretty cool. Uh, I bought myself a pair of
these actually these, um, retro five Concords and, uh, they have the new zooms, um, 90s,
95 zooms that they're sick and they're the only place that has them. No one else has
(40:32):
them. And we're in the Melbourne store. I showed my son, he looked at it and he kind
of looked funny cause I got it like a big bubble. I'll show you in a minute. Um, but
I looked on the website first before and he didn't even know. And within a few seconds,
I'm like, let's go to culture Kings. I want to have a look. I'll show you cause he had
never seen it walked in and he was blown away, but I showed him the shoes. And the other
thing that they do really well is they show you the sizes that are available at the search
(40:54):
results page, which is so good because why am I going to waste my time clicking through
to something that it's not available in my size? All right. You have no idea how like,
and I don't have bad company, but like you've became a store that came out.com and these
aren't out. These aren't, we aren't pairing these ones and bunnings. They're just a frustrating
experience. They really are. And you had all these products at Christmas time and you're
(41:17):
like, I want to check out, but there's none available. I mean, like, are you kidding me?
So that, that, that, that may not go to the search vendor per se, but we, as a vendor
will be saying, guys, you need to fix this. Like we will help you fix this, but you need
to fix it. Surface relevant inventory because if you, if you're not, you're frustrating
your customers and I know they're already frustrated by this. So, um, you know, this
(41:38):
experience here is not the same as what, what, what were you trying to say? I shouldn't say
I shouldn't recall that.
No, no, no. Well, it's, it's, it's funny because my wife, um, said she, she hates this search.
It's not, I actually was going to ask you, I wasn't going to throw a grenade your way,
but I was going to say is, is Kmart like one of yours? So here's an example of what I'm
(41:59):
talking about. So if I search for headphones without a space, there's a hundred products
available. If I search for headphones with a space, there's 103 products available. Now
it's an 80 20 principle. If, if, um, you know, 80% of my revenue is generated by 20% of my
skews and that 20% is represented by the three that I'm missing in, without the space, I'm
(42:24):
potentially missing half the conversions based on the search terms that are for headphones.
So you're thinking about this with data. I don't know. I hope people in retail land are
thinking about this with data back data, cause it's there. If you don't have the data and
you're not using the data to drive that, that data, different decision making process, then
yeah, you've lost, you're losing the game altogether. It's my assumption that most of
(42:48):
them are like, we're all about analytics, right? Yeah. Because that's why I love it.
To be honest, it's all about data. The data that's getting fed into it. Yeah. Sorry to
cut you off.
Sorry, man. No, you're right. For another time, I'll hopefully come back and talk. I
hope this has been a good session about the Chicago. I spent two years in Chicago and
the retail experience and working with retailers in platform. Amazing experience. And I think
(43:11):
there's definitely different, different markets, but yeah, lots to unpack there as well.
There was one more topic, but we're not going to have time to get onto that, but it was
really talking about the ROI. We sort of did touch on it a bit, but I think if, oh, it's
there 386%. Like, like developers, like millions of dollars in their time will basically help
your business users better. Like it's there. Like we've actually done the analysis on
(43:35):
different retail segments, literally 800, 386% return on investment is the minimum.
Yeah. So it's a no brainer. Yep. But all I'd say is you've got to balance your, the cost
of implementation. That's what we don't know. And again, you only know your business better
than you do. So let's make sure you understand that. Yeah. That will make sense. Yeah. If
(43:58):
someone was looking to fix their search before, say for example, Christmas, is it still, is
there still time? Absolutely. Yeah. Walgreens 16 weeks, 10,000 stores, the second largest
pharmacy in the U S 10,000 stores, 50,000 skews per store. And they were competing with
Amazon like we all are. Yeah. They had to deliver 30 minutes and we did, did that for
(44:21):
them. We actually helped them achieve that within 16 weeks. All right. I would say Bruce,
if they're, if they want to fix search for them to reach out to you and start a conversation,
I'll leave your details on the show notes. I'll link to, if you've got anything to, to
show the audience in relation to the ROI, if you've done that research, I can put a
(44:44):
link up on the blog. It's been awesome having you on. Thank you so much again for agreeing
to come onto the show and I'd love to have you back on. I really appreciate it. I'm so
excited to see who else you bring onto the show. And yeah, I'm always listening now.
Have a great week. Thanks buddy. Love it. Thanks mate.
Welcome to the show, Bruce. How are you going?
(00:14):
Mate Nigel, thank you so much for having me. Long time, this is our first time caller.
That's what they say, isn't it?
You're welcome, mate. I'm really excited to have you on the show today. And when I reached
out to you a few weeks ago, and I was thinking, you know, we were talking about personalization
a few weeks ago, and I was thinking to myself, search is next. And the search piece is such
(00:36):
an important piece of, you know, e-commerce and retail at the moment. So I was thinking,
I have to reach out to Bruce and we'll get him on the show. So thank you so much for
joining the show today.
No, no, I appreciate it. And again, you've hit the nail on the head with search becoming
more relevant. We've seen obviously an influx of inquiries, probably in the back of chat
and CPT and open R. That's been the people have now talked about the conversation versus,
(01:01):
you know, trying to serve them up what they have. And that's what's really exciting about
our markets. It's in this kind of flux state. And I think our goal is, you know, as a business,
I can explain what our goal is in a minute, but I think we're really well positioned to
help not only just retail, but help other call it verticals, digitize their practice
and potentially monetize it as well. So for us, we're in a really good spot at the moment
(01:24):
on the back of all this AI, Vector, ML capabilities that's actually been more accessible to everyday
consumers, which is basically driving us to drive our customers as well.
