July 23, 2025 • 39 mins
With Big Tech hyperscalers investing billions in local data centres and the Government taking a cloud-first stand, the migration of our data and applications to cloud platforms is in full swing.
But as David Reiss, my guest on this week’s episode of The Business of Tech, told me that moving to the cloud isn’t necessarily the panacea if you are looking to lower the ongoing costs of running your IT systems.
The co-CEO of 30-year-old tech services firm, Equinox IT, says costs can escalate quickly in the cloud without the controls and culture in place.
“Most organisations we talk to now are in the cloud,” Reiss said, “but they are suffering some of those problems due to the fact that they kind of evolved into it rather than necessarily having it being initially a strategic objective.”
Cloud “bill shock” is an increasing reality, not just in New Zealand but globally.
“There has definitely been a lot of higher [spending] than was expected. Some of the cloud pricing models appear very opaque, they appear very confusing, and it can be very difficult to figure out what is costing you money in the cloud environment if it’s not set up well to begin with,” Reiss said.
So what's the answer? a cultural change in our organisations where IT departments and business teams get on the same page about what they need to do in the cloud - and plan accordingly.
Listen to the full episode of The Business of Tech, powered by 2degrees Business, for a candid discussion of cloud realities, the human factors shaping IT outcomes, and how businesses can prepare for the next wave of digital change. Streaming on iHeartRadio or wherever you get your podcasts.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:03):
Welcome back to the Business of Tech powered by two
Degrees Business. I'm your host Peter Griffin, coming to you
this week from Auckland's innovation hub, The Grid. Today we're
examining a pressing issue that resonates with tech leaders across
New Zealand and beyond. Cloud bill shock. That's the moment
when an organization migrates its applications and data to the cloud,
(00:26):
only to discover that the ongoing costs are much higher
than expected, prompting budget headaches, tough conversations, and in some cases,
real financial pain. We know that tech companies like a
WUS and Microsoft are investing billions of dollars in New
Zealand building hyperscale data centers. They're touting big productivity gains
(00:49):
for our businesses and government departments moving their applications and
data to the cloud. There are genuine performance improvements and
cost savings to be had in doing so, but you
also need to know what you're doing because the cloud
can blow up in your face. Take, for example, a
recent case in Wellington were a user racked up an
(01:09):
eyewatering sixteen thousand dollars bill after experimenting with a Google
AI tool. The bill was refunded after it garnered some
media coverage. Globally, there are stories of organizations, sometimes even
large multinationals, suddenly facing monthly invoices running to tens of
millions of dollars. Its cloud usage scales faster than anyone anticipated.
(01:34):
When the Silicon Valley News website the information was leaked
some AWS documents in twenty nineteen, they showed that Apple
and other companies had seen big spikes in their cloud
cost the previous year. That was put down two so
called egress fees, the costs associated with moving data out
of the AWS cloud or to a new geographical location.
(01:57):
It's free typically to put data in these clouds, it's
expensive to get them out. Now those charges are legitimate.
It costs money in network bandwidth to shift large amounts
of data. This is the nature of demand based pricing
which underpins public cloud platforms. Again, it pays to know
(02:19):
exactly what you are doing to avoid bill shock. So
joining me this week is David Reese, co CEO of
equinox It, a stalwart of New Zealand's tech services sector
who's worked with scores of Q organizations to help avoid
or fix precisely this kind of scenario. David brings a
(02:41):
wealth of experience and not only cloud technology, but also
in helping businesses develop the right capabilities, visibility and team collaboration.
They need to ensure that cloud investments deliver true value
instead of nasty surprises. So this conversation is essential listening
for anyone read with cloud budgets. We're planning a migration
(03:03):
in the coming year. Here's the interview with Equinox. It's
David Reese. David, welcome to the business of tech. How
are you doing on a rainy day in Auckland.
Speaker 2 (03:22):
Yeah, good, thank you Dry, thankfully.
Speaker 3 (03:23):
Take us through the history of Equinox.
Speaker 1 (03:26):
I t it's been around for a long time, a
store really of the IT scene techiclarly in Auckland and Wellington.
You've got government clients, You've got a reasonable presence in Wellington.
But take us through the origin story of Equinox.
Speaker 4 (03:39):
Yeah, Equinox has been around for thirty years, proudly in
New Zealand only operated and we've literally got some of
the same clients that we did thirty years ago. And
as you say, government agency, so started in Wellington and
then decided to diversify into Auckland. So I joined the
company about eight and a half years ago now where
(03:59):
we're looking to change a little bit about how we're
approaching the Auckland market because as an organization, a team
at that point of six or seven people, we needed
to have something that we're going to be famous for.
So we said, well, look, we've been the longest in
a fact that stage only devots partner for Microsoft in
New Zealand. Let's double down on that with the cloud
(04:20):
skills we've got and actually create ourselves a position which
is very much about driving genuine business change through the
use of cloud and DevOps technologies. And so that's what
we've been doing in Auckland, and we've been making that
conversation you know, increasingly more relevant to our Wellington clients
as well.
Speaker 2 (04:36):
And so that's kind of, you know's that's our.
Speaker 4 (04:39):
Niche is where those three things combine, and a key
part of what makes that successful is our push to
always focus on the people as well as the technology
and obviously the business process because nothing we do as
an IT project is all a business project.
Speaker 1 (04:52):
Yeah, and you do a lot of training as part
of that. Thousands of people have gone through your training
programs so that we sometimes forget.
Speaker 3 (04:59):
I think we do criticized in New.
Speaker 1 (05:00):
Zealand about perceived lack of capability to use this technology
to best effect. So training as part of that will
talk a little bit more about that. But Equinox has
really evolved through that whole transition a couple of major transitions,
the rise of apps, for instance, but that transition from
(05:23):
having your data and your applications on a server in
the basement of your building.
Speaker 3 (05:29):
To the move to the cloud.
Speaker 1 (05:31):
And we've heard so much about that over the last
couple of years. We've had Microsoft debut it's cloud region
in New Zealand, We've got a WUS supposedly on the way,
but lots of investments in Frattil investing CDC, so literally
probably billions of dollars or one hundreds of millions at
least going into cloud. We've been told to be more productive,
(05:53):
we need to get to the cloud. What's your perspective
over the time you've been at Equinox, which is really
that period when it's Excel. What are some of the
trends that you've seen in terms of New Zealand companies
going on that journey.
Speaker 4 (06:05):
Yeah, Look, I think and definitely those that were first going,
and I used to say to clients, it's not really
so much that that much had been released to the cloud,
but a lot had escaped. So there was a lot
that wasn't really done with a lot of strategic thinking
in place.
Speaker 2 (06:18):
But you know that they got in and.
Speaker 4 (06:20):
They started using and they started doing, which was great.
Most organizations we talk to now are in the cloud,
but they are suffering some of those problems at the
fact that they know they kind of evolved into it
rather than necessarily having it being initially a strategic objective,
and now that it is strategic, there's a little bit
of unpicking that needs to get done. And so we
(06:40):
often come in and find organizations needing to kind of
reimagine what they're doing in the cloud and set it
up better for them for now, but also for the
future as well, and take away some of that you know,
tech jargon of technical debt, some of the things that
we did that maybe were you okay at the time,
but now we look at them and say, actually, that
wasn't our best work. We need to reimagine that and
(07:01):
redeploy it for something that's going to last us for
the next three to five years. Obviously, not much in
tech lasts three to five years, but those foundational layers
and a real understanding about what an organization is trying
to achieve, those do stay a lot more static, and
so having that stuff really understood and the foundations set
up for what the business is and should be famous
(07:23):
for is really important.
Speaker 1 (07:24):
Technically, it's really important obviously what cloud platforms you move
to have the capacity to host that apps. Your apps
might have to be adapted or rewritten for cloud environments,
so there's that aspect of it, and the so called
DevOps is.
Speaker 3 (07:43):
Central to that.
Speaker 1 (07:44):
At the same time, we've got this transition, and I
know talking to people in government departments, this is really
important for them.
Speaker 3 (07:50):
From a budget point of view.
Speaker 1 (07:51):
You're going from a sort of a capital expenditure environment
where you're spending a lot of money on hardware that
is transitioning to more of an operational experidence your model
where you're paying literally subscriptions or contracts to aws, Microsoft,
Google and others every month. So it's a real challenge
(08:14):
to get your head around it.
Speaker 3 (08:15):
I guess for.
Speaker 1 (08:17):
People who are trying to predict this is how much
capacity we're going to need to serve new Zealand citizens
or business customers. How much is this actually going to cost?
Speaker 4 (08:27):
Absolutely, And I think there's massive power in that change
to an operational expense rather than the need to buy hardware, right.
Speaker 2 (08:33):
I mean, in the bad old days, you'd go, oh.
