What comes to mind when you hear the words ‘sales data’?
Alloy.ai’s resident expert on all things related to retail data, Manfred Reiche, says there’s a whole realm of information you’re likely missing out on when it comes to the sales data bucket. In this episode, we delve into the intricacies of point-of-sale data and its transformative impact on the consumer goods industry.
Listen in as we explore the critical role of data analysis in uncovering regional problems, managing inventory, and enhancing decision-making across various company sizes. From the importance of granularity in data to the necessity of breaking down silos within organizations, this discussion promises to shed light on how businesses can leverage vast data sets to drive success.
Three Key Takeaways:
- Accumulating data isn't enough, the key is to convert it into insights that drive action and that’s where granularity and AI play pivotal roles
- Encourage cooperation between different business units - breaking down silos can lead to more synchronized decision-making and better use of POS data
- Transition from traditional methods to using real-time POS data for forecasting to better predict market behaviors and enhance supply chain decisions
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
From Alloy AI.
This is Shelf Life.
Speaker 2 (00:19):
What are some of the common mistakes brands make when
it comes to data?
Speaker 1 (00:25):
the common mistakes brands make when it comes to
data.
How can you be using point ofsale data to power other parts
of your business, like planning,promotions and supply chain
decisions?
Speaker 2 (00:31):
How can you use data to better collaborate with
retailers?
Speaker 1 (00:35):
On every episode of Shelf Life.
We answer questions like theseand more, with the help of
leaders across the consumergoods industry.
Speaker 2 (00:43):
Today we welcome our colleague, Manfred Reichi,
Alloy's subject matter expert onall things related to retail
data and products.
He spent his time at Alloybeing hands-on with some of our
most sophisticated brands in theconsumer goods industry.
Prior to joining us at Alloy,Manfred started his career as a
technical consultant at Deloitte, focused on SAP implementations
.
I'm your co-host, Joel Beal,CEO of Alloy AI.
Speaker 1 (01:07):
And I'm your co-host, Logan Ensign, Chief Customer
Officer at Alloy AI.
Speaker 2 (01:12):
We'll be back with Manfred right after this.
Speaker 3 (01:16):
As a consumer brand, you connect with dozens of
external partners and internalsystems to get a complete
picture of your business.
Each one is different.
Alloy AI makes it easy toconnect data from retailers,
e-commerce, supply chainpartners and even your own ERP,
then easily surface insights todrive sales growth.
Every day, brands use Alloy AIto see POS trends, measure
(01:37):
promotion performance and makebetter replenishment decisions
with their retail partners.
That's why we're trusted by BIC, crayola, valvoline, melissa,
doug Bosch and many more.
Get a demo at Alloyai today.
Speaker 1 (01:51):
Manfred, welcome to Shelf Life, really good to have
you, great to be here.
I'm surprised we've done asmany episodes as we have and
haven't had you yet.
Really excited to dive into oneof my favorite topics.
A lot to talk about data,that's right.
That's right.
Well, let's dive in.
I think between the three of uswe have spoken to a ton of
companies enterprise, mid-market, smb that many think have this
(02:14):
idea of point of sale datasolved already, and then when we
peel back the curtain a littlebit, we realize that maybe
that's not the case.
And so in our experience I knowtypically we do find that a lot
of companies feel like they'repretty good at data, but
actually a lot of them aren't.
So, manny, I am interested inhearing just some specific
(02:34):
example stories of kind of thisexperience that often we have.
Speaker 4 (02:38):
One of my favorite initial stories about data, and
we have this concept in Allo.
We call language policingourselves.
Right, Like words meandifferent things and when I talk
to people, either internallywith customers here at Alloy or
to my friends, about what I doand I talk about sales data,
just the word sales can be veryconfusing to many people.
I was blown away when I foundout that most big organizations
(03:00):
operate their business off ofshipments what they sold in the
retail.
Every decision is made when thetruck leaves the facility
without knowing what thecustomers did on the store level
.
And I've been in so manycustomer meetings when I come up
and I say hey, Alloy does salesdata, and someone on the call
says like oh, we already havethat.
And then I start asking likeokay, do you actually know what
(03:20):
you sold at this Walmart storein California?
Wait, sales means shipments.
And I was like no, no, no,Sales should really mean like
what happens at the registeredstore, and it's one of those
light bulb moments I often seecustomers' faces light up when
they say like wait, you can dothat.
It's not possible, it's not inmy ERP system.
And just the word sales can beconfusing between selling
(03:43):
shipments or point of saleregistered movements, and people
should really be running theirbusiness off of registered
movements.
Speaker 2 (03:50):
So, Manny, why do you say that people should run
their business off registeredmovements?
Because they are paid based on,obviously, what's ordered to
them by a retailer ordistributor, assuming we're
talking wholesale here.
So why is that point of sale soimportant?
Speaker 4 (04:03):
Well, customers, like the end consumer, is the
foundation that leads to anyretailer ordering products right
.
If you don't stock the rightshelves, if you don't have the
right product mixes at thestores, if you don't understand
the tendencies, you might beproducing the wrong thing, you
might be forecasting the wrongthing on your end and it's just
going to deviate yourself so faraway from what the consumer is
doing.
And it's such a competitivelandscape out there between
(04:26):
brands trying to fight for shelfspace, trying to keep all
consumers.
We have all these newgenerations coming up and you
have to really understand whatpeople are buying.
And if you're just stuck onyour old behavior what you were
shipping in last year and whatyou're doing you don't
understand what the customer isdoing.
It'll just ripple all the waythroughout your supply chain and
produce massive problems inwhat you're producing and
(04:47):
selling and shipping.
Speaker 2 (04:49):
My experience is maybe a little bit different.
I feel as though what I've seenwith companies I think it's
evolved over the last couple ofyears Most companies recognize
understanding what's happeningwith the end consumer of their
product is important.
They run their business offwhat they sell.
I mean that's going to drivetheir own personal P&L.
(05:09):
But I think the reason, as yousaid, is you need to understand
that kind of leading indicator.
The retailer is not going toorder more product from you if
they can't sell it.
So you better understand thatconsumer and look, there's a lot
of effort put intounderstanding that.
What I feel like I often see isa recognition that it's like yes
, there is some data out there,whether you're getting it
directly from the retailer,whether you're purchasing it,
(05:30):
syndicated data from like aNielsen or Cercana to kind of
understand what consumers arebuying and what those patterns
look like.
But it's really more of aresearch motion.
I guess I would say People arethinking of it more to
understand broader trends, to beable to talk internally about
why is SKUA doing, or is SKUAdoing well or not and should we
(05:52):
expect more, rather than thiskind of how do we operate our
business more at the modernspeed of business, because that,
to me, is what's changing isthe speed at which we are
turning over SKUs andintroducing new products and
sunsetting products etc.
Has just massively accelerated.
It's your comment, manny.
(06:12):
There's just so muchcompetition out there.
We're in a global market andit's kind of that recognition
that for a long time I use thisto kind of run my weekly sales
reports and get a high levelpulse of how my business is
performing and what businessesare going to take that next
level to start saying I am goingto use this on a daily basis
(06:36):
and it's going to drive how Iexecute and I'm going to be
getting ahead of consumer demand, or at least as responsive as I
possibly can, versus that kindof delayed, you know lagged
feedback loop that exists rightnow, where it's like you know,
the order comes in, I fill it,watch it from afar, and then the
next order comes in and ohshoot, it's lower than I
expected.
I better start adjusting.
Speaker 4 (06:58):
Yeah, I hear what you're saying and to your point,
I think using that point ofsale to drive product mix
decisions or productiondecisions is kind of that next
step.
Right, it really is, based onwhat I've seen with customers,
what people should be doing.