Yeah, yeah, no, absolutely. Yeah, it's interesting, isn't it? You're talking about digitizing
the businesses and digital transformation, right? I mean, it's really on the minds of
(01:47):
a lot of people have been talking to a few CEOs and things at the moment, CFOs and heads
of retail and so on and, and, and e-commerce. And yeah, I can see that there's, you know,
since 2020, 2021, there's more of a focus on their online experience and their, and
their websites and things. And as we know, personalization piece is a big piece, but
(02:08):
even bigger than that, I think is obviously search and discovery starts with either browsing
via the main navigation on a website by categorization, or it starts with search. And I'm finding
that, and I don't know what you're seeing, maybe you can shine some, some light on that,
but how search has really changed in the last say 10 years, when Magento first released,
(02:29):
obviously they had solo search, you would go to a website and all of the listeners would
know you go to site and you'd search for something obscure, let's just say, or not so obscure.
It didn't really matter. The results were always different. You know, you'd search for
the same thing even, you know, two times you would find different results. So we know search
is getting better and better. You know, one of the reasons I actually met or discovered
(02:53):
Algolia was in the startup area of a Magento conference. And at the time I met Gaitan and
we can go into that story in a minute, but I was looking for a solution for search and,
and nobody really had anything. I was looking for something that I could give to my developers
to say, Hey guys, why don't you check this out? Meanwhile, they're on the hamster wheel
trying to fix, you know, the search.
(03:16):
The airplane in the air with the engine. What was the catalyst for you to do that though?
Was it just simply that you weren't seeing, you knew that search was a problem because
customer feedback or, you know, was it that you recognize that if we're going to make
some quick wins, as you said, it's like, can I be able to browse in the search and they're
the kind of two vehicles of how they can engage with your platform. What was the kind of driver
(03:37):
there? Was it just simply that you're looking for?
Well, I think there's probably a handful of drives. I think the main one was the customer
experience. Obviously it's frustrating and that's the first thing you try to fix. You
keep your customers happy so that you always put them first. Then I think it was, you know,
just generally the business, you know, when you're talking to suppliers and they're searching
for their product or SKUs or part of a model number and it's not appearing in our search,
(04:00):
they think it's not on the website and then it creates this internal kind of turmoil and
it just chews up time and it's really bad for productivity. So you've got the customer
side, the retail side where they can't find your product reliably via search. And then
you've got the internal battle that's going on with trying to explain to people, unless
you say I won't name names, buyers, let's just say that no, the product is there, but
(04:26):
search is just rubbish, you know, how it was back in the day and, you know, and that we're
working on it. But obviously it's very hard to have that internal conversation about how
complicated search is. We try to explain latent semantic indexing and, you know, search algorithms
and things to someone who's just interested in buying and selling stuff and making money
in between, which is basically retail. It's quite a complex conversation to have and it's
(04:50):
very hard to sit there and justify why it keeps happening, you know, and it's just the
technology hadn't caught up to, you know, the fact that consumers were really now starting
to embrace online. So really it was, that was the key drivers, you know, consumer, so
customer experience and, you know, internal pressures from not being able to find product.
It's actually, you've kind of hit the nail on the head and I'll go back and talk you
(05:13):
through kind of the evolution of where search has come from, but those three users that
you talk about, you have the developer, they're kind of like, how do we get these guys to
build something that's going to service both our customers and our business users? That's
the, you know, the stakeholders and, you know, how do we merchandise the site? And I think
(05:34):
those three users, the customer, the business user, and really the merchandising team who
are responsible for making sure product comes up and is relevant and it's moving, right?
But I think search as it is today and where it's come from, you know, search has been
around for like the, like the 90s, I think that 70s and 80s in a decade kind of built
the first initial call it multifaceted search. But when you look at solar and elastic search,
(05:59):
the underlying premise of that is a thing called the CERN, it's Apache CERN technology.
It's basically a mathematical equation that looks for words or statements or phrases and
then puts a weight into it. And that's how you build the algorithm. The way that basically
the two founders of Algolia came up with our technology and that's where the name Algolia
(06:20):
comes from, it's algorithm. It's based on this tie breaking approach. So it's, so if
you look at search across all the technologies today and we'll get into vectors in a minute,
but it's under the premise of keywords. And obviously each SKU has a stock unit or each
product has attributes assigned to its size, whatever it is. And obviously we take those
(06:40):
attributes and make them relevant to a search inquiry. So small t-shirt, large t-shirt located
next to like close to me. But what we've noticed though, as the evolution has come through
is that A, that the quality of data is really important as in what data is being serviced
for these different products and SKUs. And then more importantly, how to understand the
customer's needs, what we're actually asking for. So keywords are great for the SKUs because
(07:03):
they're all embedded there, but trying to understand where the customers are coming
from. We used to build synonyms. So t-shirt, is it spelled T-E-E, is it spelled T--shirt,
is it spelled T-space shirt or all together? So it's like five different synonyms. Your
poor business users or the merchandisers sitting there, writing these things out and out and
(07:24):
out. And so I think where we've evolved from that era where it's been very much about weighting
words and terminology is actually helping with the automation side. And that's kind
of where that vector technology comes, which we'll go into in a minute, but I think that
would be really important to discuss because it really is an evolution that's been around
for a good 10 years now. That's only become more accessible. So now you don't have to
(07:47):
hire basically a data scientist. You don't have to host a data lake. Then you have to
host it like a DXP platform to stitch all together and then serve that. It's all kind
of been delivered as a service now. And that's really, I think where, I suppose where search
has exploded over the last probably two or three years now. COVID has demonstrated to
not only just the e-comm world, but to the business world. So the C levels of these large
(08:10):
institutions across all verticals, it doesn't matter if it's media, traditional FMC, geo
retail. Even the sports industry is now trying to monetize digital engagements. And I think
e-commerce is across the board. And I think that's where you'll find the evolutions coming
through for search. Go to the customers and they'll tell you what they want because they've
(08:31):
got Netflix, they've got Spotify, they've got Google, they've got all that rich, very
personalized content being served to them as they need it. Now we have to case you catch
up to that. And that's kind of where this platform of our goal really plays into quite
nicely.