Speaker 4 (08:36):
We need to deploy one new business application that's now
pushed us above the capacity of what we've got our
data center. Now we need to buy and you serve,
and then we need to buy new storage. You're talking
about stuff being shipped over in a worst case scenario
and you've got maybe a two month lag before you
can put one new thing in, whereas now the cloud,
of course is the exact opposite. You want something, you
(08:57):
can spin it up in minutes, so there's a huge
amount of fleet stability to it. But of course you know,
you can depreciate that hardware and you can treat it
the way you're used to treating assets. And a lot
of organizations that are set up really capital heavy in
the way they do their financing, it's a shift that
needs to get made in their thinking to be able
to really get the power out of cloud. And I think,
(09:17):
as you kind of mentioned, there was a lot of
stuff in their own data center. You know, the great
thing about KIWIS is our number eight wire mentality. Sometimes
the risk about QWIS is our number eight wire mentality,
where you have business critical systems running on a PC
under someone's desk, or you've got, you know, some assets
that are really being sweated a lot longer than they
(09:39):
should be, a lot longer than their manufacturers would recommend.
And that stuff, of course always seems like a great
idea until something goes wrong, and I think there's that
insurance policy that you can get by having that stuff,
having that kind of raw infrastructure looked after by the
cloud providers, and obviously you know, you do have other
options with data center hosting and that sort of thing
(09:59):
as well. They generally don't give you, you know, generally
what I call weapons grades, security and privacy and scalability,
which you do get out of the hypervisors like your
Microsoft's and AWS and Googles look by.
Speaker 1 (10:11):
And which I think the track record has been pretty
solid in New Zealand in terms of reliability off the cloud.
You know, they're configured in ways we do have natural
disasters in our country, but they typically have three points
of presents at least in different regions of a city
or sometimes in Hamilton and Auckland, for instance. So the
(10:32):
model seems to have proven itself. What I am hearing
though from businesses, I'm really keen to get your insights
into this is a bit of alarm. When companies do
get to the cloud at the ongoing costs are often
higher than they anticipated, so called bill shock. And this
is not a New Zealand saying there are companies all
over the world that we've heard stories about that are
(10:54):
literally paying tens of millions of dollars in cloud costs
and freaking out about it. Then, of course we're going
to talk specifically about artificial intelligence, where you have these
data centers full of GPUs that are doing really sophisticated
AI staff. But you know the capacity required and the
resources are significant, so you've got to pay for that.
(11:17):
But let's talk us through about what you've seen in
terms of customers. Maybe you having to come into help
customers optimize that spend and in some cases sort of
rescue them from a situation.
Speaker 3 (11:30):
That's really bad.
Speaker 2 (11:31):
Yeah.
Speaker 4 (11:31):
Look, obviously, I mean there are some high profile cases
of cloud costs really going feral.
Speaker 2 (11:36):
We haven't experienced any of those ourselves.
Speaker 4 (11:38):
And our clients haven't, but there has definitely been a
lot of much higher spend than was expected, and I think,
you know, the recent economic situation has caused a lot
of organizations to actually take a look at their spending
quite closely, and often it's been the first time that
a huge amount of detail has been applied to cloud
costs because they've realized that actually, some of the cloud
(12:00):
pricing models.
Speaker 2 (12:01):
Appear very opaque, they appear.
Speaker 4 (12:03):
Very confusing, and it can be very difficult to kind
of unpick them and figure out what is costing you
the money in the cloud environment if it's not set
up well to begin with, which kind of takes us
back to that point I made before about things evolving
over time without the appropriate guardrails and controls in place,
without the appropriate visibility means that you end up with
(12:25):
a bit of a spaghetti mess to work through. So
when you kind of see those problems at the headline
level of cost and you know you need to manage
that cost down and you only just realize that you've
got a cost problem from the headline, that's when things
do get pretty challenging, and organizations like ours, you know,
come in and help our clients to identify and their
(12:46):
manage that cost better. But you know what we much
prefer to do as be involved at the start of
a thing or be involved in helping an organization to
really look at how they could be using cloud better
to get business value and have those conversations about what
will be coming in the coming months and years and
set them up well with everything. And obviously a big
(13:08):
part of that is having appropriate visibility and control around
things like cost.
Speaker 1 (13:13):
Yeah, and you know from looking at what Equinox does
talked about the training, it is really about about people
and this is.
Speaker 3 (13:21):
Where it projects usually go wrong.
Speaker 1 (13:23):
The technology might be world class and have the ability
to do what you want to do, whether it's you know,
selling tickets on a New Zealand or a banking system
or whatever, it's the people within the organization and often
that where it falls down is the culture and it
not gelling with the business or the actual business culture
the people who are responsible for driving the business forward.
Speaker 3 (13:46):
What are some of the red.
Speaker 1 (13:47):
Flags you see when you go into an organization that
might be struggling a little bit with its cloud spend.
Some of the things you see we go oh that
instantly needs addressing that's where the core of your problem is.
Speaker 4 (14:00):
Look, it really is interesting how much an organization culture
can permeate into the way that systems get built, managed
and operated.
Speaker 2 (14:08):
So anytime that you do see teams within.
Speaker 4 (14:11):
The IT team not getting on very well, as you
say that that very traditional disconnect between the IT and
the business, those will usually play out in an IT
environment that is not set up the way it's best
for the business.
Speaker 2 (14:24):
You end up with.
Speaker 4 (14:27):
A lot of people within the area of control optimizing
for their own team or optimizing for their own business unit,
their own budget. Whereas really because of the fact that
a lot of what you have in cloud covers everything
you do, but then you've got little bits that you
need to be specialist for certain areas, you need to
have that approach of optimizing for the whole and that's
how you get by far the best value out of
(14:48):
the cloud and you manage to keep your overall costs down.
The other interesting thing in IT is I guess looking
at cost and cost right, and so total cost of
ownership being what you genuinely influenced with cloud headline cost
not necessarily and in fact, if you're comparing your server
that you've got under your desk to a serve you
have hosted in weapons great infrastructure in Azure or Google
(15:12):
or a US.
Speaker 2 (15:13):
Obviously you're not going to get.
Speaker 4 (15:16):
The same cost comparison there, But what you are going
to get is insurance from risk, and you are going
to get reliability and resilience and inability to recover from
a failure. So you have to know how to attach
a value to those things to actually get that view
of that total cost. There's a great little analogy online
of pizza as a service that kind of talks about
the different cloud models and likens them to a pizza
(15:38):
delivery business.
Speaker 2 (15:39):
At one end, you've.
Speaker 4 (15:41):
Got someone making all the pizza and all the ingredients
and putting it together for you and cooking it and
delivering it and putting it on your table. And at
the other end, you're making your own dough and you're
putting all the ingredients on, you're making your own pizza.
And I think it's really useful to kind of think
of that in terms of that layering that actually comes
into the cost for managing it, and unless you are
really clear about how you strip out some of that cost,
(16:03):
and usually that's just about putting an ento higher, higher
value tasks. That's when you can start doing a clear
comparison between cloud and on prem.
Speaker 1 (16:12):
A lot of this is about human behavior and human nature.
I guess the trickiest thing is when you go into
an organization and you just see a dysfunctional relationship there
between maybe the CIO or the IT team and what
the product owners are, the people who are driving the
business forward bringing.
Speaker 3 (16:31):
In the revenue.
Speaker 1 (16:33):
So do you sometimes literally sort of have to almost
be like therapists to them, get them all in a
room and sort of say, look, you all need to
get on the same team.
Speaker 3 (16:41):
How do you approach that?
Speaker 2 (16:42):
Yeah, very much.
Speaker 4 (16:43):
Sometimes it feels like we've got the customers on the
couch when we're talking them through things, and look and
we understand both sides of that equation. So we can
be genuinely empathetic with the situation that people are. And
it's one of those usual things, right. People don't come
to work think they're going to do a bad job.
People come to work with the confines of the role
that they are in, wanting to do the best that
they can do. And so we've got some of these
(17:05):
tensions or conflicts. Is not ever really down to the people.
It's down to the position that they are in and
maybe a lack of understanding about how they could do
things better together.
Speaker 2 (17:14):
But some of those traditional.
Speaker 4 (17:15):
Dysfunctions are You've got a product team or project teams
going and creating a whole bunch of things, and they
don't have within their remits, sometimes not even within their
business case, what the genuine long term cost to operate
the thing is going to be, and so they don't
worry about that because that's not something that they're tasked with.
There's not a KPI around them understanding that stuff, and
(17:36):
so they throw it over the fence to the operations teams,
and their role is purely about keeping things reliable and
performant and ready when people need them. And they've now
effectively adopted a large amount of costs that they wouldn't
assigned up for if they.
Speaker 2 (17:50):
Had the choice.
Speaker 4 (17:51):
So neither of those two functions has done the wrong thing.
One has been tasked with driving business value, getting more customers,
making efficiency done it and the business elements. The other
team is tasked with keeping the costs down, keeping things
reliable and going, and they're doing that as best they can,
and it's just about making those two groups of people
able to have a good conversation and make sure that
(18:14):
what is delivered as something that achieves both of those objectives.