But I can actually tell you Ialmost have these cautionary
tales of companies that arefully operating on sell-in and
(07:20):
the risks that can happen.
Like you said, most companiescare about shipments because
that is how they make money.
That is where the transactionhappens and we had this one
customer, really exciting brand.
I personally really liked theproducts and I found that they
used to pay commissions to theirsales teams off of shipments.
They only care about what madeit to the store.
(07:40):
They were not looking under thehood that the return rate was
like 85%, so people were buyingstuff on the shelf but returning
it.
They were looking at grosssales, not net sales, but they
were actually only looking atshipments and what we started
seeing in the data was theseretailers were stuck.
They were not keeping track ofreturns.
Three months later they wentbankrupt.
(08:02):
There was just a completesqueeze on their finances
because they were paying peopleoff of sell-in.
They had no clue what washappening at the store and I
think it opened my eyes.
If I ever ran a consumerproducts company, the first
thing I'd do is make sure that Iknow what's happening at the
stores and have tight control onthis, because it just ripples
(08:23):
everywhere.
Speaker 2 (08:24):
You're certainly reminding me, we have had a
couple, I think, customers orsales prospects and we can get
this data live very quickly.
And you look at it and you'relike oof, that does not look
like a healthy balance betweenwhat you're shipping in and what
you're selling out and you cananticipate a lot of problems.
And I know in some of thosecases those are not companies
(08:46):
that made it long-term, soalways a little painful.
Not the news you want to showup or you want to share when you
first sign somebody up oryou're trying to close a new
deal.
Speaker 4 (08:54):
Yeah, and we'll probably get to that right In a
little later in the podcast.
But what you just mentionedcomparing shipments and point of
sale can sound trivial, butmost people have that data.
If they have point of sale, itis in a completely different
system.
Right Point of sale, to beclear, does not live in SAP.
It does not live in your ERPsystems.
Erp covers how much inventoryyou have and what you are
(09:17):
shipping.
The orders that you have goingto retail.
Point of sale is reported byretailers in different portals
and different Excel files and ifyou don't have the right tools
to put them in the same placeand to leverage that chip to
consumption, that you justmentioned is a super important
dashboard that is actuallynon-tributed build just because
systems do not talk to eachother and this data is usually
(09:38):
in a completely different datasilo well, manfred would love to
kind of unpack this concept alittle bit more because again I
know, yeah, you in particularhave a ton of experience in this
space and I think, as we kindof work with our customers and
talking about the role of pointof sale data, you have lots of
different philosophies andperspectives.
Speaker 1 (10:00):
One of the most common mistakes we've
highlighted is organizations nottalking enough about what they
want to do with point of saleand therefore maybe making some
poor design decisions about howto integrate and kind of take
advantage of that data.
So I'd be curious in you kindof helping unpack this concept
and how organizations may orshould be thinking about that
(10:20):
role of point of sale data andsort of the architecture of how
it flows into other systems andhow to take best advantage of it
.
Speaker 4 (10:25):
One of the hardest parts in starting to use the
point of sale data and sort ofthe architecture of how it flows
into other systems and how totake best advantage of it.
One of the hardest parts instarting to use the point of
sale data is just.
I like to talk and refer tothis as, like different data
languages, every data systemspeaks their own language, right
, they're all databases andthey're all keyed with a
different identity.
So when you talk to Target andyou talk to Walmart and you talk
to Amazon, they have differentkeys for what they call your
(10:46):
products.
The few companies that I'veseen are dipping their toe in a
point of sale might haveindividual teams that have their
own data sets and they havetheir own spreadsheets.
It's all run in Excel becausemost of these companies share
data with Excel and you startrunning into problems.
If you're a VP of sales whowants to keep a pulse of the
system across a drug channel ora whole retail channel, you
(11:08):
can't consolidate it right.
You might think you're readybecause you have to.
Oh yeah, they have the Excelreports.
They're tracking stuff everyweek, but then you might be
asked for a report of OK, it'sApril 4th, we just closed March.
What was the total point ofsale in the drug channel, you
might need to wait two weeks foran analyst to put that together
(11:29):
because they have to translatebetween all these different
languages to give you thatreport and if you're operating
your business two or three weeksafter the fact, right by the
time you're done now, aprilclosed and you're on to May and
like it just slows you down.
So I think having a system oneof the proudest things I think
we built at Alloy is thattranslation layer where we can
speak retail languages and webasically translate that into
(11:52):
your own language in one placethat allow you to have that
insight.
April 1st you know whathappened in March and you can
have a completely aggregatedview of the point of sale across
all your product mixes, allyour retailers.
I find that foundational andI'm blown away by how many
companies do not have that today.
Speaker 2 (12:12):
Manny on the topic of frequency.
So there you're, using anexample of a month close how to
do on the month.
What do you see in the marketaround, how frequently people
are looking at this data, whatlevel of granularity, what do
you see and what should peoplebe doing or what's possible?
Speaker 4 (12:30):
Yeah, it's a good question.
Granularity and frequency aretwo very important things that
I'll probably talk aboutseparately, because when you're
running in an Excel sheet andyou're running with humans, the
amount of detail you can go intois limited.
Excel can only handle a millionrows and then you're out right,
and it just becomes completelyslow.
So what I found is, when youtalk about data granularity the
(12:52):
level that detail at which theyrun their business people might
just look at things by skew.
They don't go down to the storeand there's 50 states in the
country.
Walmart has 8,000 locations youmight need to know and you're
just limited by tools if you'rerunning on reports.
So location granularity isimportant.
You talked about timegranularity.
(13:15):
Most people that tell me theyuse point of sale data.
I find that they run monthlyreports because, again, they
rely on humans to consolidatestuff and you don't have enough
time to be able to operate this.
If you're doing it every weeknow, you need someone to do
reports four times a month andlet alone daily level.
Right, Most of these bigretailers give you data for
(13:36):
yesterday by 8 am Eastern.
You know what you sold, youknow inventory levels across the
board, but I find that manypeople don't have the resources
that they're relying on humansto utilize that data.
So it's so important if you'retrying to run the business to be
able to run more frequently.
You can do that with softwareright.
Computers run faster thanhumans.
(13:56):
It is a repeatable process andwhen you can unlock, going
deeper than a SKU level, you cananalyze stuff by state, by
warehouse.
It's so much more powerful andthe insights you can get at
those levels.
Speaker 1 (14:11):
Well, manny, I know you've sort of addressed this
already, but I am curious.
You're talking about a lot ofdata and a lot of depth, okay,
every day, every store.
Could that be perceived asoverkill?
Why should organizations careabout doing this at such a high
degree of granularity and whyare these shortcuts potentially
(14:31):
not the best approach whendealing with point of sale data
it?
Speaker 4 (14:35):
can certainly be overkill if you're trying to
analyze this as a human rightAgain, millions of data points.
Imagine you have 100 SKUs thatyou sell at Walmart, 8,000
locations, and you're trying toanalyze every day.
You're talking massive amountsof data for a semi-small company
.
You're not going to be havingpeople read millions of lines of
(14:56):
rows to find an insight.
Read millions of lines of rowsto find an insight.
So the power comes, like.
I see this detailed data as thefoundation to then build
insights, right.
So once you have all thisinformation and you're out of an
Excel sheet, you have adatabase at your disposal.
What we can do is start usingand empowering metrics at the
store level, at the day level,to tell you exactly where you're
(15:19):
hurting to surface a problem,right.
Tell you exactly where you'rehurting to surface a problem
right, an exception report thatit's actually this particular
item, regional in California,that's struggling.
One of your biggest markets islow and it's an insight that if
you think you're doing point ofsale data at the surface you
wouldn't have because you're notdown to that level, right.