On the point of knowing or understanding what your customers are asking for. And I think
that's a, that's a really good point because people think a lot of times businesses will
(08:56):
say to me, well, when I search for this, I'm expecting this. I'm saying, yeah, but hang
on a second. You're expecting it because you're, you're predicting what the future outcome
is going to be for that search term. But a customer who's in a discovery phase, they
might be searching around the topic. They're not actually, they haven't got to the point
of where they're searching for that particular term you're, you're trying to, you're searching
(09:17):
for now. So there is, and this is starts a conversation with how search actually works
and the intense signals and things that people actually show before they get to your site
and collecting that information, using it to understand. And I think for me, a good
example would be we looked at JB Hi-Fi. We did a whole bunch of, you know, our research
(09:39):
and I was leading that research. And one of the things we noticed was like, if you search
for radio head, right? So JB have media content, obviously they have albums and yep, media
and radio head would show up. But if you just quickly searching for radio, expecting to
see radio head, cause all you look at JB Hi-Fi is for example, their albums and so forth.
(09:59):
You're really not interested in digital radios, right? There's a lot of different use cases
we'd worked out that those intense signals aren't just the primer for the search path.
It's also what have they previously looked at in the past, wherever they come from on
Google search, for example, you know, all of these different data points to understand,
(10:20):
okay, when they do come, if they do search for something, it's more than likely going
to be weighted across the media section of the business, for example, albums and records
and DVDs or whatever. So I recommend you should go back to their website in about a month's
time staples.ca the staples can, uh, Canadian store. They've got this, it's kind of our
(10:41):
example side, but we've, we've just created a thing called query categorization. Um, and
so it's, it's in session experience. So the concept is, you know, if you're engaged with
JB and you've, uh, you're looking for albums or we see a lot of use case for coffee. Are
you looking for ground coffee or beans? It was the coffee machines, coffee pods, coffee
(11:03):
tables. And so the idea that you have all that product, how do you categorize it so
that the next engagement that you have with the platform is it's predicting that intent.
And that's kind of really what's the evolution is. So all those buying signals that you've
already captured, be it their net new customer, they're accustomed with a, like a very weak
(11:24):
profile or they're a well-known customer to you. How do you service those three customers?
And you know, they're all about different personas and buying habits. And that's what
we're trying to do now is, uh, I'm trying to be the, the end state, which you're your
full customer profile, because that, that runs the risk of us holding onto data that
we shouldn't do. But I think we can definitely supplement that with either third party data
(11:45):
or the stuff, the buying signals that we capture and which you would know about.
Yeah. All the tags and things. Yep. That, that idea of understanding what your, what
your intent is, um, be it cause you're a net new customer, there's no understanding
about who you are. That's a really important thing to, for a search provider to do because
it's surfacing relevant results. And that's the most important thing. That's what we designed
(12:06):
to do, right? Solve your problem, solve your need, help you be understood to this retailer
and then deliver, deliver a product service capability back then really quickly.
Yeah, that's right. I mean, if you go, just on that example, if you go search for coffee
and you've been looking at coffee beans, but you also sell machines and grinders and everything
else, like you, you, you, and you're constantly buying coffee off that particular retailer.
(12:29):
For example, the last thing you want to do is refactor those results again and put to
the top coffee machines when you know that all they, all they buy off you is coffee.
You'd want to put them further on down the list or highlight coffee machines and say,
by the way, did you know we've got an offer here? We can give you 20% off a coffee machine.
Is your coffee machine good enough? Maybe start them down that path. If you want to
(12:50):
do as a marketer, you could quite easily do that. But I think that's a really, really
good example. Yeah. Back to your point about Gaetan and I'll touch on that story. It's
quite interesting. So please tell us location. Tell us, okay, so Magento conference 2016.
I think it was 20. Magento. Were they owned by Adobe by then? No, no, no, no. This was
(13:13):
um, yeah. Yeah. Yeah. Yeah. Yeah. Yeah. So it was a great conference. You know, he had
this little tiny booth and Gaetan was there and I'd messaged him and I said, Hey, I don't
think it was at his booth. I think he was with a client. So, so we sat down and we had
a coffee and he was telling me about search. And when I was explaining to him all these
issues and like he'd heard it all before. So he knew exactly what to expect. He goes,
(13:34):
well, look, I can solve your problems, but I'm going to have to show you on my little
laptop and he had like a little laptop and he just had his phone and it was tethered
to like a, it was 3G I think connection. And it was, and I was having issues too. It was
three and 4G, right? So this is how long ago it was. There was no 5G. I'm thinking how
the hell is he going to tether his, his notebook to his phone and then run a search query and
(13:56):
get back anything in enough time to, to impress a customer. So I literally saw him, he dialed
in pulled up a query screen and you could see that it was live. There was no caching
or anything like that. And he goes to something he was searching over something. It's a huge
amount and he searched and I had to go as well, but he searched for a couple of search
terms and just, it was instant. And I looked at it. I looked at him and I was just, I was
(14:20):
just gobsmacked. I'm like, where's the spinny wheel thing? You know, the spinning wheel,
like Ajax rubber. I'm like, where's that? Where's that gone? And I'm thinking, what
this is blowing me away. And he's like, Oh no, no, you know, he started and look, I,
on purpose, I brought my propeller, my propeller hat. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah.