Speaker 1 (18:16):
And I think we've got to a point with the
cloud now where you can estimate, you know, if you
actually understand where your business is going and how it's growing,
or how new products are going to impact on your
digital footprint and all that sort of thing, you can
sort of you can estimate your costs over the coming
years relatively accurately, and there are volume discounts and things
(18:39):
to minimize your cost the vendors are offering that are
they good enough at giving you enough visibility into your
costs and giving you enough understanding? Are there enough tools
and dashboards and methods of actually understanding what you're paying?
Speaker 4 (18:56):
A couple of answers that I guess one. I mean,
the tools are great. The problem with all these things
is not the technology. There are tools that will do
it if you use them well and if you've got
your cloud set up well to use them right. And
again that's that technical debt thing where it can be
quite hard for a lot of organizations to see why
they should invest in going through and retrofitting some of
the better practices and better visibility metrics and that sort
(19:18):
of thing into some of their projects that they completed
and they'll put down and the team that used to
work on those isn't there anymore. It's quite hard to
get people to refocus on the old stuff because it's
not the exciting, new, sexy things and need to get
out the door. So there's that element to it in
terms of the tools are there but not necessarily being used.
(19:40):
But the other element to it is just those tools
and just the accountability for cost.
Speaker 2 (19:47):
And the even.
Speaker 4 (19:50):
The responsibility and desire to look at it often sits
within the team and the I T team only, and
so that information that they have isn't perm added to
the rest of the business and they are only able
to make the decisions that they can within their kind
of scope and knowledge, if that greater level of visibility
around cost and performance and what resources are there, how
(20:12):
they're being used, and ultimately what the conversion rate is
between some services and a cloud provider and a number
of new clients to get in the door, or the
increase in cross sell and upsell because of an ability
to get a new and exciting service out to the
customer until you can kind of draw a line in
your budget between the business value metrics and the baseline
(20:33):
costs of the infrastructure and the people and the partners
that you need to support that infrastructure, which again needs
to be optimized and made best for cloud. Until you
can get that kind of steel thread, it's really really
hard to optimize for value.
Speaker 2 (20:46):
So the tools are there, it's the cultures.
Speaker 4 (20:49):
That aren't quite there at the moment. It's the processes
that aren't there in the guardrails. Everything from the light
bulb going off for someone imagining an idea to the
end of the life cycle of of.
Speaker 2 (21:00):
An IT system.
Speaker 1 (21:01):
And you know, this is really sort of complex, especially
if you're a business. We've got a lot of digital
first businesses and ual software businesses like zero that have
done really well at exporting internationally. So for them, big
chunk of their cost is literally development as well as
hosting millions of transactions.
Speaker 3 (21:23):
Every day across their servers. So this is really important.
Speaker 1 (21:26):
It's actually led to the rise of a whole discipline,
the phinnops discipline, which you guys are very active in
as well.
Speaker 3 (21:33):
Talk us through how that works.
Speaker 4 (21:34):
And it's interesting you bring up the case of the
software created because I think a lot of the tools
and a lot of the practices that you know the
right thing to do if all of your business and
therefore a large percentage of your cost is that it infrastructure.
Obviously it gets a good amount of attention, but for
businesses where the main thing they do is retail, it
just so happens. They need IT systems to make those happen,
(21:56):
it's a lot harder to I guess, get the priority
around some of these things. And look, and that's WHEREOPS
or cloud cost optimization really does come in, and it
is about creating a group of people who are able
to collaborate on maximizing business value out of technology. Actually
(22:16):
it is not just cloud's kind of in some ways
a bit easier. You've got some more levers you can
pull to make it so that your costs better value
the better match the value that you get. But you
can still do that stuff on traditional tools as well.
So the approach TOOPS is very much one that, yes,
there's some technology enablement that needs to get done. There
(22:39):
are some tools that you can put in place, there
are some different technical practices that you could use, but
the line share of it is actually around the business process,
around business casing, around forecasting, and about turning some of
those again what look to be maybe over complex pricing
models compared to the old way work for you rather
(23:01):
than work against you.
Speaker 2 (23:02):
So I think there's a huge.
Speaker 4 (23:04):
Amount of power that can be gained from a cost
benefit point of view if you embrace the way cloud works.
Speaker 2 (23:11):
But that means a lot of change.
Speaker 4 (23:13):
That means change for the people, change for the skills
that people have, and changed the processes that the organizations have.
Speaker 1 (23:19):
I've interviewed some companies that have done incredibly well. I'm
thinking of a company like Volpower Health recently sold to
a Korean medical company for three hundred million dollars. Founded
in Wellington, Ralph heinem who we had on the podcast
about a month ago.
Speaker 3 (23:36):
Incredibly successful, very.
Speaker 1 (23:37):
Much built on as you're on the Microsoft platform and
as a result of that was able to leverage all
the tools, the AI sort of tools. So there is
incredible power, as you say, if you really understand and
sort of a cloud centric organization, it can really pay off.
I mentioned AI. So that's the next phase. We're going
(23:58):
through a lot of hype around it. Out of social intelligence,
the right of AI agents, and again, you know a
lot of companies are going, wow, okay, I want to
spin up agents and allow my customers to use agents,
and there are platforms to make that easy to do.
Speaker 3 (24:15):
But there's a lot of consumption based business models emerging.
Speaker 1 (24:19):
And for New Zealand companies that are trying to build
and use their own agents, I guess getting their head
around what are the ongoing implications costs wise of that
is a real challenge.
Speaker 4 (24:31):
Yeah, absolutely, and it is technically complex the way things
are priced at the moment. Some are comparatively simple. But
the second you want to get more than you kind
of get from the consumer and consumer like versions of
the AI tools, the second you want to start pumping
your business darter into them and getting some genuine insights
that you can use, and you want to start using
(24:51):
AI to optimize workflows that you have and automate those.
That's when you start getting into the territory of some
of those reasons you need to deploy that are going
to you.
Speaker 2 (25:02):
Know, potentially massively expand without letting you know first.
Speaker 4 (25:05):
There are some services that I would argue probably should
be opt in rather than opt out, but by default
they turn on because the vendors think that you obviously
want all the AI power you could possibly ever get, therefore,
best make it easy to consume. But I think the
other problem that we are seeing because of the fact
there is such a groundswell of interest around AI, is
(25:28):
that a lot of people are going into it and
feeling almost a bit of desperation to be doing it
and to be seen to be doing it, and maybe
not putting in some of the business discipline that you
can still do even with this fast moving stuff. So,
you know, we like to think of having a living
business case, so you don't need to go all the
way down the path of completing a business case before
(25:50):
you do anything. You don't need to analyze everything before
you start doing anything. But what you should be doing
is factoring in what you believe the tangible business value
you're going to get out of your experiment is from
the get go, and it's part of the experiment is learning, oh, no,
our costs are going to be slightly higher than we
thought they were going to be, or actually the value
that we thought we were going to get isn't But
now that I look at it, we're going to get.
Speaker 2 (26:10):
A different piece of value.
Speaker 4 (26:12):
So until you do tie the genuine business outcomes into
the business outcomes and the business costs into what you're
doing within AI. There are definitely risks and costs going
well north of what a good business decision would indicate
they should.
Speaker 3 (26:26):
If you're in a team or an individual in a mid.
Speaker 1 (26:29):
Size or organization that's responsible for finops and for Equinox's
team as well, increasingly, you've got to have pretty good
AI expertise in those teams to actually understand what the
cost implications of the company rolling out AI actually are.
Speaker 4 (26:44):
Yeah, definitely, and looking there are some traditional sort of
high risk workloads for finops. Those are the ones that
we do focus on, you know, AI obviously, and given
that it's emergent and the pricing models are emergent and
they're changing regularly, it's to read a key ones to
make sure that is understood.
Speaker 2 (27:02):
And big data is.
Speaker 4 (27:03):
The other one of the other major workloads that would
often you know, as you allow things to get into
the hands of citizen developers and people can start creating
their own data sources or merging data sources and then
extrapolating from that to get more value, it can be
quite hard to kind of put constraints on those things.
When the old world again, you'd have a sequel server
(27:23):
that the IT team would look after.
Speaker 2 (27:25):
If that started hitting capacity.
Speaker 4 (27:26):
Well you slow down and you can't do things today
of course without.
Speaker 2 (27:32):
Some controls in place.
Speaker 4 (27:34):
You know, the cloud based big data solutions will scale
infinitely because you know, they want you to buy more
of their products, but also because they assume that what
you're doing is valuable. So it is the onus is
on the organization to put those limits in place on
you know, the maximum expected value that's going to be
gotten from something and not just allow endless consumption.
Speaker 1 (27:53):
Yeah, and one thing I do hear a lot oft
and I think this speaks to the fact that we've
been a little bit behind in getting our data house
in order. In New Zealand is where the big bills
rack up is constantly shuttling data between different silos because
you need to bring all of that data in free use,
and you haven't done the groundwork to put all that
(28:14):
place in one area. You know, we've heard of data
lakes and lake houses and all these terms to describe that.
But there's been this massive push in the industry get
your data in one place so we can query that
and you're not having to shell it around and rack
up cloud builds.