So we don't ask people to gothrough and scroll through all
(15:41):
8,000 stores, but we have thatin our back pocket to service
that insight, to tell you it'sskew A in this region.
You're losing $100,000 a day.
Speaker 2 (15:50):
Take action now, you know, something I've seen is
retailers have scorecards fortheir suppliers so that they can
go in and say, okay, this ishow much I'm selling, this is
how much it's comping year overyear, how fast the inventory is
turning over out of stock rates,et cetera, and how easy it can
be for those high levelindicators to mask issues.
(16:12):
So someone may go in and say,well, I've got 98% in stock
target, I'm doing just fine.
And yet when you roll that back, it doesn't mean that you might
not have chronic issues incertain regions or locations
that you're just losing money on, and you might look at that top
level number and be like thatlooks fine and not recognize
(16:34):
that there actually are veryaddressable issues.
So that's one of those thingswhere people, I think, are very
used to those numbers.
They that's one of those thingswhere people, I think, are very
used to those numbers.
They have line reviews, theylook at them, but oftentimes
they're actually maskingunderlying things that are just
kind of glossed over becausepeople assume that the issue is
fine.
Speaker 1 (16:52):
Great point.
You remind me of one of ourcustomers, and they're actually
quite a small company.
I don't know, manny, if you werethinking along the same vein,
where you kind of look at theprofile of this company and you
can say, on paper, this isn't acompany ready to sift through
daily store SKU level databecause they were very small.
But they had an emergingrelationship with Best Buy and
(17:14):
Best Buy had set in-stocktargets, like you're describing,
joel, and very quickly in Alloy, this company was able to see
that although they were hittingtheir on-shelf availability
targets, there were only 10stores that represented an
insane amount of their volumesbecause they sold a product that
would sell in sort of highadventure areas.
And they were able to come toBest Buy and say, yeah, we're
(17:36):
hitting our in-stock rates, butour Honolulu and our San Diego
and our Miami stores go out ofstock the day after they
replenish.
Let's actually adjust thereorder point for just these 10
stores so that we can stay instock and have the velocity that
we need.
And this is a small brand thatwas really trying to get
traction at a particularretailer and you can appreciate.
(17:58):
Of course they're not going tosift through every store every
day.
But without that granularityyou're losing opportunity right,
and there's a lot that isunlocked again to your point,
mandy, like a databasetechnology perspective that
you're leaving on the table.
Speaker 4 (18:13):
I remember that client exactly.
It was one of my first meetingswhen, I think you and I went on
site and it stuck in my brain,right, Because, going back to
Joel's question aboutgranularity, you're right that
we were able to detect that theproblem was in Honolulu, right,
so that's already one levelabove when we can detect the
region underneath this.
You were already helping, butwe went one level deeper, right.
(18:35):
Going back to this weeklyversus daily data set, Joel's
right, they were runningbusiness reviews with Best Buy.
They're running them on Fridaysand every week everything in
stocks looked good across theboard, because weekly in stocks
were fine.
But we found out that theyreplenished on Thursday and they
were actually stocking outMonday, Tuesday, Wednesday.
(18:56):
The shelves were out of stockand you don't see that if you're
only running the business byyour week, right.
But we could tell them.
In fact, last week you were not100% in stock at the end of the
week, you were 50% in stockthroughout the week and we think
you should be selling twicethis.
So we gave them thatrecommendation.
So it's very early days.
They actually told Best Buy hey, I think you can order twice as
(19:18):
much.
Early days they actually toldBest Buy.
Hey, I think you can ordertwice as much.
Best Buy was like yeah, OK,let's see Right, it's mutually
beneficial for them to alsoorder more if they're going to
sell more.
And it stuck.
Honolulu started selling twicethe volume and then they applied
that across the nation becausethey realized they were stocking
out.
Speaker 2 (19:34):
Here's an example of a smaller.
It sounds like brand relativelynew to Best Buy.
Here's an example of a smaller.
It sounds like brand relativelynew to Best Buy.
Do we see a difference betweenyour large global CPGs, for
example, that have thousands ofpeople that are much more
sophisticated, and then you gotthese small brands that are
maybe going into retail for thefirst time?
So do all people benefit fromthis data?
(19:56):
Similarly, do we find that theydo different things with it at
different sizes?
Curious, manny, what you'veseen on this front.
Speaker 4 (20:03):
So I think you're spot on that.
Bigger companies, right, carrymore weight behind them and they
have big retailers almost onephone call away.
They can provide a call.
They have more influence overbuying decisions, right, just
because of the nature, theyprobably have a longer history,
(20:24):
they probably have a personalrelationship there.
But I found that even smallcompanies, when you bring the
right data insight right, no onecan dispute a data-backed
recommendation.
It could take a few emails toget there.
This particular example is agood one because it took a few
emails.
We had a very specific insightat one location, one data point,
and it landed right.
So it gave this small brandkind of a reputation of someone
(20:44):
who is data-driven.
The Best Buy team doesn't haveenough time to track every
single SKU that they have intheir portfolio, but there are,
again, mutually aligned andmutually beneficial to want to
maintain shelves that arestocked.
So, yes, there's more sway ifyou're a bigger company, but
that's the power of data.
Speaker 1 (21:05):
No one can dispute an insight.
If you can convince the rightpeople that it's in their
interest, You're spot on from anexecution perspective in that
distinction.
I think in my experience aswell.
Here, Joel, you find theselarger organizations having
other systems, other processes,investing a lot in AI
initiatives, investing a lot inplanning applications and supply
chain optimization, and I thinkoften in the enterprise we find
(21:26):
that that's another area, thatpoint of sale data can help
deliver a ton of value.
Right, I think AI is probablythe most interesting concept
right now.
A common question is when youhave massive data sets, how do
you make sense of them?
Ai can play a key role there,and so, as we sort of advise
enterprises, one of the piecesof advice that we have is hey,
(21:48):
maybe 10 years ago thisgranularity we're talking about
store level, day, SKU levelwasn't as important.
But now, as AI sort of advancesand you can get value out of
that data at scale even thesereally big enterprises that's
something we really push on.
Speaker 2 (22:04):
So getting that data ready, sanitized, harmonized, at
that level of granularity, Ithink in the enterprise gets
pretty exciting at the size ofdata sets and if you're looking
as we talked about at thebeginning, you know kind of the
traditional way that a wholesaleat least the wholesale side of
(22:24):
the business is going to work isthey're going to be like well,
I've got my orders that I maketo my factory, I've got the
inventory across, however manywarehouses that I'm putting my
product at, and then I'm fillingorders as they come in and
those are kind of the datapoints that you have.
So maybe you have fivedifferent warehouses.
You're fulfilling these ordersthat come in every week, or what
(22:44):
have you from various retailersand distributors.
You go to the point of sale andall of a sudden, the volume of
data goes up exponentially.
You're now not looking at fivelocations and a thousand SKUs,
you're looking at a hundredthousand locations.
It can be overwhelming.
I think we've certainly seenthat.
And how do you distill thatdown?
(23:05):
How do you use, you know, alertpeople to the things that are
going to be the most interestingso they don't have to sift
through it, which nobody wouldbe able to do.
But as AI advances, you knowits ability to kind of augment
that and say, hey, the more datathat it has, the more powerful
it's going to be in terms ofgetting those insights.
So yeah, size of data sets Iknow we alluded to that earlier,
(23:26):
but it definitely goes beyondanything that an individual, or
humans in general, can kind ofmanage.
You've got to use, you know,tools to take advantage of it,
and so then it's probably just aquestion of like, how much
incremental value is there inall these nuances which I think
the stories we're telling is?
There's a lot, as long as youcan find it Well.