(14:41):
I should put my propeller hat on right now. So I can start spinning, but yeah, it brought
back all these results. And I searched a few things. I think it was a, a taxation library
of, you know, all this law or something. So it was all text-based. And it was just instant
and then all the refactoring there. And it had the most popular pages and it had, and
(15:01):
I'm thinking, and I looked at him and at the time I said, how much is it? And he told me
and I went, sign me up. That's it. I'm done. You're in. I mean, that was it. What was the
result though? Like, did you see a, like it was an increase in revenue? Was it a, was
it a conversions? Everything. I mean, it's conversions, revenue conversions, but average
order order amount. But what I was basing the success metric for me obviously was conversions,
(15:25):
but it was actually the number of products displayed per search term inside a category
that I knew was our core business. So what did I do? Because we didn't have any analysts
or anything where I was working at the time. So what I did was I took all of the keyword
search terms out of our Google account, out of our AdWords account. I then overlaid that
(15:46):
with the number of search or the number of products returning results on the site based
on that, on the ad spend the ad keywords that we had. I expected all brand terms, all the
store locator terms. It was just more product terms. So you actually waiting to terms as
well because of the cost of them. Yeah. So I aligned our margins per category. So the
(16:09):
category, so I went from the, the, I worked back from the margins and the popularity of
the products. So stock turns, then margin, margin, so availability, the margin, availability.
Yeah. Correct. So availability was, was third. All right. So because the thing he did, this
was a debate because you don't always want to show things that are in stock over things
(16:32):
that are out of stock if they're not what the intent of the search is. So for example,
if for example, Sony TV number one has the most stock turns and has the most margin,
but he's out of stock and I put that third or fourth, it's not likely to result in a
conversion, but if I put it first, they'll might sign up for when it comes back into
stock or go in store and place an order in store. Okay. And white. And that's a lot of,
(16:55):
a lot of retailers didn't realize that because a lot of people were just looking online and
going in store. So I would wait it based on that. And I could see what we were selling.
So I had all the reports so I could see the impact. It was very hard for me to show the
business what that impact was because Google would just have a row as multiplier. Yeah.
And then I worked back from that. And then I said, okay, well I've only got five products
(17:16):
showing there's 50 products in that particular category, but I'm search on that search term.
Our click through rate on our ads is really high. The quality score is really high on
our ads. And I can see, you know, we've got a dedicated landing page, but the search results
when they get there, they hit the landing page, but then they go to search. For example,
I start looking at the range and so I did it that way. So I worked back in, in that,
(17:37):
in that way to work out how to work out the ROI. Yeah. Yeah. I mean, that's, that's exactly
what I mean, that is literally the business model today. Tell us your business metrics.
Are you trying to move your stuff? You have built your own, basically you're your own
BA and your own almost yeah, like almost auto management systems to a certain degree, right?
(17:58):
Cause you're looking at inventory in real time and what the data, but what the data
piece is exactly the, that is exactly what we're about. We're the second largest search
vendor behind Google by a fair way. But if you add up all the other search vendors like
your Yahoo's and your bins times and by four, and that's kind of the size of our goal. Yeah.
Wow. That's huge. Yeah. We're seeing this wealth of information. What you just said
(18:21):
there about all these conversion metrics. If we capture that from all our customers,
we can predict what's happening globally. Now for us as a very, very PI and GDPR company,
we will never sell that data. That's part of our DPA. Google does though. And that's
how Google does their evolution. They take their analytic traffic and then they use that
(18:41):
to influence understanding. And that's what drives it. So it's a completely different
approach. Google's for the masses. We're kind of for you as the consumer. And that's
why we've kind of made our system quite closed. We don't, you can influence that by third
party buying signals and third party data, like a Google analytics or a GTM tagging and
things like that. And be extensible. But the reality is you have to understand your business
(19:05):
the best so that we can help make the relevancy change. If you need to move more product,
more margin, if you've got customer profiles, we'll help understand that for you and give
you lots of ways to do it. But you were bleeding edge man way back then doing that, looking
at the buying signals and then augmenting or merchandising the e-commerce experience
(19:25):
and the search experience and the results that are being rendered, which is amazing.
Yeah. Thanks man. Yeah. I mean, it's just retail in the day. But now, I mean, if you
look back over the past 10 years, Algolia has really been just on the point of your,
you know, the number of search queries. It's been basically part of the evolution of search
(19:45):
and it's changed the industry, the e-commerce industry globally. I think that's as profound
the impact that Algolia has actually had. And you see a lot of spin-offs now, different
products going, but they don't have the underlying purpose because it's purpose that, I don't
want to sound like a line from the matrix, but it's purpose that drives everybody, right?