Speaker 4 (28:30):
Absolutely, and that was one of the earliest kind of
bill shock instances really was when organizations read the fine
print and realize that you get charged to take data
out of the cloud platforms. So the second you kind
of half pregnant in cloud, that's kind of where your
cost problems really kick in because you've got some stuff
on premise, you move some things into cloud, but you
never move enough into the cloud that you can turn
(28:51):
off what you've got on site, So you're actually double
paying fear infrastructure at that point, in fact sometimes more
because you go, well, we need to be secure, so
we'll put security things both on site and also in
the cloud. And oh, now actually we've got these two
systems need to talk to each other. We're better make
our network more expensive to cater for it, and then
of course more expensive times ten because the cloud vendors
(29:12):
are charging you for the way out as well. So
definitely that half pregnant approach is really problematic, and AI
is only adding to that. And I think where we see,
you know, I think you need to be careful about
where you experiment on AI.
Speaker 2 (29:27):
So AI is going to.
Speaker 4 (29:29):
Be pervasive across every piece of software that you use today. Therefore,
if the vendor of that software is going to be
building their own AI capability, should look at whether you're
better to wait for them to do it. Because they
will build it, they will package into they're pricing, there
won't be those unknowns. Rather than going, oh, we need
to build something that's completely unique to us, and that
(29:50):
involves extracting a whole bunch of data, standing up a
whole bunch of compute resource yourself, doing a lot that's
going to cost you a lot of money to then
in six months find that actually that you know, unique
thing that you built is actually now available as part
of the product. And we have seen some of those
examples or organizations have spun up something AI that really,
(30:10):
you know, it will be table stakes at some stage
in the coming.
Speaker 1 (30:13):
Months, particularly because you know, I sort of here, particularly
from vendors like IBM, are saying, oh, the future is
hybrid cloud.
Speaker 3 (30:22):
It seems logical to put everything with one vendor.
Speaker 1 (30:25):
You're going to you're going to get a better overall
price the more business you have with one vendor. But
it's not as simple as that there are some people
still have stuff on their premises. There are others who
will have multiple cloud instances. They might have something on
as you're on something on IBM's cloud because they want
to use Watson.
Speaker 3 (30:46):
Or something like that.
Speaker 1 (30:47):
So the reality is a little bit messier, I guess
is hybrid cloud really what is sort of has emerged
as a new norm and new Zealand.
Speaker 4 (30:54):
I mean hybrid cloud, especially if you include on premise
in that. Yes, that's the default for a lot of organism. Again,
there's maybe not in some cases not enough intentionality about
how it's been done. And I think there's a couple
of different things. Is taking the best out of a
cloud provider right and realizing the particular solution that you
(31:15):
need that's going to have a lot of business value
is best to live it out of aws where most
of the other stuff you've got is an azure that
to my mind, is a useful model. That's an approach
that works. Instead saying oh, we've got you know, we
need to have a data lake, Oh let's have another
data lake and another provider, or saying oh we need
(31:36):
to have our virtual machines in one place, it's have
a whole bunch of virtual machines in another cloud provider.
Speaker 2 (31:40):
That makes a lot less sense to me.
Speaker 1 (31:42):
And is it getting more competitive as more capacity comes
into the market and more players are offering the so
called hyperscale clouds. Are you seeing a good competitive tension
out there between all of these big multinationals a little?
Speaker 4 (31:56):
I mean, it's interesting in New Zealand because we've traditionally
been so heavily into the micro Soft ecosystem for everything,
and so when AWS came into New Zealand, they absolutely
you beat Microsoft to the punch around that timeframe, and
so the cloud volume for a WS increased very very quickly.
In Microsoft, I think it's still probably playing catch up.
So I think the interesting thing for New Zealand is
(32:19):
that now Microsoft has caught up. From a feature's point
of view, it kind of is the path of least
resistance for most organizations to move into Asia, so government right,
well exactly, yeah, yeah, And therefore it becomes you know,
niche positions are taken by other players. But I guess
you know, with organizations that are in a brown fields environment. Again,
(32:42):
it's kind of it's the difference between the volparers of
the world and an organization that's been running for you know,
twenty five years, and they've got a whole bunch of
legacy and data centers and that sort of thing. In
a brown field's environment, it's very very hard to go
pure into one thing or another. It's very hard for
organizations to turn off those last things that sitting on
their sites.
Speaker 2 (33:01):
It's very hard to.
Speaker 4 (33:02):
Say no, no, we're not going to pack aws for
something and Google for.
Speaker 2 (33:05):
Something else, or Microsoft for something else.
Speaker 4 (33:08):
But it's just about I think, being really clear where
some of those costs and complexity drivers.
Speaker 2 (33:13):
Are going to be.
Speaker 4 (33:13):
That mean that what looks like a ten percent better
solution functionality wise in a second cloud might be double
the cost compared to having it in something that an
environment you already have.
Speaker 1 (33:25):
Yeah, So what is your advice to those sorts of
brown fields organizations that they might be in the cloud
and becoming a bit more aligned at the ongoing cost
of what they're doing, or they may be on that
journey and about to do a migration. What are the
sort of things that you sit down and talk to
these sorts of organizations about what are the principles or
(33:46):
the fundamentals they need to consider.
Speaker 4 (33:48):
Well, look, amend does all start with a business case, right,
and it's going to be really very clear on what
the outcomes are that are trying to be achieved, and
it has to be a very conscious decision, and one
is made with a real understanding of the total cost,
not of the.
Speaker 2 (34:04):
Cost of servers.
Speaker 4 (34:06):
And if it's only going to be done with a
top line cost, then you know, cloud might not be
the right solution. It might not be able to be
justified for doing so. But for an organization that has
stuff that is in legacy land and can't be moved
into the cloud, I think you need to realize that
(34:26):
at some point you're going to have to make that change.
And it's kind of one of those you know, the
best time to do it was ten years.
Speaker 2 (34:32):
Ago, next best time is now.
Speaker 4 (34:34):
If you can't modernize your system so they can run
in today's infrastructure, then you are going to hold the
business back. It's going to be an anchor that is
going to be stopping the business in a whole bunch
of ways. Because the second everything is powered by AI
and whatever the next set of new technologies are, and
the next you know, exciting ways of engaging staff and
(34:55):
engaging customers if you can't take advantage of those, you
can lose competitive advantage anything that touches consumer. For example,
you know, before too long, if your services don't show
up on an AI search and AI query, you will
cease to exist. You will be invisible to everybody because
the number of eyeballs are going to be going to
websites to look at things is going to rapidly decline
(35:17):
is AI takes over and replaces search. So you know,
if you've got an old legacy system that you could
never plug meaningfully into AI, you can't take advantage of
cloud's ability to scale.
Speaker 2 (35:28):
You need to look at how you can.
Speaker 4 (35:30):
You know, even if it is a multi year program
to modernize and move away from that application, then that
should be absolutely on the table for consideration. And it's
just about looking at the things that you need to
be doing, looking at the things that are genuinely business
value generating invest in those, modernize those, get them to
(35:52):
where they need to be. And other things where you
know you've built something, or you're using something that maybe
could be done another way by assas products or something
that you could consume off the shelf.
Speaker 2 (36:01):
Any of those kinds of movies you can make the better.
Speaker 1 (36:03):
So it sounds like it's you know, it's not something
that's going to be pigeonholed in the IT department of
a New Zealand business. This is something that the CEO,
you know, the top executives need to have real clear
visibility into and understanding of.
Speaker 2 (36:18):
Absolutely.
Speaker 4 (36:20):
I mean, I think the days of a business strategy
being handed to the I T team to then form
an IT plan, you know, is very backward. I think
if you look at your traditional smack influences of social mobility,
analytics in cloud, if technology and technology innovation and disruption
isn't factored into business strategy, then you know, there aren't
(36:42):
very many industries where you know, that's a viable strategy anymore.
I think you need to be taken into account the
massive changes they're going to make for consumer buying habits
and B to B buying habits, and just the ability
to genuinely compete in whatever an organization does is going
to be increasing reliant on good use of technology, so
(37:02):
it has to be baked into every level of the
organization from the boad doown.
Speaker 1 (37:12):
That wraps up another episode of the Business of Tech
My thanks to David Reese from equinox It for sharing
practical wisdom on tackling cloud bill shock, including why the
hardest problems guess what are more often found in company
culture and processes than in the technology itself.
Speaker 3 (37:32):
Look.
Speaker 1 (37:32):
I've been writing about the cloud for years, and every
time I cover the tech company's claims of the benefits
of migrating to the cloud, which are legit, I usually
get one or two readers contacting me to say, Basically,
it may look that way, but the costs really do
escalate over time, and it's actually more cost effective in
many cases to maintain your own infrastructure, which is an
(37:56):
approach that's sort of gone out of fashion. I've heard
stories of companies leave the cloud due to costs. The
company thirty seven Signals is a very high profile case
internationally of that. I link to the CEO's blog post
on how the company moved onto its own hosted hardware
to cut its three million dollar a year cloud bill.