Speaker 1 (23:47):
Manny, shifting up to kind of other functions and
organizations and kind of wherePOS data plays.
I'm curious from an executionperspective.
You know we often don't thinkabout that as a data problem,
but could you speak a little bitto where you should be using
daily highly granular point ofsale data across your business
(24:08):
right when you see kind of bestin class taking advantage?
We've talked aboutcollaborative replenishment and
kind of collaborating with thoseretailers providing
perspectives, but what otherways do you recommend or see
being an effective place todeploy these kind of large,
detailed data sets?
Speaker 4 (24:24):
That's a great question because we've talked a
lot about data as thatfoundational layer for you to
run your business right, anddata is only as good as the
people who use it.
You can have the best analytics, you can have the best
visualization tools and allthese insights, but if you don't
have people that can actuallytake action, then they're
worthless to you as a businessright.
(24:44):
And one of the most importantthings you know I was talking a
little earlier about data silosyou have a sales team in an
organization that usually hastheir POS reports in their own
thing and they're completelydisconnected from your supply
chain team, and supply chainoperates in a different way.
They meet once a month.
If you have an SNOP processwhere they kind of talk about
production forecast, things likethat, there's no communication
(25:06):
throughout the month.
And we were just talking aboutthat story.
Right, Honolulu, we recommendedthat they should be ordering
twice as much into a particularSKU, into that volume.
Where the execution right.
I call that collaborativereplenishment.
We brought them a collaborativereplenishment insight that the
sales team at this company, thatthey can take to the retailer
to take action.
Imagine if they don't havetheir own inventory ready to
(25:30):
fulfill twice the amount oforders.
Imagine if that sales team hadnot collaborated with their
supply chain team before they,you know, recommended this To be
sure, right.
What we found is that at Alloy,when you can put people on the
same platform to collaborateright, when you can connect that
point of sale insight with yourown inventory, you can make
sure that you're breaking onthat data silo.
(25:51):
You could have sent them anemail to ask, hey, do I have
enough inventory?
And wait a few days to hearback.
If you have the right insightin the right place, you know
immediately.
So I think collaborationreplenishment is a great example
of when you execute to sellmore, but it's super important
that you coordinate internallybetween all your teams to make
sure that you are able todeliver on what you're promising
(26:14):
.
Speaker 1 (26:15):
That, I think, is a fascinating example to kind of
round out that part of the story.
So we talked about evenimproving that collaborative
replenishment, avoiding mistakes, uncomfortable situations.
But I also am curious because Iknow again, you get to touch
lots of different functions andteams.
Other examples of maybe howthis can be deployed and kind of
helping deliver value for theorganization Absolutely.
Speaker 4 (26:38):
So this example we talked about, right, was a sales
team going to their supplychain team to ask for help.
But in my experience it couldstart the opposite way.
It is very rare.
It is very rare.
Even if an organization isusing point of sale data, it is
extremely rare for the supplychain team to be acting or to
have visibility to that data.
And supply chain team operatesin truckloads right, they're
(27:00):
producing, they're moving stuffbetween warehouses.
They're never understandingwhat a customer is doing at the
store.
And I remember we had onecustomer that was pretty
visionary and he said I want mysupply chain team to use point
of sale data because they thinkit's important and I think that
it's a relevant piece for them.
And I remember going to thesecalls and like it was a little
bit apprehensive.
That team, the supply chainteam, was apprehensive.
(27:22):
They had enough data pointsalready.
Why do they need more datapoints?
Then we started talking aboutsomething every supply chain
team dreads, which is the A wordallocation.
Right, you produced, you didn'tproduce enough and you have a
thousand units.
You've got orders that total5,000 units for that item and
(27:43):
you have to allocate it.
You have to pick who you sendit to.
That's a supply chain problem.
Right, like you're running low,you're at risk.
You have to make a decisionthat's gonna impact the sales
team.
Many companies today willprobably just do FIFO right.
First in, first out.
Who placed the first order.
It goes that way.
But we started providinginsight.
That said, hold up.
(28:03):
We actually know whichretailers need this product
because we know what theconsumer is doing.
And if I could tell you thatWalmart has six weeks of
inventory, target has two,amazon is at one, you can now
start prioritizing yourallocation decisions in a way
that could try to protect yourrelationship with the retailer
and, even more important,prevent lost sales at the store.
(28:26):
Right, because these 1,000units you have are pretty
precious.
And if you're just doing firstin, first out, going back to
data-driven decisions, you don'thave any data.
But when you connect the pointof sale all the way up there,
you can bring that insight tothe supply chain team.
Now everyone can live moreharmoniously and you're making a
better execution decisionacross the board pretty
(28:46):
seamlessly if you could do it inone system.
Speaker 1 (28:48):
I was going to ask you a question, joel.
It's a topic that's, I think,passionate on your end is this
kind of planning process andplanning and forecasting process
and how that intersects withpoint of sale data.
I know you're workingpartnerships for us and have a
lot of experience in that space.
Could you speak a little bit tokind of what you see as best in
class and taking advantage ofpoint of sale data in that
(29:09):
planning domain, of what you seeas best in class and taking
advantage of point of sale datain that planning domain?
Speaker 2 (29:11):
Yeah, I think we're kind of in the midst of an
evolution, as we've talked abouta couple times earlier in this
podcast.
Historically, brands have usedtheir shipment data to really
forecast out what sales aregoing to be.
I mean, that's been kind of theprimary historical data point.
(29:32):
Then they're going to layer inall sorts of other things on top
of that.
Models have gotten a lot moresophisticated.
A lot of people are usingmachine learning, but if you
really break it down, it's likethe primary driver of predicting
future sales performance isgoing to be prior shipments.
I think for a long time there'sbeen a feeling well, point of
(29:52):
sale should be a betterindicator and it shouldn't just
be point of sale itself.
You also need to understandwhat inventory levels are like
at a retailer.
If you're selling a lot ofproduct but there's tons of
inventory that's piled up,that's very different than
you're selling a lot of productand the retailer is about to run
out.
But, as we talked about earlier, it's just been so hard to
bring that in.
So I think what we've seen isan interest.
I think over the last couple ofyears, more and more brands are
(30:16):
trying to figure out how theybring that into their models.
I think there's interestingkind of statistical questions
about how they actuallyincorporate that data.
But I know, logan, you and I wewere talking to a very large
company.
We're talking to them right nowabout this idea.
I mean, this is a brandeverybody would know and they
are very intent on saying wereally think that using point of
(30:39):
sale is going to be a betterpredictor than looking at
shipments and that's somethingthat they want to shift to.
And so I think everybody assumesthat it's going to switch over
at some point.
It's a question of when and howquickly that is.
I think everybody assumes thatit's going to switch over at
some point.
It's a question of when and howquickly that is.
And this gets to again.
The core problem we've beentalking about is point of sale
has been out there for a while.
It's been getting better.
People are recognizing withtechnology you can really
(31:01):
incorporate this.
It doesn't just need to be kindof a sideshow, but these shifts
take time.
Speaker 1 (31:06):
Well, really fascinating stuff.
Joel Manny, thank you both foryour insights, perspectives on
data.
I know this group's quitepassionate about it.
So great conversation andthanks for joining us, manny
Absolutely.
Speaker 4 (31:19):
It's one of my favorite things to talk about
right Data.
People think they know whatthey're doing and they're not
right when they talk about salesor talk about shipments or unit
sales.
Big decision.
Even if they are talking aboutunit sales at the store, is it
weekly, is it daily, like?
Are they actually knowing this?
Are the right teams gettingthis insight or is it kind of
siloed in their own world?
There's so much under the hoodthat is important.