You've got to have a purpose. Algolia was for the developers. I remember sitting at,
(20:10):
and I think it was a year, a couple of years after when I met Gaetan and it was already
customer and it had proven that it was working, you know, and there's always these little
teething things that we had, but that was sorted straight away pretty much. And that's
important, you know, as a product and as a retailer, you can slam your hand down on the
desk and say, oh, you know, this is not acceptable, but you've got to work with people as well.
(20:30):
You've got to work and you've got to be prepared to put the right, sell it internally and have
the right mindset of how search is. It's never perfect. It aims to be perfect. But as you
say, with the advent of AI and everything else, eventually it will be fully AI. It will
understand what you're looking for, what your preferences are in terms of brand. Like it
will, it will just know in order to just happen. And like, what do they say? What's that saying?
(20:54):
And technology can be indistinguishable from magic. And I think eventually you just go
to a site and expect it. And it'll just be, as you say, you've got all of this data. It'll
just work. It'll just have the similar search experience across most websites. I think that's
probably one of the most frustrating things. You go to a site and the search is not as
(21:15):
good as another site that you've been to. And it kind of, it sort of falls short in
your expectations that you want to shop there because it just becomes really frustrating.
Yeah. Thanks man. That's super interesting.
No, no. Even just unpacking that a little bit more. Like if you look at our customer
base and it's heavily weighed towards retail. So JB Hi-Fi, Bingley, Culture Kings, we've
(21:39):
got Woolworths as a customer. And then when I look at our in-store base, it's like 20,000
customers. I think we say 17,000, but some of these customers have multiple customers
and they've won. And so all that 60% of those retailers, they're helping us innovate. Our
customers send feature requests like it's no one's business because they do, they expect
(22:01):
that right? Their customers are asking them and they're asking us. And so that only makes
us fight more for that capability with a development base of like 350 odd team. We've got so many
developers and different capabilities trying to work on different strategies to help improve
that. One of those ones we'll talk about, which is SearchIO. We acquired them in Australian
(22:23):
business and I would highly recommend you get Hamish on. Hamish was one of the co-founders
of SearchIO. He's kind of like the godfather for us when it comes to vector search and
this idea of natural language, large language models. He is amazing in terms of what his
perception is. But when I say perception about meeting the customer where they are, I just
(22:44):
did a session in commerce tools, like an e-com for Shopify or Magento. And they were talking
about the advent of the mobile device being the full platform offering you through WeChat
in China. You buy your groceries, you pay homeless people on the street if you want
to tip them. It's all QR codes. People just want to be understood. And for us, you said
(23:10):
it before. When you think about their experiences, be it in-store experience where they're trying
to be understood searching for something, I need to find the toilet. I need to find
Nike's that fit a kid that's aged five, putting size in there. I'm looking for a lilac sweater
or a lavender sweater. Now, if you don't have purple or mauve or whatever the color
(23:31):
of shade it is, from a keyword, you're screwed. They're not going to come back with any relevancy.
It's going to give you a jumper, hoodie, jacket, pullover, whatever it is. And so I think that
understanding to whatever platform it is, be it a point of sale device, be it in-store
experience, be it digital apps, be it for business users as well, trying to do searches
(23:51):
on give us the latest inventory levels and what products are moving quickest like what
you did before. That's where we need to help the customer be better serviced.
The browse experience and content that's being served, I think that also is important because
those digital natives that are coming from Instagram, Facebook, and only these social
platforms, they want to feel that same brand and trust come through to whatever platform
(24:15):
or brand they're touching. And so I think getting that consistent message that you were
saying before, going to one site, getting let down and then going to another site and
being, wow, for the audience, look, search converts. There's no doubt about it. For every
person that engages with search and look at Google Analytics, jump on there and say, people
that are having search sessions, what's the conversion rate now compared to browse sessions?
(24:38):
It's that simple. If you're doing it right, it should be in the realm of like five to
six X. That's five to six times conversions through search than it is to browse. Because
customers are coming to the website either as a known quantity or an unknown quantity
and you're just trying to work out which bucket you're going to fit in and then help them
convert for you. One thing I would say though is when you think about those pillars, you
(25:00):
need to be AI movement and where you buy one platform or one capability and it solves all
your problems. I think you got to be aware that there's you as a consumer and you as
a vendor, as a brand, you know your customers better than anyone. You can't expect some
third party coming in and say, just turn on the tweaks of AI and it'll run itself. You
(25:23):
still need to have some level of input to help drive your signals, your buying signals.
Because we know everyone wants that Nike because on sale and there's none of it left. So they
want to keep coming back and asking for it. And to the idea of, I know you're looking
for the size 10 and I know you're coming from this location. So how do I help you find it
in store or how do I help? That may not serve us their needs, but at least it gives them
(25:46):
opportunities to help discover what else we have to offer for you and see what else you're
interested in. And then once you've got them in that buying mode and try and help them
discover what you have, they're like, oh wow, you've got heaps of stuff, right? This is
feels like an endless aisle experience kind of approach. When you think about Algolia's
technology, that's actually fundamentally our biggest offering in terms of capability.
When it comes to speed and the speed of those results, not only to index data really quickly,
(26:10):
but more importantly to surface that data back to mobile device. When we talk about
speed, we're talking about like sub 20 milliseconds. If you're the other spectrum, which is a hundred
milliseconds or even like the inability to load a page and a part of that page might be
simply recommendations and you're taking like 300 milliseconds to serve that up. That's
money walking out that door for every point one of a second. It's like, like it reduces
(26:33):
your percentage of conversion by 8% black Friday coming up. How do we get the season,
the right seasonal data into play? And I think that's really fundamentally important as well.
So I'm hoping on that a bit.