(38:17):
The cloud's true value can be unlocked, but only when
business leaders, it and finance teams work together with transparency,
shared goals, and the right guard rails long before the
invoices start to arrive. Whether you're scaling up AI or
just managing day to day cloud operations. A strategic and
(38:38):
collaborative approach will protect you from those nasty surprises and
help ensure your investment actually pays off. Equinox is doing
a lot of work in this area. I recently helped
them write a white paper on this whole topic of
cloud cost optimization. I'll link to it in the show
notes at Businessdesk dot co dot nz. You'll find them
there and my weekly tech reading list. Just go the
(39:00):
podcast section. Thanks for listening. If you enjoyed this episode,
don't forget to subscribe and share it with your colleagues
on iHeartRadio or wherever you get your podcasts. I'm Peter Griffin.
See you next time on the Business of Tech.
Welcome back to the Business of Tech powered by two
Degrees Business. I'm your host Peter Griffin, coming to you
this week from Auckland's innovation hub, The Grid. Today we're
examining a pressing issue that resonates with tech leaders across
New Zealand and beyond. Cloud bill shock. That's the moment
when an organization migrates its applications and data to the cloud,
(00:26):
only to discover that the ongoing costs are much higher
than expected, prompting budget headaches, tough conversations, and in some cases,
real financial pain. We know that tech companies like a
WUS and Microsoft are investing billions of dollars in New
Zealand building hyperscale data centers. They're touting big productivity gains
(00:49):
for our businesses and government departments moving their applications and
data to the cloud. There are genuine performance improvements and
cost savings to be had in doing so, but you
also need to know what you're doing because the cloud
can blow up in your face. Take, for example, a
recent case in Wellington were a user racked up an
(01:09):
eyewatering sixteen thousand dollars bill after experimenting with a Google
AI tool. The bill was refunded after it garnered some
media coverage. Globally, there are stories of organizations, sometimes even
large multinationals, suddenly facing monthly invoices running to tens of
millions of dollars. Its cloud usage scales faster than anyone anticipated.
(01:34):
When the Silicon Valley News website the information was leaked
some AWS documents in twenty nineteen, they showed that Apple
and other companies had seen big spikes in their cloud
cost the previous year. That was put down two so
called egress fees, the costs associated with moving data out
of the AWS cloud or to a new geographical location.
(01:57):
It's free typically to put data in these clouds, it's
expensive to get them out. Now those charges are legitimate.
It costs money in network bandwidth to shift large amounts
of data. This is the nature of demand based pricing
which underpins public cloud platforms. Again, it pays to know
(02:19):
exactly what you are doing to avoid bill shock. So
joining me this week is David Reese, co CEO of
equinox It, a stalwart of New Zealand's tech services sector
who's worked with scores of Q organizations to help avoid
or fix precisely this kind of scenario. David brings a
(02:41):
wealth of experience and not only cloud technology, but also
in helping businesses develop the right capabilities, visibility and team collaboration.
They need to ensure that cloud investments deliver true value
instead of nasty surprises. So this conversation is essential listening
for anyone read with cloud budgets. We're planning a migration
(03:03):
in the coming year. Here's the interview with Equinox. It's
David Reese. David, welcome to the business of tech. How
are you doing on a rainy day in Auckland.
Speaker 2 (03:22):
Yeah, good, thank you Dry, thankfully.
Speaker 3 (03:23):
Take us through the history of Equinox.
Speaker 1 (03:26):
I t it's been around for a long time, a
store really of the IT scene techiclarly in Auckland and Wellington.
You've got government clients, You've got a reasonable presence in Wellington.
But take us through the origin story of Equinox.
Speaker 4 (03:39):
Yeah, Equinox has been around for thirty years, proudly in
New Zealand only operated and we've literally got some of
the same clients that we did thirty years ago. And
as you say, government agency, so started in Wellington and
then decided to diversify into Auckland. So I joined the
company about eight and a half years ago now where
(03:59):
we're looking to change a little bit about how we're
approaching the Auckland market because as an organization, a team
at that point of six or seven people, we needed
to have something that we're going to be famous for.
So we said, well, look, we've been the longest in
a fact that stage only devots partner for Microsoft in
New Zealand. Let's double down on that with the cloud
(04:20):
skills we've got and actually create ourselves a position which
is very much about driving genuine business change through the
use of cloud and DevOps technologies. And so that's what
we've been doing in Auckland, and we've been making that
conversation you know, increasingly more relevant to our Wellington clients
as well.
Speaker 2 (04:36):
And so that's kind of, you know's that's our.
Speaker 4 (04:39):
Niche is where those three things combine, and a key
part of what makes that successful is our push to
always focus on the people as well as the technology
and obviously the business process because nothing we do as
an IT project is all a business project.
Speaker 1 (04:52):
Yeah, and you do a lot of training as part
of that. Thousands of people have gone through your training
programs so that we sometimes forget.
Speaker 3 (04:59):
I think we do criticized in New.
Speaker 1 (05:00):
Zealand about perceived lack of capability to use this technology
to best effect. So training as part of that will
talk a little bit more about that. But Equinox has
really evolved through that whole transition a couple of major transitions,
the rise of apps, for instance, but that transition from
(05:23):
having your data and your applications on a server in
the basement of your building.
Speaker 3 (05:29):
To the move to the cloud.
Speaker 1 (05:31):
And we've heard so much about that over the last
couple of years. We've had Microsoft debut it's cloud region
in New Zealand, We've got a WUS supposedly on the way,
but lots of investments in Frattil investing CDC, so literally
probably billions of dollars or one hundreds of millions at
least going into cloud. We've been told to be more productive,
(05:53):
we need to get to the cloud. What's your perspective
over the time you've been at Equinox, which is really
that period when it's Excel. What are some of the
trends that you've seen in terms of New Zealand companies
going on that journey.
Speaker 4 (06:05):
Yeah, Look, I think and definitely those that were first going,
and I used to say to clients, it's not really
so much that that much had been released to the cloud,
but a lot had escaped. So there was a lot
that wasn't really done with a lot of strategic thinking
in place.
Speaker 2 (06:18):
But you know that they got in and.
Speaker 4 (06:20):
They started using and they started doing, which was great.
Most organizations we talk to now are in the cloud,
but they are suffering some of those problems at the
fact that they know they kind of evolved into it
rather than necessarily having it being initially a strategic objective,
and now that it is strategic, there's a little bit
of unpicking that needs to get done. And so we
(06:40):
often come in and find organizations needing to kind of
reimagine what they're doing in the cloud and set it
up better for them for now, but also for the
future as well, and take away some of that you know,
tech jargon of technical debt, some of the things that
we did that maybe were you okay at the time,
but now we look at them and say, actually, that
wasn't our best work. We need to reimagine that and
(07:01):
redeploy it for something that's going to last us for
the next three to five years. Obviously, not much in
tech lasts three to five years, but those foundational layers
and a real understanding about what an organization is trying
to achieve, those do stay a lot more static, and
so having that stuff really understood and the foundations set
up for what the business is and should be famous
(07:23):
for is really important.
Speaker 1 (07:24):
Technically, it's really important obviously what cloud platforms you move
to have the capacity to host that apps. Your apps
might have to be adapted or rewritten for cloud environments,
so there's that aspect of it, and the so called
DevOps is.
Speaker 3 (07:43):
Central to that.
Speaker 1 (07:44):
At the same time, we've got this transition, and I
know talking to people in government departments, this is really
important for them.
Speaker 3 (07:50):
From a budget point of view.
Speaker 1 (07:51):
You're going from a sort of a capital expenditure environment
where you're spending a lot of money on hardware that
is transitioning to more of an operational experidence your model
where you're paying literally subscriptions or contracts to aws, Microsoft,
Google and others every month. So it's a real challenge
(08:14):
to get your head around it.
Speaker 3 (08:15):
I guess for.
Speaker 1 (08:17):
People who are trying to predict this is how much
capacity we're going to need to serve new Zealand citizens
or business customers. How much is this actually going to cost?
Speaker 4 (08:27):
Absolutely, And I think there's massive power in that change
to an operational expense rather than the need to buy hardware, right.
Speaker 2 (08:33):
I mean, in the bad old days, you'd go, oh.
Speaker 4 (08:36):
We need to deploy one new business application that's now
pushed us above the capacity of what we've got our
data center. Now we need to buy and you serve,
and then we need to buy new storage. You're talking
about stuff being shipped over in a worst case scenario
and you've got maybe a two month lag before you
can put one new thing in, whereas now the cloud,
of course is the exact opposite. You want something, you
(08:57):
can spin it up in minutes, so there's a huge
amount of fleet stability to it. But of course you know,
you can depreciate that hardware and you can treat it
the way you're used to treating assets. And a lot
of organizations that are set up really capital heavy in
the way they do their financing, it's a shift that
needs to get made in their thinking to be able
to really get the power out of cloud. And I think,
(09:17):
as you kind of mentioned, there was a lot of
stuff in their own data center. You know, the great
thing about KIWIS is our number eight wire mentality. Sometimes
the risk about QWIS is our number eight wire mentality,
where you have business critical systems running on a PC
under someone's desk, or you've got, you know, some assets
that are really being sweated a lot longer than they
(09:39):
should be, a lot longer than their manufacturers would recommend.