(31:40):
I'd love to talk about this and, more than talking about it, I
love all the value we can bringto customers when we can get
them on our best practice.
Speaker 1 (31:52):
You've been listening to Manfred Reichi, subject
matter expert at Alloy AI.
That's all for this week.
See you next time on Shelf Life.
From Alloy AI.
This is Shelf Life.
Speaker 2 (00:19):
What are some of the common mistakes brands make when
it comes to data?
Speaker 1 (00:25):
the common mistakes brands make when it comes to
data.
How can you be using point ofsale data to power other parts
of your business, like planning,promotions and supply chain
decisions?
Speaker 2 (00:31):
How can you use data to better collaborate with
retailers?
Speaker 1 (00:35):
On every episode of Shelf Life.
We answer questions like theseand more, with the help of
leaders across the consumergoods industry.
Speaker 2 (00:43):
Today we welcome our colleague, Manfred Reichi,
Alloy's subject matter expert onall things related to retail
data and products.
He spent his time at Alloybeing hands-on with some of our
most sophisticated brands in theconsumer goods industry.
Prior to joining us at Alloy,Manfred started his career as a
technical consultant at Deloitte, focused on SAP implementations
.
I'm your co-host, Joel Beal,CEO of Alloy AI.
Speaker 1 (01:07):
And I'm your co-host, Logan Ensign, Chief Customer
Officer at Alloy AI.
Speaker 2 (01:12):
We'll be back with Manfred right after this.
Speaker 3 (01:16):
As a consumer brand, you connect with dozens of
external partners and internalsystems to get a complete
picture of your business.
Each one is different.
Alloy AI makes it easy toconnect data from retailers,
e-commerce, supply chainpartners and even your own ERP,
then easily surface insights todrive sales growth.
Every day, brands use Alloy AIto see POS trends, measure
(01:37):
promotion performance and makebetter replenishment decisions
with their retail partners.
That's why we're trusted by BIC, crayola, valvoline, melissa,
doug Bosch and many more.
Get a demo at Alloyai today.
Speaker 1 (01:51):
Manfred, welcome to Shelf Life, really good to have
you, great to be here.
I'm surprised we've done asmany episodes as we have and
haven't had you yet.
Really excited to dive into oneof my favorite topics.
A lot to talk about data,that's right.
That's right.
Well, let's dive in.
I think between the three of uswe have spoken to a ton of
companies enterprise, mid-market, smb that many think have this
(02:14):
idea of point of sale datasolved already, and then when we
peel back the curtain a littlebit, we realize that maybe
that's not the case.
And so in our experience I knowtypically we do find that a lot
of companies feel like they'repretty good at data, but
actually a lot of them aren't.
So, manny, I am interested inhearing just some specific
(02:34):
example stories of kind of thisexperience that often we have.
Speaker 4 (02:38):
One of my favorite initial stories about data, and
we have this concept in Allo.
We call language policingourselves.
Right, Like words meandifferent things and when I talk
to people, either internallywith customers here at Alloy or
to my friends, about what I doand I talk about sales data,
just the word sales can be veryconfusing to many people.
I was blown away when I foundout that most big organizations
(03:00):
operate their business off ofshipments what they sold in the
retail.
Every decision is made when thetruck leaves the facility
without knowing what thecustomers did on the store level
.
And I've been in so manycustomer meetings when I come up
and I say hey, Alloy does salesdata, and someone on the call
says like oh, we already havethat.
And then I start asking likeokay, do you actually know what
(03:20):
you sold at this Walmart storein California?
Wait, sales means shipments.
And I was like no, no, no,Sales should really mean like
what happens at the registeredstore, and it's one of those
light bulb moments I often seecustomers' faces light up when
they say like wait, you can dothat.
It's not possible, it's not inmy ERP system.
And just the word sales can beconfusing between selling
(03:43):
shipments or point of saleregistered movements, and people
should really be running theirbusiness off of registered
movements.
Speaker 2 (03:50):
So, Manny, why do you say that people should run
their business off registeredmovements?
Because they are paid based on,obviously, what's ordered to
them by a retailer ordistributor, assuming we're
talking wholesale here.
So why is that point of sale soimportant?
Speaker 4 (04:03):
Well, customers, like the end consumer, is the
foundation that leads to anyretailer ordering products right
.
If you don't stock the rightshelves, if you don't have the
right product mixes at thestores, if you don't understand
the tendencies, you might beproducing the wrong thing, you
might be forecasting the wrongthing on your end and it's just
going to deviate yourself so faraway from what the consumer is
doing.
And it's such a competitivelandscape out there between
(04:26):
brands trying to fight for shelfspace, trying to keep all
consumers.
We have all these newgenerations coming up and you
have to really understand whatpeople are buying.
And if you're just stuck onyour old behavior what you were
shipping in last year and whatyou're doing you don't
understand what the customer isdoing.
It'll just ripple all the waythroughout your supply chain and
produce massive problems inwhat you're producing and
(04:47):
selling and shipping.
Speaker 2 (04:49):
My experience is maybe a little bit different.
I feel as though what I've seenwith companies I think it's
evolved over the last couple ofyears Most companies recognize
understanding what's happeningwith the end consumer of their
product is important.
They run their business offwhat they sell.
I mean that's going to drivetheir own personal P&L.
(05:09):
But I think the reason, as yousaid, is you need to understand
that kind of leading indicator.
The retailer is not going toorder more product from you if
they can't sell it.
So you better understand thatconsumer and look, there's a lot
of effort put intounderstanding that.
What I feel like I often see isa recognition that it's like yes
, there is some data out there,whether you're getting it
directly from the retailer,whether you're purchasing it,
(05:30):
syndicated data from like aNielsen or Cercana to kind of
understand what consumers arebuying and what those patterns
look like.
But it's really more of aresearch motion.
I guess I would say People arethinking of it more to
understand broader trends, to beable to talk internally about
why is SKUA doing, or is SKUAdoing well or not and should we
(05:52):
expect more, rather than thiskind of how do we operate our
business more at the modernspeed of business, because that,
to me, is what's changing isthe speed at which we are
turning over SKUs andintroducing new products and
sunsetting products etc.
Has just massively accelerated.
It's your comment, manny.
(06:12):
There's just so muchcompetition out there.
We're in a global market andit's kind of that recognition
that for a long time I use thisto kind of run my weekly sales
reports and get a high levelpulse of how my business is
performing and what businessesare going to take that next
level to start saying I am goingto use this on a daily basis
(06:36):
and it's going to drive how Iexecute and I'm going to be
getting ahead of consumer demand, or at least as responsive as I
possibly can, versus that kindof delayed, you know lagged
feedback loop that exists rightnow, where it's like you know,
the order comes in, I fill it,watch it from afar, and then the
next order comes in and ohshoot, it's lower than I
expected.
I better start adjusting.
Speaker 4 (06:58):
Yeah, I hear what you're saying and to your point,
I think using that point ofsale to drive product mix
decisions or productiondecisions is kind of that next
step.
Right, it really is, based onwhat I've seen with customers,
what people should be doing.
But I can actually tell you Ialmost have these cautionary
tales of companies that arefully operating on sell-in and
(07:20):
the risks that can happen.
Like you said, most companiescare about shipments because
that is how they make money.
That is where the transactionhappens and we had this one
customer, really exciting brand.
I personally really liked theproducts and I found that they
used to pay commissions to theirsales teams off of shipments.
They only care about what madeit to the store.
(07:40):
They were not looking under thehood that the return rate was
like 85%, so people were buyingstuff on the shelf but returning
it.
They were looking at grosssales, not net sales, but they
were actually only looking atshipments and what we started
seeing in the data was theseretailers were stuck.
They were not keeping track ofreturns.