No, no, I mean, it's, it's absolutely spot on. One of the things I've told many people
who've asked me about search is that, you know, it's a bit like buying a Ferrari and
(26:54):
not taking it on the racetrack. If you don't put the right sort of structure around your
search, you know, the thing is it's just making sure that whatever team you put around your
search, you know, it's the right team. So you've got to, when you pull into the pit
lane, you want to be in and out really quickly. If something goes wrong, you know, and I'm
using that analogy to simple, simplify it down for the general person out there that's
(27:14):
listening who's a shopper and not really running an e-commerce website, for example, you know,
you've got to, you've got to put the right team around it to make it work for you because
it is such a powerful tool. But at the same time, if you want to be hands off, you can,
you know, engage with you guys and get it all set up and, and, and do that. And it's
a really good point about, you know, black Friday and a lot of the revenue, actually,
(27:40):
I think a lot of retailers are banking on having a big final quarter. And I think in
my mind, what I used to do, and it's a, I don't know if you're thinking about it or
not out there in listening land, but if you've got products, you know, that you've got coming
back this year, it's not a category you normally delve in, but you've got, you know, quantities
(28:01):
that you want to sell for black Friday. I would make sure that you check your search,
make sure that your search is picking up, you got your metadata structured correctly
and go back and have a look at last year's search data in Algolia, because there is so
many good reports and analytics. There's so much intelligence in the Algolia platform.
If you are already a customer, if you've got Google analytics and you're just using that,
(28:26):
that's fine too. You can just go in and have a look at your search reports for last year
and have a look at the queries, run those queries again, and make sure that something
is appearing in those search terms and you're not serving up white pages of, sorry, there's
no products in that category or we can't find what you're looking for. Yeah, no results
or whatever. And cause that's just a market is nightmare. And from a business owner's
point of view, if I was a business owner looking at that and not finding the products that
(28:49):
I know I've got sitting in the warehouse, for example, in search, cause someone along
the line hasn't put the right content in or, you know, it's not in the right category because
it always happens. It can go wrong, it will go wrong. I would say a tip from me would
be to make sure you have, you know, your Black Friday search data looked at and make sure
that it's, you know, you've got product appearing in those categories because it is super important.
(29:13):
You know, you get conversions cause yeah, a lot of retailers are banking on that big
quarter for the rest of the year, you know. So search IO, so in Australia, our goal is
47 staff now in total. Okay. And a chunk of those came from search. Yeah. Well, a large
portion as in data scientists, engineers all came from the search IO acquisition back in
(29:35):
September. I met with search IO, sorry, a while ago. I remember they, they had a cool
feature which was, I think, and correct me if I'm wrong, they had a, like a visual search.
You could search over your database and it would find the black t-shirt. It would find
into our product roadmap. So visual neural is what we refer to. And that may have to
(29:57):
like put something visual. What you just, you said that exactly what it is, right? It's
the idea that we are looking for opportunities to fill in the gaps of the data that's missing.
We're not going to be granular in a sense of, we know that there's a small shoe here.
We're going to surface a small shoe where you don't have in stock, but the reality is
through search IO plus our own engine, we're basically bringing those two worlds together.
(30:23):
And the real nut that search IO cracked that we, we as an organization, we're leaning towards
this is the idea of creating basically vector hashes or these neural hashes. And what they
are are pockets or packets of data specific to your individual use case that allow us
to basically provide the keyword and like a closest proximity to that and bind them
(30:48):
together and service it up. So when, and again, vector search and like the big retailers across
the globe have tried this for a long time. They recognize that running a data warehouse
is very expensive. And so the search IO team have been done an amazing job to build this
is neural hashing. What that means is it allows us to do stuff at scale. So we, when we build
(31:11):
our profiles for customers, it is a learning model. It's a large language model that we
use from the, you know, that everyone uses basically the large language model that is
sourced by the internet. But what we've done is almost applied our own logic of AI top
of it and said, okay, well for this retailer, we know they don't sell fishing rods. So why
(31:31):
is that part of a large language model, for example, or whatever it is, but combine those
keywords and this vector search and neural hashing allows us to provide a really efficient
fast. So we don't lose the speed. We improve the relevancy and to turn it on is literally
a toggle. So what you just said there about Black Friday, we've taken all our customers
now is like for all of those results that don't get, you know, like a competing result
(31:57):
of buying signal against it for no results pages for those category pages that are obscure
or the way that you're putting in your terminology, the statistics are in, you know, we've gone
from two years ago, it was 2.6 words. Now it's 8.9 or something like that words per query.
So really expressions are getting, well, that's, that's interesting. Yeah. Yeah. Super interesting.