And that stuff, of course always seems like a great
idea until something goes wrong, and I think there's that
insurance policy that you can get by having that stuff,
having that kind of raw infrastructure looked after by the
cloud providers, and obviously you know, you do have other
options with data center hosting and that sort of thing
(09:59):
as well. They generally don't give you, you know, generally
what I call weapons grades, security and privacy and scalability,
which you do get out of the hypervisors like your
Microsoft's and AWS and Googles look by.
Speaker 1 (10:11):
And which I think the track record has been pretty
solid in New Zealand in terms of reliability off the cloud.
You know, they're configured in ways we do have natural
disasters in our country, but they typically have three points
of presents at least in different regions of a city
or sometimes in Hamilton and Auckland, for instance. So the
(10:32):
model seems to have proven itself. What I am hearing
though from businesses, I'm really keen to get your insights
into this is a bit of alarm. When companies do
get to the cloud at the ongoing costs are often
higher than they anticipated, so called bill shock. And this
is not a New Zealand saying there are companies all
over the world that we've heard stories about that are
(10:54):
literally paying tens of millions of dollars in cloud costs
and freaking out about it. Then, of course we're going
to talk specifically about artificial intelligence, where you have these
data centers full of GPUs that are doing really sophisticated
AI staff. But you know the capacity required and the
resources are significant, so you've got to pay for that.
(11:17):
But let's talk us through about what you've seen in
terms of customers. Maybe you having to come into help
customers optimize that spend and in some cases sort of
rescue them from a situation.
Speaker 3 (11:30):
That's really bad.
Speaker 2 (11:31):
Yeah.
Speaker 4 (11:31):
Look, obviously, I mean there are some high profile cases
of cloud costs really going feral.
Speaker 2 (11:36):
We haven't experienced any of those ourselves.
Speaker 4 (11:38):
And our clients haven't, but there has definitely been a
lot of much higher spend than was expected, and I think,
you know, the recent economic situation has caused a lot
of organizations to actually take a look at their spending
quite closely, and often it's been the first time that
a huge amount of detail has been applied to cloud
costs because they've realized that actually, some of the cloud
(12:00):
pricing models.
Speaker 2 (12:01):
Appear very opaque, they appear.
Speaker 4 (12:03):
Very confusing, and it can be very difficult to kind
of unpick them and figure out what is costing you
the money in the cloud environment if it's not set
up well to begin with, which kind of takes us
back to that point I made before about things evolving
over time without the appropriate guardrails and controls in place,
without the appropriate visibility means that you end up with
(12:25):
a bit of a spaghetti mess to work through. So
when you kind of see those problems at the headline
level of cost and you know you need to manage
that cost down and you only just realize that you've
got a cost problem from the headline, that's when things
do get pretty challenging, and organizations like ours, you know,
come in and help our clients to identify and their
(12:46):
manage that cost better. But you know what we much
prefer to do as be involved at the start of
a thing or be involved in helping an organization to
really look at how they could be using cloud better
to get business value and have those conversations about what
will be coming in the coming months and years and
set them up well with everything. And obviously a big
(13:08):
part of that is having appropriate visibility and control around
things like cost.
Speaker 1 (13:13):
Yeah, and you know from looking at what Equinox does
talked about the training, it is really about about people
and this is.
Speaker 3 (13:21):
Where it projects usually go wrong.
Speaker 1 (13:23):
The technology might be world class and have the ability
to do what you want to do, whether it's you know,
selling tickets on a New Zealand or a banking system
or whatever, it's the people within the organization and often
that where it falls down is the culture and it
not gelling with the business or the actual business culture
the people who are responsible for driving the business forward.
Speaker 3 (13:46):
What are some of the red.
Speaker 1 (13:47):
Flags you see when you go into an organization that
might be struggling a little bit with its cloud spend.
Some of the things you see we go oh that
instantly needs addressing that's where the core of your problem is.
Speaker 4 (14:00):
Look, it really is interesting how much an organization culture
can permeate into the way that systems get built, managed
and operated.
Speaker 2 (14:08):
So anytime that you do see teams within.
Speaker 4 (14:11):
The IT team not getting on very well, as you
say that that very traditional disconnect between the IT and
the business, those will usually play out in an IT
environment that is not set up the way it's best
for the business.
Speaker 2 (14:24):
You end up with.
Speaker 4 (14:27):
A lot of people within the area of control optimizing
for their own team or optimizing for their own business unit,
their own budget. Whereas really because of the fact that
a lot of what you have in cloud covers everything
you do, but then you've got little bits that you
need to be specialist for certain areas, you need to
have that approach of optimizing for the whole and that's
how you get by far the best value out of
(14:48):
the cloud and you manage to keep your overall costs down.
The other interesting thing in IT is I guess looking
at cost and cost right, and so total cost of
ownership being what you genuinely influenced with cloud headline cost
not necessarily and in fact, if you're comparing your server
that you've got under your desk to a serve you
have hosted in weapons great infrastructure in Azure or Google
(15:12):
or a US.
Speaker 2 (15:13):
Obviously you're not going to get.
Speaker 4 (15:16):
The same cost comparison there, But what you are going
to get is insurance from risk, and you are going
to get reliability and resilience and inability to recover from
a failure. So you have to know how to attach
a value to those things to actually get that view
of that total cost. There's a great little analogy online
of pizza as a service that kind of talks about
the different cloud models and likens them to a pizza
(15:38):
delivery business.
Speaker 2 (15:39):
At one end, you've.
Speaker 4 (15:41):
Got someone making all the pizza and all the ingredients
and putting it together for you and cooking it and
delivering it and putting it on your table. And at
the other end, you're making your own dough and you're
putting all the ingredients on, you're making your own pizza.
And I think it's really useful to kind of think
of that in terms of that layering that actually comes
into the cost for managing it, and unless you are
really clear about how you strip out some of that cost,
(16:03):
and usually that's just about putting an ento higher, higher
value tasks. That's when you can start doing a clear
comparison between cloud and on prem.
Speaker 1 (16:12):
A lot of this is about human behavior and human nature.
I guess the trickiest thing is when you go into
an organization and you just see a dysfunctional relationship there
between maybe the CIO or the IT team and what
the product owners are, the people who are driving the
business forward bringing.
Speaker 3 (16:31):
In the revenue.
Speaker 1 (16:33):
So do you sometimes literally sort of have to almost
be like therapists to them, get them all in a
room and sort of say, look, you all need to
get on the same team.
Speaker 3 (16:41):
How do you approach that?
Speaker 2 (16:42):
Yeah, very much.
Speaker 4 (16:43):
Sometimes it feels like we've got the customers on the
couch when we're talking them through things, and look and
we understand both sides of that equation. So we can
be genuinely empathetic with the situation that people are. And
it's one of those usual things, right. People don't come
to work think they're going to do a bad job.
People come to work with the confines of the role
that they are in, wanting to do the best that
they can do. And so we've got some of these
(17:05):
tensions or conflicts. Is not ever really down to the people.
It's down to the position that they are in and
maybe a lack of understanding about how they could do
things better together.
Speaker 2 (17:14):
But some of those traditional.
Speaker 4 (17:15):
Dysfunctions are You've got a product team or project teams
going and creating a whole bunch of things, and they
don't have within their remits, sometimes not even within their
business case, what the genuine long term cost to operate
the thing is going to be, and so they don't
worry about that because that's not something that they're tasked with.
There's not a KPI around them understanding that stuff, and
(17:36):
so they throw it over the fence to the operations teams,
and their role is purely about keeping things reliable and
performant and ready when people need them. And they've now
effectively adopted a large amount of costs that they wouldn't
assigned up for if they.
Speaker 2 (17:50):
Had the choice.
Speaker 4 (17:51):
So neither of those two functions has done the wrong thing.
One has been tasked with driving business value, getting more customers,
making efficiency done it and the business elements. The other
team is tasked with keeping the costs down, keeping things
reliable and going, and they're doing that as best they can,
and it's just about making those two groups of people
able to have a good conversation and make sure that
(18:14):
what is delivered as something that achieves both of those objectives.
Speaker 1 (18:16):
And I think we've got to a point with the
cloud now where you can estimate, you know, if you
actually understand where your business is going and how it's growing,
or how new products are going to impact on your
digital footprint and all that sort of thing, you can
sort of you can estimate your costs over the coming
years relatively accurately, and there are volume discounts and things
(18:39):
to minimize your cost the vendors are offering that are
they good enough at giving you enough visibility into your
costs and giving you enough understanding? Are there enough tools
and dashboards and methods of actually understanding what you're paying?
Speaker 4 (18:56):
A couple of answers that I guess one. I mean,
the tools are great. The problem with all these things
is not the technology. There are tools that will do
it if you use them well and if you've got
your cloud set up well to use them right. And
again that's that technical debt thing where it can be
quite hard for a lot of organizations to see why
they should invest in going through and retrofitting some of
the better practices and better visibility metrics and that sort
(19:18):
of thing into some of their projects that they completed
and they'll put down and the team that used to
work on those isn't there anymore. It's quite hard to
get people to refocus on the old stuff because it's
not the exciting, new, sexy things and need to get
out the door. So there's that element to it in
terms of the tools are there but not necessarily being used.