Three months later they wentbankrupt.
(08:02):
There was just a completesqueeze on their finances
because they were paying peopleoff of sell-in.
They had no clue what washappening at the store and I
think it opened my eyes.
If I ever ran a consumerproducts company, the first
thing I'd do is make sure that Iknow what's happening at the
stores and have tight control onthis, because it just ripples
(08:23):
everywhere.
Speaker 2 (08:24):
You're certainly reminding me, we have had a
couple, I think, customers orsales prospects and we can get
this data live very quickly.
And you look at it and you'relike oof, that does not look
like a healthy balance betweenwhat you're shipping in and what
you're selling out and you cananticipate a lot of problems.
And I know in some of thosecases those are not companies
(08:46):
that made it long-term, soalways a little painful.
Not the news you want to showup or you want to share when you
first sign somebody up oryou're trying to close a new
deal.
Speaker 4 (08:54):
Yeah, and we'll probably get to that right In a
little later in the podcast.
But what you just mentionedcomparing shipments and point of
sale can sound trivial, butmost people have that data.
If they have point of sale, itis in a completely different
system.
Right Point of sale, to beclear, does not live in SAP.
It does not live in your ERPsystems.
Erp covers how much inventoryyou have and what you are
(09:17):
shipping.
The orders that you have goingto retail.
Point of sale is reported byretailers in different portals
and different Excel files and ifyou don't have the right tools
to put them in the same placeand to leverage that chip to
consumption, that you justmentioned is a super important
dashboard that is actuallynon-tributed build just because
systems do not talk to eachother and this data is usually
(09:38):
in a completely different datasilo well, manfred would love to
kind of unpack this concept alittle bit more because again I
know, yeah, you in particularhave a ton of experience in this
space and I think, as we kindof work with our customers and
talking about the role of pointof sale data, you have lots of
different philosophies andperspectives.
Speaker 1 (10:00):
One of the most common mistakes we've
highlighted is organizations nottalking enough about what they
want to do with point of saleand therefore maybe making some
poor design decisions about howto integrate and kind of take
advantage of that data.
So I'd be curious in you kindof helping unpack this concept
and how organizations may orshould be thinking about that
(10:20):
role of point of sale data andsort of the architecture of how
it flows into other systems andhow to take best advantage of it
.
Speaker 4 (10:25):
One of the hardest parts in starting to use the
point of sale data and sort ofthe architecture of how it flows
into other systems and how totake best advantage of it.
One of the hardest parts instarting to use the point of
sale data is just.
I like to talk and refer tothis as, like different data
languages, every data systemspeaks their own language, right
, they're all databases andthey're all keyed with a
different identity.
So when you talk to Target andyou talk to Walmart and you talk
to Amazon, they have differentkeys for what they call your
(10:46):
products.
The few companies that I'veseen are dipping their toe in a
point of sale might haveindividual teams that have their
own data sets and they havetheir own spreadsheets.
It's all run in Excel becausemost of these companies share
data with Excel and you startrunning into problems.
If you're a VP of sales whowants to keep a pulse of the
system across a drug channel ora whole retail channel, you
(11:08):
can't consolidate it right.
You might think you're readybecause you have to.
Oh yeah, they have the Excelreports.
They're tracking stuff everyweek, but then you might be
asked for a report of OK, it'sApril 4th, we just closed March.
What was the total point ofsale in the drug channel, you
might need to wait two weeks foran analyst to put that together
(11:29):
because they have to translatebetween all these different
languages to give you thatreport and if you're operating
your business two or three weeksafter the fact, right by the
time you're done now, aprilclosed and you're on to May and
like it just slows you down.
So I think having a system oneof the proudest things I think
we built at Alloy is thattranslation layer where we can
speak retail languages and webasically translate that into
(11:52):
your own language in one placethat allow you to have that
insight.
April 1st you know whathappened in March and you can
have a completely aggregatedview of the point of sale across
all your product mixes, allyour retailers.
I find that foundational andI'm blown away by how many
companies do not have that today.
Speaker 2 (12:12):
Manny on the topic of frequency.
So there you're, using anexample of a month close how to
do on the month.
What do you see in the marketaround, how frequently people
are looking at this data, whatlevel of granularity, what do
you see and what should peoplebe doing or what's possible?
Speaker 4 (12:30):
Yeah, it's a good question.
Granularity and frequency aretwo very important things that
I'll probably talk aboutseparately, because when you're
running in an Excel sheet andyou're running with humans, the
amount of detail you can go intois limited.
Excel can only handle a millionrows and then you're out right,
and it just becomes completelyslow.
So what I found is, when youtalk about data granularity the
(12:52):
level that detail at which theyrun their business people might
just look at things by skew.
They don't go down to the storeand there's 50 states in the
country.
Walmart has 8,000 locations youmight need to know and you're
just limited by tools if you'rerunning on reports.
So location granularity isimportant.
You talked about timegranularity.
(13:15):
Most people that tell me theyuse point of sale data.
I find that they run monthlyreports because, again, they
rely on humans to consolidatestuff and you don't have enough
time to be able to operate this.
If you're doing it every weeknow, you need someone to do
reports four times a month andlet alone daily level.
Right, Most of these bigretailers give you data for
(13:36):
yesterday by 8 am Eastern.
You know what you sold, youknow inventory levels across the
board, but I find that manypeople don't have the resources
that they're relying on humansto utilize that data.
So it's so important if you'retrying to run the business to be
able to run more frequently.
You can do that with softwareright.
Computers run faster thanhumans.
(13:56):
It is a repeatable process andwhen you can unlock, going
deeper than a SKU level, you cananalyze stuff by state, by
warehouse.
It's so much more powerful andthe insights you can get at
those levels.
Speaker 1 (14:11):
Well, manny, I know you've sort of addressed this
already, but I am curious.
You're talking about a lot ofdata and a lot of depth, okay,
every day, every store.
Could that be perceived asoverkill?
Why should organizations careabout doing this at such a high
degree of granularity and whyare these shortcuts potentially
(14:31):
not the best approach whendealing with point of sale data
it?
Speaker 4 (14:35):
can certainly be overkill if you're trying to
analyze this as a human rightAgain, millions of data points.
Imagine you have 100 SKUs thatyou sell at Walmart, 8,000
locations, and you're trying toanalyze every day.
You're talking massive amountsof data for a semi-small company
.
You're not going to be havingpeople read millions of lines of
(14:56):
rows to find an insight.
Read millions of lines of rowsto find an insight.
So the power comes, like.
I see this detailed data as thefoundation to then build
insights, right.
So once you have all thisinformation and you're out of an
Excel sheet, you have adatabase at your disposal.
What we can do is start usingand empowering metrics at the
store level, at the day level,to tell you exactly where you're
(15:19):
hurting to surface a problem,right.
Tell you exactly where you'rehurting to surface a problem
right, an exception report thatit's actually this particular
item, regional in California,that's struggling.
One of your biggest markets islow and it's an insight that if
you think you're doing point ofsale data at the surface you
wouldn't have because you're notdown to that level, right.
So we don't ask people to gothrough and scroll through all
(15:41):
8,000 stores, but we have thatin our back pocket to service
that insight, to tell you it'sskew A in this region.
You're losing $100,000 a day.
Speaker 2 (15:50):
Take action now, you know, something I've seen is
retailers have scorecards fortheir suppliers so that they can
go in and say, okay, this ishow much I'm selling, this is
how much it's comping year overyear, how fast the inventory is
turning over out of stock rates,et cetera, and how easy it can
be for those high levelindicators to mask issues.
(16:12):
So someone may go in and say,well, I've got 98% in stock
target, I'm doing just fine.