(32:19):
Yeah. I didn't know that. So the idea that you're searching for, head to my sentences
now really it's a sentence. Yeah. It's a full sentence. Headphones to wear in the gym. I
sweat a lot small. When you think about that, this, that's a statement. Yep. There's no
keywords that the headphones is it, what else are you going to do? Like, you know, you could
be showing sweaters and headphones. Absolutely. I mean, you know, so what we do with that
(32:45):
and we call that is completely different. Yeah, exactly. And there are, if what you
did before, when you're saying you're looking back at your Google search results, there
are millions and billions of queries just like this and you're going on serve because
there's no keyword matching. And so what our goalie has done is not just address that,
but we've applied the same neural module to the front end to improve the query. So understanding
(33:09):
intent to help with the retrieval. So we understand that it can be served by keyword vector. Let's
apply another vector model on top of it to provide a better result from retrieval. If
you want to see in like page nation, you want to see it in different modeling. And then
of course, the most probably the most important part of the business is what do you want to
do? Do you want to sell more of them? Do you want to sell less of them? Do you want to
(33:30):
sell more margin? All that kind of business around that. So no more rules, the ability
for you to build custom rules and logic when you're surfacing. If someone puts in their
headphone or head, you don't want to be surfacing phones or head mask or whatever it is. You
want to understand the full content. So you get rid of rules, get rid of synonyms. You
get rid of the idea that keywords have to be improved or attributes have to be added
(33:53):
to support that. And there's no more kind of hacking of the platform. You don't have
to natural languages are already included with it. So you're basically bridging that
gap to helping a customer being understood. But that technology, it's just amazing. And
this is, this is, this is homegrown stuff, right? It's insane. Sorry. I geek out, but
it's a really amazing shift.
No, no, no. Look, that's why I've got my hat here, man. I was going to literally put
(34:15):
it on so the propeller could start spinning, but I love it. It gets my propeller spinning
too, mate. Don't worry about that. I love it. I think, and a lot of our audience out
there, you know, who are retailers, they want to understand more of this. It's a perfect
way for us to just give them a bit of a, a bit of insight into the importance of search
because consumers are always a moving target. And I speak about this quite a bit from a,
(34:40):
from a customer perspective and brands perspective. So there's two sides of it. You've got the
brand who's always trying to maintain relevance. I mean, there is no better way to maintain
the relevance with your audience and making sure your search results on your website match
what people are thinking about looking for and what they're expecting. If you're not,
(35:01):
you should be already in there and looking at it and keep it top of mind because it speaks
to future revenue. It speaks to future delegate with what you're saying. So you, you will
know that, right? If you're an executive and you're seeing your ad spend go up, that's
a leading indicator of one or two things. One is they don't trust the site. So they'll
(35:21):
go back to Google and search vendor this and they'll go, okay, Google has called my site
better and serves me better relevance. And you'll see that come through and they'll go.
And so that should be a big leading indicator that they're not searching on site because
they don't trust it or they can't find it. Right?
Well yeah, early days. I remember there were a lot of competitors to Algolia and their
(35:42):
search were actually better optimized than the websites were. Have you got any brands
or any examples of brands? Maybe one story of where they've introduced Algolia and seen
some really good, you don't have to speak to the exact results, but probably the best
one. The first one that comes to mind is probably culture king. So he's an Australian brand,
well-known Shopify account. You know, they got acquired by AKA brands two years ago.
(36:07):
So Simon's no longer, he's obviously still investing, but they still involved in business.
They launched in Vegas last year, Black Friday for them. They saw one like 5.2% increase
in conversions just through some baseline tweaks. And on this website here, it says
trending and shop by new. That's our recommend platform as well. So because we helped them
(36:32):
make money, they actually reinvested in us. So they said like, we actually were helping
them drive their business case for them to continue to invest. So when you add to cart,
the idea that you're getting frequently brought together, the idea that you can also now do
things where you can add capabilities inside and gauge your customer is really what we've
done for them. So the culture kings example is probably the one that comes first to mind
(36:54):
because they're Australian brand going international. And that excites me because they basically
drive innovation for us. When you think about metrics though, like what we care for them
is conversion is the most important through. So conversion lifts through searching conversions,
average order values, exit rate, time on site, time in app, how do we help them from a net
(37:15):
promoter score or minimizing churn? They all come part of the kind of complete package
of us. And the call out for this one is that the customer success manager, Captain Thorogood,
that basically works with the culture kings team. She is their SME when it comes to this
stuff, the subject matter expert. She helps them using other customer examples and shows
them what to do. And I think that's the real, the gel here is that we just want you to sell
(37:40):
your platforms and leave it all to yourself. And I think you probably have done it in the
past. The new way is don't do stuff that's going to take you months, get stuff that our
goal is going out of the box and it'll take you days, days to implement and maybe a week
to do regression testing and make sure it works and then put it out in the wild and
A B test and see the results. That's what you need to do. Iterate, iterate, iterate.