(19:40):
But the other element to it is just those tools
and just the accountability for cost.
Speaker 2 (19:47):
And the even.
Speaker 4 (19:50):
The responsibility and desire to look at it often sits
within the team and the I T team only, and
so that information that they have isn't perm added to
the rest of the business and they are only able
to make the decisions that they can within their kind
of scope and knowledge, if that greater level of visibility
around cost and performance and what resources are there, how
(20:12):
they're being used, and ultimately what the conversion rate is
between some services and a cloud provider and a number
of new clients to get in the door, or the
increase in cross sell and upsell because of an ability
to get a new and exciting service out to the
customer until you can kind of draw a line in
your budget between the business value metrics and the baseline
(20:33):
costs of the infrastructure and the people and the partners
that you need to support that infrastructure, which again needs
to be optimized and made best for cloud. Until you
can get that kind of steel thread, it's really really
hard to optimize for value.
Speaker 2 (20:46):
So the tools are there, it's the cultures.
Speaker 4 (20:49):
That aren't quite there at the moment. It's the processes
that aren't there in the guardrails. Everything from the light
bulb going off for someone imagining an idea to the
end of the life cycle of of.
Speaker 2 (21:00):
An IT system.
Speaker 1 (21:01):
And you know, this is really sort of complex, especially
if you're a business. We've got a lot of digital
first businesses and ual software businesses like zero that have
done really well at exporting internationally. So for them, big
chunk of their cost is literally development as well as
hosting millions of transactions.
Speaker 3 (21:23):
Every day across their servers. So this is really important.
Speaker 1 (21:26):
It's actually led to the rise of a whole discipline,
the phinnops discipline, which you guys are very active in
as well.
Speaker 3 (21:33):
Talk us through how that works.
Speaker 4 (21:34):
And it's interesting you bring up the case of the
software created because I think a lot of the tools
and a lot of the practices that you know the
right thing to do if all of your business and
therefore a large percentage of your cost is that it infrastructure.
Obviously it gets a good amount of attention, but for
businesses where the main thing they do is retail, it
just so happens. They need IT systems to make those happen,
(21:56):
it's a lot harder to I guess, get the priority
around some of these things. And look, and that's WHEREOPS
or cloud cost optimization really does come in, and it
is about creating a group of people who are able
to collaborate on maximizing business value out of technology. Actually
(22:16):
it is not just cloud's kind of in some ways
a bit easier. You've got some more levers you can
pull to make it so that your costs better value
the better match the value that you get. But you
can still do that stuff on traditional tools as well.
So the approach TOOPS is very much one that, yes,
there's some technology enablement that needs to get done. There
(22:39):
are some tools that you can put in place, there
are some different technical practices that you could use, but
the line share of it is actually around the business process,
around business casing, around forecasting, and about turning some of
those again what look to be maybe over complex pricing
models compared to the old way work for you rather
(23:01):
than work against you.
Speaker 2 (23:02):
So I think there's a huge.
Speaker 4 (23:04):
Amount of power that can be gained from a cost
benefit point of view if you embrace the way cloud works.
Speaker 2 (23:11):
But that means a lot of change.
Speaker 4 (23:13):
That means change for the people, change for the skills
that people have, and changed the processes that the organizations have.
Speaker 1 (23:19):
I've interviewed some companies that have done incredibly well. I'm
thinking of a company like Volpower Health recently sold to
a Korean medical company for three hundred million dollars. Founded
in Wellington, Ralph heinem who we had on the podcast
about a month ago.
Speaker 3 (23:36):
Incredibly successful, very.
Speaker 1 (23:37):
Much built on as you're on the Microsoft platform and
as a result of that was able to leverage all
the tools, the AI sort of tools. So there is
incredible power, as you say, if you really understand and
sort of a cloud centric organization, it can really pay off.
I mentioned AI. So that's the next phase. We're going
(23:58):
through a lot of hype around it. Out of social intelligence,
the right of AI agents, and again, you know a
lot of companies are going, wow, okay, I want to
spin up agents and allow my customers to use agents,
and there are platforms to make that easy to do.
Speaker 3 (24:15):
But there's a lot of consumption based business models emerging.
Speaker 1 (24:19):
And for New Zealand companies that are trying to build
and use their own agents, I guess getting their head
around what are the ongoing implications costs wise of that
is a real challenge.
Speaker 4 (24:31):
Yeah, absolutely, and it is technically complex the way things
are priced at the moment. Some are comparatively simple. But
the second you want to get more than you kind
of get from the consumer and consumer like versions of
the AI tools, the second you want to start pumping
your business darter into them and getting some genuine insights
that you can use, and you want to start using
(24:51):
AI to optimize workflows that you have and automate those.
That's when you start getting into the territory of some
of those reasons you need to deploy that are going
to you.
Speaker 2 (25:02):
Know, potentially massively expand without letting you know first.
Speaker 4 (25:05):
There are some services that I would argue probably should
be opt in rather than opt out, but by default
they turn on because the vendors think that you obviously
want all the AI power you could possibly ever get, therefore,
best make it easy to consume. But I think the
other problem that we are seeing because of the fact
there is such a groundswell of interest around AI, is
(25:28):
that a lot of people are going into it and
feeling almost a bit of desperation to be doing it
and to be seen to be doing it, and maybe
not putting in some of the business discipline that you
can still do even with this fast moving stuff. So,
you know, we like to think of having a living
business case, so you don't need to go all the
way down the path of completing a business case before
(25:50):
you do anything. You don't need to analyze everything before
you start doing anything. But what you should be doing
is factoring in what you believe the tangible business value
you're going to get out of your experiment is from
the get go, and it's part of the experiment is learning, oh, no,
our costs are going to be slightly higher than we
thought they were going to be, or actually the value
that we thought we were going to get isn't But
now that I look at it, we're going to get.
Speaker 2 (26:10):
A different piece of value.
Speaker 4 (26:12):
So until you do tie the genuine business outcomes into
the business outcomes and the business costs into what you're
doing within AI. There are definitely risks and costs going
well north of what a good business decision would indicate
they should.
Speaker 3 (26:26):
If you're in a team or an individual in a mid.
Speaker 1 (26:29):
Size or organization that's responsible for finops and for Equinox's
team as well, increasingly, you've got to have pretty good
AI expertise in those teams to actually understand what the
cost implications of the company rolling out AI actually are.
Speaker 4 (26:44):
Yeah, definitely, and looking there are some traditional sort of
high risk workloads for finops. Those are the ones that
we do focus on, you know, AI obviously, and given
that it's emergent and the pricing models are emergent and
they're changing regularly, it's to read a key ones to
make sure that is understood.
Speaker 2 (27:02):
And big data is.
Speaker 4 (27:03):
The other one of the other major workloads that would
often you know, as you allow things to get into
the hands of citizen developers and people can start creating
their own data sources or merging data sources and then
extrapolating from that to get more value, it can be
quite hard to kind of put constraints on those things.
When the old world again, you'd have a sequel server
(27:23):
that the IT team would look after.
Speaker 2 (27:25):
If that started hitting capacity.
Speaker 4 (27:26):
Well you slow down and you can't do things today
of course without.
Speaker 2 (27:32):
Some controls in place.
Speaker 4 (27:34):
You know, the cloud based big data solutions will scale
infinitely because you know, they want you to buy more
of their products, but also because they assume that what
you're doing is valuable. So it is the onus is
on the organization to put those limits in place on
you know, the maximum expected value that's going to be
gotten from something and not just allow endless consumption.
Speaker 1 (27:53):
Yeah, and one thing I do hear a lot oft
and I think this speaks to the fact that we've
been a little bit behind in getting our data house
in order. In New Zealand is where the big bills
rack up is constantly shuttling data between different silos because
you need to bring all of that data in free use,
and you haven't done the groundwork to put all that
(28:14):
place in one area. You know, we've heard of data
lakes and lake houses and all these terms to describe that.
But there's been this massive push in the industry get
your data in one place so we can query that
and you're not having to shell it around and rack
up cloud builds.
Speaker 4 (28:30):
Absolutely, and that was one of the earliest kind of
bill shock instances really was when organizations read the fine
print and realize that you get charged to take data
out of the cloud platforms. So the second you kind
of half pregnant in cloud, that's kind of where your
cost problems really kick in because you've got some stuff
on premise, you move some things into cloud, but you
never move enough into the cloud that you can turn
(28:51):
off what you've got on site, So you're actually double
paying fear infrastructure at that point, in fact sometimes more
because you go, well, we need to be secure, so
we'll put security things both on site and also in
the cloud. And oh, now actually we've got these two
systems need to talk to each other. We're better make
our network more expensive to cater for it, and then
of course more expensive times ten because the cloud vendors
(29:12):
are charging you for the way out as well. So
definitely that half pregnant approach is really problematic, and AI
is only adding to that. And I think where we see,
you know, I think you need to be careful about
where you experiment on AI.