And yet when you roll that back, it doesn't mean that you might
not have chronic issues incertain regions or locations
that you're just losing money on, and you might look at that top
level number and be like thatlooks fine and not recognize
(16:34):
that there actually are veryaddressable issues.
So that's one of those thingswhere people, I think, are very
used to those numbers.
They that's one of those thingswhere people, I think, are very
used to those numbers.
They have line reviews, theylook at them, but oftentimes
they're actually maskingunderlying things that are just
kind of glossed over becausepeople assume that the issue is
fine.
Speaker 1 (16:52):
Great point.
You remind me of one of ourcustomers, and they're actually
quite a small company.
I don't know, manny, if you werethinking along the same vein,
where you kind of look at theprofile of this company and you
can say, on paper, this isn't acompany ready to sift through
daily store SKU level databecause they were very small.
But they had an emergingrelationship with Best Buy and
(17:14):
Best Buy had set in-stocktargets, like you're describing,
joel, and very quickly in Alloy, this company was able to see
that although they were hittingtheir on-shelf availability
targets, there were only 10stores that represented an
insane amount of their volumesbecause they sold a product that
would sell in sort of highadventure areas.
And they were able to come toBest Buy and say, yeah, we're
(17:36):
hitting our in-stock rates, butour Honolulu and our San Diego
and our Miami stores go out ofstock the day after they
replenish.
Let's actually adjust thereorder point for just these 10
stores so that we can stay instock and have the velocity that
we need.
And this is a small brand thatwas really trying to get
traction at a particularretailer and you can appreciate.
(17:58):
Of course they're not going tosift through every store every
day.
But without that granularityyou're losing opportunity right,
and there's a lot that isunlocked again to your point,
mandy, like a databasetechnology perspective that
you're leaving on the table.
Speaker 4 (18:13):
I remember that client exactly.
It was one of my first meetingswhen, I think you and I went on
site and it stuck in my brain,right, Because, going back to
Joel's question aboutgranularity, you're right that
we were able to detect that theproblem was in Honolulu, right,
so that's already one levelabove when we can detect the
region underneath this.
You were already helping, butwe went one level deeper, right.
(18:35):
Going back to this weeklyversus daily data set, Joel's
right, they were runningbusiness reviews with Best Buy.
They're running them on Fridaysand every week everything in
stocks looked good across theboard, because weekly in stocks
were fine.
But we found out that theyreplenished on Thursday and they
were actually stocking outMonday, Tuesday, Wednesday.
(18:56):
The shelves were out of stockand you don't see that if you're
only running the business byyour week, right.
But we could tell them.
In fact, last week you were not100% in stock at the end of the
week, you were 50% in stockthroughout the week and we think
you should be selling twicethis.
So we gave them thatrecommendation.
So it's very early days.
They actually told Best Buy hey, I think you can order twice as
(19:18):
much.
Early days they actually toldBest Buy.
Hey, I think you can ordertwice as much.
Best Buy was like yeah, OK,let's see Right, it's mutually
beneficial for them to alsoorder more if they're going to
sell more.
And it stuck.
Honolulu started selling twicethe volume and then they applied
that across the nation becausethey realized they were stocking
out.
Speaker 2 (19:34):
Here's an example of a smaller.
It sounds like brand relativelynew to Best Buy.
Here's an example of a smaller.
It sounds like brand relativelynew to Best Buy.
Do we see a difference betweenyour large global CPGs, for
example, that have thousands ofpeople that are much more
sophisticated, and then you gotthese small brands that are
maybe going into retail for thefirst time?
So do all people benefit fromthis data?
(19:56):
Similarly, do we find that theydo different things with it at
different sizes?
Curious, manny, what you'veseen on this front.
Speaker 4 (20:03):
So I think you're spot on that.
Bigger companies, right, carrymore weight behind them and they
have big retailers almost onephone call away.
They can provide a call.
They have more influence overbuying decisions, right, just
because of the nature, theyprobably have a longer history,
(20:24):
they probably have a personalrelationship there.
But I found that even smallcompanies, when you bring the
right data insight right, no onecan dispute a data-backed
recommendation.
It could take a few emails toget there.
This particular example is agood one because it took a few
emails.
We had a very specific insightat one location, one data point,
and it landed right.
So it gave this small brandkind of a reputation of someone
(20:44):
who is data-driven.
The Best Buy team doesn't haveenough time to track every
single SKU that they have intheir portfolio, but there are,
again, mutually aligned andmutually beneficial to want to
maintain shelves that arestocked.
So, yes, there's more sway ifyou're a bigger company, but
that's the power of data.
Speaker 1 (21:05):
No one can dispute an insight.
If you can convince the rightpeople that it's in their
interest, You're spot on from anexecution perspective in that
distinction.
I think in my experience aswell.
Here, Joel, you find theselarger organizations having
other systems, other processes,investing a lot in AI
initiatives, investing a lot inplanning applications and supply
chain optimization, and I thinkoften in the enterprise we find
(21:26):
that that's another area, thatpoint of sale data can help
deliver a ton of value.
Right, I think AI is probablythe most interesting concept
right now.
A common question is when youhave massive data sets, how do
you make sense of them?
Ai can play a key role there,and so, as we sort of advise
enterprises, one of the piecesof advice that we have is hey,
(21:48):
maybe 10 years ago thisgranularity we're talking about
store level, day, SKU levelwasn't as important.
But now, as AI sort of advancesand you can get value out of
that data at scale even thesereally big enterprises that's
something we really push on.
Speaker 2 (22:04):
So getting that data ready, sanitized, harmonized, at
that level of granularity, Ithink in the enterprise gets
pretty exciting at the size ofdata sets and if you're looking
as we talked about at thebeginning, you know kind of the
traditional way that a wholesaleat least the wholesale side of
(22:24):
the business is going to work isthey're going to be like well,
I've got my orders that I maketo my factory, I've got the
inventory across, however manywarehouses that I'm putting my
product at, and then I'm fillingorders as they come in and
those are kind of the datapoints that you have.
So maybe you have fivedifferent warehouses.
You're fulfilling these ordersthat come in every week, or what
(22:44):
have you from various retailersand distributors.
You go to the point of sale andall of a sudden, the volume of
data goes up exponentially.
You're now not looking at fivelocations and a thousand SKUs,
you're looking at a hundredthousand locations.
It can be overwhelming.
I think we've certainly seenthat.
And how do you distill thatdown?
(23:05):
How do you use, you know, alertpeople to the things that are
going to be the most interestingso they don't have to sift
through it, which nobody wouldbe able to do.
But as AI advances, you knowits ability to kind of augment
that and say, hey, the more datathat it has, the more powerful
it's going to be in terms ofgetting those insights.
So yeah, size of data sets Iknow we alluded to that earlier,
(23:26):
but it definitely goes beyondanything that an individual, or
humans in general, can kind ofmanage.
You've got to use, you know,tools to take advantage of it,
and so then it's probably just aquestion of like, how much
incremental value is there inall these nuances which I think
the stories we're telling is?
There's a lot, as long as youcan find it Well.
Speaker 1 (23:47):
Manny, shifting up to kind of other functions and
organizations and kind of wherePOS data plays.
I'm curious from an executionperspective.
You know we often don't thinkabout that as a data problem,
but could you speak a little bitto where you should be using
daily highly granular point ofsale data across your business
(24:08):
right when you see kind of bestin class taking advantage?
We've talked aboutcollaborative replenishment and
kind of collaborating with thoseretailers providing
perspectives, but what otherways do you recommend or see
being an effective place todeploy these kind of large,
detailed data sets?
Speaker 4 (24:24):
That's a great question because we've talked a
lot about data as thatfoundational layer for you to
run your business right, anddata is only as good as the
people who use it.