(38:01):
And I think that's the most important. That's exactly what culture kings have done with
us in their strategy. For those listening, I'm doing a search and I'm searching for balls
hat, which is the type of hat I'm wearing at the moment because I'm a huge Michael Jordan
fan. And it's, you can see it's literally immediate the moment you search. And if you
(38:21):
just do a search for balls, for example, I've just do a misspelling where I use two U's
straight away. You are exactly what I'm searching for. All that stuff should be table stakes
and we, I think we gloss over it, but auto suggestion, you'd have to be logged in to
know that you're an existing customer. So the idea that there's, there's personalization
(38:42):
out of the box. That stuff should be like 101. And I say that like, liberally that,
you know, when you, when you are looking at search, make sure they do the, the, the baseline
stuff good because if they haven't thought about that and they're trying to use AI to
fudge it, then you don't have any control over what's being produced. Culture kings
will drill us. They'll go, why is this coming up here? What's the merchandising strategy
(39:04):
here? And we'll help them guide them through what data they've given us. It's very much
a great job. I mean, this site looks really good, Bruce. And I think that the, the fact
that they've got our goal is search running on there just makes it so easy because you've
got, when you do a search for those listening, you know, you've got all the categorization
on the left. You've got genders, you've got brand size, color, discount price buckets,
(39:27):
size sliders. So you've got different ways to interact with the results as you would
expect. But the fact that it is so quick, you don't feel fatigued. You know, sometimes
you go to websites, you do some searching and then after you kind of want to go and
have a glass of wine and sit down, it's just like, Oh my God, I feel exhausted. Still search
for Jordan's off. I don't even misspell it, but bang straightaway. Jordan's bang. It's
got a landing page. Jordan's, if you're looking for Jordan says keyword data to re to reaffirm
(39:51):
that you're at the right place. And let's look at this website. I'm saying right now,
the reason why there's a shirt, two shirts come up in that first tile. It's been merchandise
that way, right? Yeah. You know, they're products that they want to move. I imagine I'm just
making a statement, but the reason why it's not coming up with all Jordan's is because
Jordan is a brand. It has, has clothing as well as exactly. Yeah. Yes. If I, if I, if
(40:12):
I search for Jordan shoes, well it has on the left footwear and they have 184. So I'm
medium. Well, that's, I'm impressed. That's pretty cool. Uh, I bought myself a pair of
these actually these, um, retro five Concords and, uh, they have the new zooms, um, 90s,
95 zooms that they're sick and they're the only place that has them. No one else has
(40:32):
them. And we're in the Melbourne store. I showed my son, he looked at it and he kind
of looked funny cause I got it like a big bubble. I'll show you in a minute. Um, but
I looked on the website first before and he didn't even know. And within a few seconds,
I'm like, let's go to culture Kings. I want to have a look. I'll show you cause he had
never seen it walked in and he was blown away, but I showed him the shoes. And the other
thing that they do really well is they show you the sizes that are available at the search
(40:54):
results page, which is so good because why am I going to waste my time clicking through
to something that it's not available in my size? All right. You have no idea how like,
and I don't have bad company, but like you've became a store that came out.com and these
aren't out. These aren't, we aren't pairing these ones and bunnings. They're just a frustrating
experience. They really are. And you had all these products at Christmas time and you're
(41:17):
like, I want to check out, but there's none available. I mean, like, are you kidding me?
So that, that, that, that may not go to the search vendor per se, but we, as a vendor
will be saying, guys, you need to fix this. Like we will help you fix this, but you need
to fix it. Surface relevant inventory because if you, if you're not, you're frustrating
your customers and I know they're already frustrated by this. So, um, you know, this
(41:38):
experience here is not the same as what, what, what were you trying to say? I shouldn't say
I shouldn't recall that.
No, no, no. Well, it's, it's, it's funny because my wife, um, said she, she hates this search.
It's not, I actually was going to ask you, I wasn't going to throw a grenade your way,
but I was going to say is, is Kmart like one of yours? So here's an example of what I'm
(41:59):
talking about. So if I search for headphones without a space, there's a hundred products
available. If I search for headphones with a space, there's 103 products available. Now
it's an 80 20 principle. If, if, um, you know, 80% of my revenue is generated by 20% of my
skews and that 20% is represented by the three that I'm missing in, without the space, I'm
(42:24):
potentially missing half the conversions based on the search terms that are for headphones.
So you're thinking about this with data. I don't know. I hope people in retail land are
thinking about this with data back data, cause it's there. If you don't have the data and
you're not using the data to drive that, that data, different decision making process, then
yeah, you've lost, you're losing the game altogether. It's my assumption that most of
(42:48):
them are like, we're all about analytics, right? Yeah. Because that's why I love it.
To be honest, it's all about data. The data that's getting fed into it. Yeah. Sorry to
cut you off.
Sorry, man. No, you're right. For another time, I'll hopefully come back and talk. I
hope this has been a good session about the Chicago. I spent two years in Chicago and
the retail experience and working with retailers in platform. Amazing experience. And I think
(43:11):
there's definitely different, different markets, but yeah, lots to unpack there as well.
There was one more topic, but we're not going to have time to get onto that, but it was
really talking about the ROI. We sort of did touch on it a bit, but I think if, oh, it's
there 386%. Like, like developers, like millions of dollars in their time will basically help
your business users better. Like it's there. Like we've actually done the analysis on
(43:35):
different retail segments, literally 800, 386% return on investment is the minimum.
Yeah. So it's a no brainer. Yep. But all I'd say is you've got to balance your, the cost
of implementation. That's what we don't know. And again, you only know your business better
than you do. So let's make sure you understand that. Yeah. That will make sense. Yeah. If
(43:58):
someone was looking to fix their search before, say for example, Christmas, is it still, is
there still time? Absolutely. Yeah. Walgreens 16 weeks, 10,000 stores, the second largest
pharmacy in the U S 10,000 stores, 50,000 skews per store. And they were competing with
Amazon like we all are. Yeah. They had to deliver 30 minutes and we did, did that for
(44:21):
them. We actually helped them achieve that within 16 weeks. All right. I would say Bruce,
if they're, if they want to fix search for them to reach out to you and start a conversation,
I'll leave your details on the show notes. I'll link to, if you've got anything to, to
show the audience in relation to the ROI, if you've done that research, I can put a
(44:44):
link up on the blog. It's been awesome having you on. Thank you so much again for agreeing
to come onto the show and I'd love to have you back on. I really appreciate it. I'm so
excited to see who else you bring onto the show. And yeah, I'm always listening now.
Have a great week. Thanks buddy. Love it. Thanks mate.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
The Clifford Show
The Clifford Show with Clifford Taylor IV blends humor, culture, and behind-the-scenes sports talk with real conversations featuring athletes, creators, and personalities—spotlighting the grind, the growth, and the opportunities shaping the next generation of sports and culture.