Speaker 2 (29:27):
So AI is going to.
Speaker 4 (29:29):
Be pervasive across every piece of software that you use today. Therefore,
if the vendor of that software is going to be
building their own AI capability, should look at whether you're
better to wait for them to do it. Because they
will build it, they will package into they're pricing, there
won't be those unknowns. Rather than going, oh, we need
to build something that's completely unique to us, and that
(29:50):
involves extracting a whole bunch of data, standing up a
whole bunch of compute resource yourself, doing a lot that's
going to cost you a lot of money to then
in six months find that actually that you know, unique
thing that you built is actually now available as part
of the product. And we have seen some of those
examples or organizations have spun up something AI that really,
(30:10):
you know, it will be table stakes at some stage
in the coming.
Speaker 1 (30:13):
Months, particularly because you know, I sort of here, particularly
from vendors like IBM, are saying, oh, the future is
hybrid cloud.
Speaker 3 (30:22):
It seems logical to put everything with one vendor.
Speaker 1 (30:25):
You're going to you're going to get a better overall
price the more business you have with one vendor. But
it's not as simple as that there are some people
still have stuff on their premises. There are others who
will have multiple cloud instances. They might have something on
as you're on something on IBM's cloud because they want
to use Watson.
Speaker 3 (30:46):
Or something like that.
Speaker 1 (30:47):
So the reality is a little bit messier, I guess
is hybrid cloud really what is sort of has emerged
as a new norm and new Zealand.
Speaker 4 (30:54):
I mean hybrid cloud, especially if you include on premise
in that. Yes, that's the default for a lot of organism. Again,
there's maybe not in some cases not enough intentionality about
how it's been done. And I think there's a couple
of different things. Is taking the best out of a
cloud provider right and realizing the particular solution that you
(31:15):
need that's going to have a lot of business value
is best to live it out of aws where most
of the other stuff you've got is an azure that
to my mind, is a useful model. That's an approach
that works. Instead saying oh, we've got you know, we
need to have a data lake, Oh let's have another
data lake and another provider, or saying oh we need
(31:36):
to have our virtual machines in one place, it's have
a whole bunch of virtual machines in another cloud provider.
Speaker 2 (31:40):
That makes a lot less sense to me.
Speaker 1 (31:42):
And is it getting more competitive as more capacity comes
into the market and more players are offering the so
called hyperscale clouds. Are you seeing a good competitive tension
out there between all of these big multinationals a little?
Speaker 4 (31:56):
I mean, it's interesting in New Zealand because we've traditionally
been so heavily into the micro Soft ecosystem for everything,
and so when AWS came into New Zealand, they absolutely
you beat Microsoft to the punch around that timeframe, and
so the cloud volume for a WS increased very very quickly.
In Microsoft, I think it's still probably playing catch up.
So I think the interesting thing for New Zealand is
(32:19):
that now Microsoft has caught up. From a feature's point
of view, it kind of is the path of least
resistance for most organizations to move into Asia, so government right,
well exactly, yeah, yeah, And therefore it becomes you know,
niche positions are taken by other players. But I guess
you know, with organizations that are in a brown fields environment. Again,
(32:42):
it's kind of it's the difference between the volparers of
the world and an organization that's been running for you know,
twenty five years, and they've got a whole bunch of
legacy and data centers and that sort of thing. In
a brown field's environment, it's very very hard to go
pure into one thing or another. It's very hard for
organizations to turn off those last things that sitting on
their sites.
Speaker 2 (33:01):
It's very hard to.
Speaker 4 (33:02):
Say no, no, we're not going to pack aws for
something and Google for.
Speaker 2 (33:05):
Something else, or Microsoft for something else.
Speaker 4 (33:08):
But it's just about I think, being really clear where
some of those costs and complexity drivers.
Speaker 2 (33:13):
Are going to be.
Speaker 4 (33:13):
That mean that what looks like a ten percent better
solution functionality wise in a second cloud might be double
the cost compared to having it in something that an
environment you already have.
Speaker 1 (33:25):
Yeah, So what is your advice to those sorts of
brown fields organizations that they might be in the cloud
and becoming a bit more aligned at the ongoing cost
of what they're doing, or they may be on that
journey and about to do a migration. What are the
sort of things that you sit down and talk to
these sorts of organizations about what are the principles or
(33:46):
the fundamentals they need to consider.
Speaker 4 (33:48):
Well, look, amend does all start with a business case, right,
and it's going to be really very clear on what
the outcomes are that are trying to be achieved, and
it has to be a very conscious decision, and one
is made with a real understanding of the total cost,
not of the.
Speaker 2 (34:04):
Cost of servers.
Speaker 4 (34:06):
And if it's only going to be done with a
top line cost, then you know, cloud might not be
the right solution. It might not be able to be
justified for doing so. But for an organization that has
stuff that is in legacy land and can't be moved
into the cloud, I think you need to realize that
(34:26):
at some point you're going to have to make that change.
And it's kind of one of those you know, the
best time to do it was ten years.
Speaker 2 (34:32):
Ago, next best time is now.
Speaker 4 (34:34):
If you can't modernize your system so they can run
in today's infrastructure, then you are going to hold the
business back. It's going to be an anchor that is
going to be stopping the business in a whole bunch
of ways. Because the second everything is powered by AI
and whatever the next set of new technologies are, and
the next you know, exciting ways of engaging staff and
(34:55):
engaging customers if you can't take advantage of those, you
can lose competitive advantage anything that touches consumer. For example,
you know, before too long, if your services don't show
up on an AI search and AI query, you will
cease to exist. You will be invisible to everybody because
the number of eyeballs are going to be going to
websites to look at things is going to rapidly decline
(35:17):
is AI takes over and replaces search. So you know,
if you've got an old legacy system that you could
never plug meaningfully into AI, you can't take advantage of
cloud's ability to scale.
Speaker 2 (35:28):
You need to look at how you can.
Speaker 4 (35:30):
You know, even if it is a multi year program
to modernize and move away from that application, then that
should be absolutely on the table for consideration. And it's
just about looking at the things that you need to
be doing, looking at the things that are genuinely business
value generating invest in those, modernize those, get them to
(35:52):
where they need to be. And other things where you
know you've built something, or you're using something that maybe
could be done another way by assas products or something
that you could consume off the shelf.
Speaker 2 (36:01):
Any of those kinds of movies you can make the better.
Speaker 1 (36:03):
So it sounds like it's you know, it's not something
that's going to be pigeonholed in the IT department of
a New Zealand business. This is something that the CEO,
you know, the top executives need to have real clear
visibility into and understanding of.
Speaker 2 (36:18):
Absolutely.
Speaker 4 (36:20):
I mean, I think the days of a business strategy
being handed to the I T team to then form
an IT plan, you know, is very backward. I think
if you look at your traditional smack influences of social mobility,
analytics in cloud, if technology and technology innovation and disruption
isn't factored into business strategy, then you know, there aren't
(36:42):
very many industries where you know, that's a viable strategy anymore.
I think you need to be taken into account the
massive changes they're going to make for consumer buying habits
and B to B buying habits, and just the ability
to genuinely compete in whatever an organization does is going
to be increasing reliant on good use of technology, so
(37:02):
it has to be baked into every level of the
organization from the boad doown.
Speaker 1 (37:12):
That wraps up another episode of the Business of Tech
My thanks to David Reese from equinox It for sharing
practical wisdom on tackling cloud bill shock, including why the
hardest problems guess what are more often found in company
culture and processes than in the technology itself.
Speaker 3 (37:32):
Look.
Speaker 1 (37:32):
I've been writing about the cloud for years, and every
time I cover the tech company's claims of the benefits
of migrating to the cloud, which are legit, I usually
get one or two readers contacting me to say, Basically,
it may look that way, but the costs really do
escalate over time, and it's actually more cost effective in
many cases to maintain your own infrastructure, which is an
(37:56):
approach that's sort of gone out of fashion. I've heard
stories of companies leave the cloud due to costs. The
company thirty seven Signals is a very high profile case
internationally of that. I link to the CEO's blog post
on how the company moved onto its own hosted hardware
to cut its three million dollar a year cloud bill.
(38:17):
The cloud's true value can be unlocked, but only when
business leaders, it and finance teams work together with transparency,
shared goals, and the right guard rails long before the
invoices start to arrive. Whether you're scaling up AI or
just managing day to day cloud operations. A strategic and
(38:38):
collaborative approach will protect you from those nasty surprises and
help ensure your investment actually pays off. Equinox is doing
a lot of work in this area. I recently helped
them write a white paper on this whole topic of
cloud cost optimization. I'll link to it in the show
notes at Businessdesk dot co dot nz. You'll find them
there and my weekly tech reading list. Just go the
(39:00):
podcast section. Thanks for listening. If you enjoyed this episode,
don't forget to subscribe and share it with your colleagues
on iHeartRadio or wherever you get your podcasts. I'm Peter Griffin.
See you next time on the Business of Tech.
Popular Podcasts
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.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
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