You can have the best analytics, you can have the best
visualization tools and allthese insights, but if you don't
have people that can actuallytake action, then they're
worthless to you as a businessright.
(24:44):
And one of the most importantthings you know I was talking a
little earlier about data silosyou have a sales team in an
organization that usually hastheir POS reports in their own
thing and they're completelydisconnected from your supply
chain team, and supply chainoperates in a different way.
They meet once a month.
If you have an SNOP processwhere they kind of talk about
production forecast, things likethat, there's no communication
(25:06):
throughout the month.
And we were just talking aboutthat story.
Right, Honolulu, we recommendedthat they should be ordering
twice as much into a particularSKU, into that volume.
Where the execution right.
I call that collaborativereplenishment.
We brought them a collaborativereplenishment insight that the
sales team at this company, thatthey can take to the retailer
to take action.
Imagine if they don't havetheir own inventory ready to
(25:30):
fulfill twice the amount oforders.
Imagine if that sales team hadnot collaborated with their
supply chain team before they,you know, recommended this To be
sure, right.
What we found is that at Alloy,when you can put people on the
same platform to collaborateright, when you can connect that
point of sale insight with yourown inventory, you can make
sure that you're breaking onthat data silo.
(25:51):
You could have sent them anemail to ask, hey, do I have
enough inventory?
And wait a few days to hearback.
If you have the right insightin the right place, you know
immediately.
So I think collaborationreplenishment is a great example
of when you execute to sellmore, but it's super important
that you coordinate internallybetween all your teams to make
sure that you are able todeliver on what you're promising
(26:14):
.
Speaker 1 (26:15):
That, I think, is a fascinating example to kind of
round out that part of the story.
So we talked about evenimproving that collaborative
replenishment, avoiding mistakes, uncomfortable situations.
But I also am curious because Iknow again, you get to touch
lots of different functions andteams.
Other examples of maybe howthis can be deployed and kind of
helping deliver value for theorganization Absolutely.
Speaker 4 (26:38):
So this example we talked about, right, was a sales
team going to their supplychain team to ask for help.
But in my experience it couldstart the opposite way.
It is very rare.
It is very rare.
Even if an organization isusing point of sale data, it is
extremely rare for the supplychain team to be acting or to
have visibility to that data.
And supply chain team operatesin truckloads right, they're
(27:00):
producing, they're moving stuffbetween warehouses.
They're never understandingwhat a customer is doing at the
store.
And I remember we had onecustomer that was pretty
visionary and he said I want mysupply chain team to use point
of sale data because they thinkit's important and I think that
it's a relevant piece for them.
And I remember going to thesecalls and like it was a little
bit apprehensive.
That team, the supply chainteam, was apprehensive.
(27:22):
They had enough data pointsalready.
Why do they need more datapoints?
Then we started talking aboutsomething every supply chain
team dreads, which is the A wordallocation.
Right, you produced, you didn'tproduce enough and you have a
thousand units.
You've got orders that total5,000 units for that item and
(27:43):
you have to allocate it.
You have to pick who you sendit to.
That's a supply chain problem.
Right, like you're running low,you're at risk.
You have to make a decisionthat's gonna impact the sales
team.
Many companies today willprobably just do FIFO right.
First in, first out.
Who placed the first order.
It goes that way.
But we started providinginsight.
That said, hold up.
(28:03):
We actually know whichretailers need this product
because we know what theconsumer is doing.
And if I could tell you thatWalmart has six weeks of
inventory, target has two,amazon is at one, you can now
start prioritizing yourallocation decisions in a way
that could try to protect yourrelationship with the retailer
and, even more important,prevent lost sales at the store.
(28:26):
Right, because these 1,000units you have are pretty
precious.
And if you're just doing firstin, first out, going back to
data-driven decisions, you don'thave any data.
But when you connect the pointof sale all the way up there,
you can bring that insight tothe supply chain team.
Now everyone can live moreharmoniously and you're making a
better execution decisionacross the board pretty
(28:46):
seamlessly if you could do it inone system.
Speaker 1 (28:48):
I was going to ask you a question, joel.
It's a topic that's, I think,passionate on your end is this
kind of planning process andplanning and forecasting process
and how that intersects withpoint of sale data.
I know you're workingpartnerships for us and have a
lot of experience in that space.
Could you speak a little bit tokind of what you see as best in
class and taking advantage ofpoint of sale data in that
(29:09):
planning domain, of what you seeas best in class and taking
advantage of point of sale datain that planning domain?
Speaker 2 (29:11):
Yeah, I think we're kind of in the midst of an
evolution, as we've talked abouta couple times earlier in this
podcast.
Historically, brands have usedtheir shipment data to really
forecast out what sales aregoing to be.
I mean, that's been kind of theprimary historical data point.
(29:32):
Then they're going to layer inall sorts of other things on top
of that.
Models have gotten a lot moresophisticated.
A lot of people are usingmachine learning, but if you
really break it down, it's likethe primary driver of predicting
future sales performance isgoing to be prior shipments.
I think for a long time there'sbeen a feeling well, point of
(29:52):
sale should be a betterindicator and it shouldn't just
be point of sale itself.
You also need to understandwhat inventory levels are like
at a retailer.
If you're selling a lot ofproduct but there's tons of
inventory that's piled up,that's very different than
you're selling a lot of productand the retailer is about to run
out.
But, as we talked about earlier, it's just been so hard to
bring that in.
So I think what we've seen isan interest.
I think over the last couple ofyears, more and more brands are
(30:16):
trying to figure out how theybring that into their models.
I think there's interestingkind of statistical questions
about how they actuallyincorporate that data.
But I know, logan, you and I wewere talking to a very large
company.
We're talking to them right nowabout this idea.
I mean, this is a brandeverybody would know and they
are very intent on saying wereally think that using point of
(30:39):
sale is going to be a betterpredictor than looking at
shipments and that's somethingthat they want to shift to.
And so I think everybody assumesthat it's going to switch over
at some point.
It's a question of when and howquickly that is.
I think everybody assumes thatit's going to switch over at
some point.
It's a question of when and howquickly that is.
And this gets to again.
The core problem we've beentalking about is point of sale
has been out there for a while.
It's been getting better.
People are recognizing withtechnology you can really
(31:01):
incorporate this.
It doesn't just need to be kindof a sideshow, but these shifts
take time.
Speaker 1 (31:06):
Well, really fascinating stuff.
Joel Manny, thank you both foryour insights, perspectives on
data.
I know this group's quitepassionate about it.
So great conversation andthanks for joining us, manny
Absolutely.
Speaker 4 (31:19):
It's one of my favorite things to talk about
right Data.
People think they know whatthey're doing and they're not
right when they talk about salesor talk about shipments or unit
sales.
Big decision.
Even if they are talking aboutunit sales at the store, is it
weekly, is it daily, like?
Are they actually knowing this?
Are the right teams gettingthis insight or is it kind of
siloed in their own world?
There's so much under the hoodthat is important.
(31:40):
I'd love to talk about this and, more than talking about it, I
love all the value we can bringto customers when we can get
them on our best practice.
Speaker 1 (31:52):
You've been listening to Manfred Reichi, subject
matter expert at Alloy AI.
That's all for this week.
See you next time on Shelf Life.
Popular Podcasts
Las Culturistas with Matt Rogers and Bowen Yang
Ding dong! Join your culture consultants, Matt Rogers and Bowen Yang, on an unforgettable journey into the beating heart of CULTURE. Alongside sizzling special guests, they GET INTO the hottest pop-culture moments of the day and the formative cultural experiences that turned them into Culturistas. Produced by the Big Money Players Network and iHeartRadio.
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
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