Your Data Anywhere

Yellowbrick
For Crying Out Cloud: How to Avoid Poor ROI from Your Cloud Analytics

For Crying Out Cloud: How to Avoid Poor ROI from Your Cloud Analytics

Cloud transformation efforts

For many enterprises, accelerating cloud transformation efforts has become an essential part of operating budgets. As economic factors force businesses to rethink spending across all areas of operations, many are frustrated with cloud data warehouse consumption-based pricing after years of skyrocketing costs.

Research from the FinOps Foundation found that about 30% of respondents identified forecasting their cloud costs for a given initiative as the biggest challenge. How can enterprises find a better, more affordable cloud data warehouse solution without sacrificing performance?

In this LinkedIn Live discussion, Kevin Petrie, VP of Research at Eckerson Group joins Yellowbrick Data’s Mark Cusack, CTO, and Heather Brodbeck, VP RevOps, to discuss how enterprises can change the norm in cloud analytics spending and strategies to consider for managing costs in the cloud.

Transcript:

Heather Brodbeck:
All right, here we are. Thank you everyone for joining the LinkedIn Live today with us. We’ve got a great topic; a couple of great panelists here; hope to learn a lot from them today. So our topic is avoiding poor ROI from your cloud analytics. And so I will first introduce myself, Heather Brodbeck; I’m the VP of RevOps with Yellowbrick. And we have first Mark Cusack here. He’s our Chief Technology Officer at Yellowbrick. Mark, do you want to introduce yourself?

Mark Cusack:
Thanks very much, Heather, and nice to see you again. Yeah, so my name’s Mark Cusack. I’m the CTO at Yellowbrick. I’ve been in the data warehousing business probably for the last 20 years. At Yellowbrick now, as CTO; but previously at Teradata, and I’ve worked at startups and in government.

I’m very excited to talk about this topic, which I know is top of mind for many CFOs as they’re looking at their cloud budget in a challenging economic environment today, and with ESG concerns starting to become first and foremost in many enterprises’ minds.

Heather Brodbeck:
All right, thanks, Mark. And our second panelist here, we have Kevin Petrie from Eckerson. So would you like to introduce yourself?

Kevin Petrie:
Great, thank you, Heather and Mark. Great to join this conversation. I run the research group here at Eckerson Group. We’re a boutique research and consulting firm focused on data analytics, and cover a lot of topics related to cloud and data modernization. FinOps or cloud cost governance is a big part of that. So very pleased to have the opportunity to join this discussion.

Heather Brodbeck:
All right, thank you. All right. So yes, like I said before, today’s discussion, we’re going to talk around cloud investments; how organizations are migrating to cloud; how they’re managing those costs; talk about some pain points; and then have some guidance from these two on what businesses can do about that. So if there are questions along the way, you can put them into the chat. We’ll try to get those to those at the end. So yeah, let’s jump into it.

So initially, Kevin, we’ll start with you. Many enterprises accelerating the transformation efforts, cloud has become essential, big part of operating budgets. So let’s first discuss managing costs in the cloud. So how does the cost and budget process really evolve when moving to the cloud?

Kevin Petrie:
Well, Heather, it’s a great question. It’s a good way to frame the conversation. I think the quick analogy is that when folks move from their own data center to the cloud, they start renting resources rather than buying them. It gives a lot of financial flexibility, and it helps accommodate changing projects and so forth. But it can create risks in terms of unknowns about consumption-related costs.

And what we’ve seen on the consulting side is that as a lot of organizations adopt advanced analytics, as they democratize the consumption of data, they have a lot of moving parts. And so their budget outlay at the start of a quarter or the start of a year related to OpEx, which is what the cloud entails, can start to fluctuate because they could have a new very compelling data science project that comes in and requires all kinds of compute to train new machine learning models, and that starts to push costs above the budget.

Heather Brodbeck:
Okay, thanks. Mark, did you want to add anything to that?

Mark Cusack:
Yeah, sure. I think what’s interesting in this move from a CapEx picture to OpEx in the cloud, and I think we’re discovering now again in this more difficult economic climate there we’re discovering is, if you look at where we were in the past with CapEx spend on-prem, companies that already made a long-term investment. And so it’s effectively like a sunk cost when you’re looking at an on-prem multi-year data warehouse investment, for example.

So when times get more difficult economically, you can kind of ride through that CapEx spend because you’ve already made that investment, right? I think one of the challenges we’re beginning to see with conversations we’re having with CFOs today is with OpEx, you have an opportunity to decide how much of my spending cloud is actually discretionary versus mandatory.

And now we’re seeing, particularly amongst many vendors that rely purely on an on-demand consumption-based model, they don’t have the kind of runway that a CapEx or subscription model would buy them. And we are seeing enterprises really start to think, well, where can we tighten our belts around the discretionary OpEx kind of spend we’re making in the cloud.

Heather Brodbeck:
Okay, great. Thanks. And so to add onto that then, is this really considered a big problem when we think about data warehousing and analytics? I know you just highlighted some of the different facets of it, but maybe give us a sense of how pervasive that challenge is.

Kevin Petrie:
Sure, and I could chime in there. I would love to hear Mark thoughts too, but we definitely see it coming up with enterprises. I put out a few metrics in terms of an unscientific poll on LinkedIn. I asked data analytics leaders: Do you see the cloud costs limiting the value you can deliver with analytics? And 48% said yes. So there are a lot of organizations that really do feel they are constrained because of these surprises on the consumption side.

So it is a challenge, and what we see is that it’s giving rise to a real focus on governance for cloud data warehousing, where if you can put a real program in place to help business and IT stakeholders look across the life cycle of how they’re consuming applications, storage, compute network, and how they’re tying up people’s time, then they can start to make smarter decisions at each point in the life cycle. So we see a number of organizations acknowledging the pain and starting to get more serious about the culture, the level of collaboration and the accountability that’s required for cloud costs.

Heather Brodbeck:
Okay. Mark, anything to add on?

Mark Cusack:
That? Yeah, I think that does make sense, taking a closer iterative look to how you’re controlling spend. But what I have observed actually in some cases is the magnitude of the cloud spend isn’t necessarily such a problem, providing businesses are getting the return on investment out of it. The cloud’s providing the flexibility to address brand-new use cases that you probably wouldn’t look at before. Discovery, exploratory use cases. And so as long as you’re getting that return, perhaps that spend magnitude isn’t a problem.

And again from what is interesting, where on the flip side where it can be a problem, is to some extent from some of those always on workloads when they move from on-prem to the cloud. Because if you go back in the dates when everyone’s running on-prem data warehousing, they’re tuning their on-prem data warehouses to absolutely max out the capacity they have in their data center.

They need to do that to drive the cost per query down. Now when they get to the cloud, what I’m seeing in the data warehousing and analytics industry, in general, is that a lot of vendors in this space have forgotten that message of efficiency that we drove on-prem. And so the way that you kind of get the scale and the capacity that you need to drive your use cases in the cloud is expanding, is scaling out and consuming more and more and more.

So I think in this business in general, in the analytics business, in the cloud, we are forgetting around that needs to be coupled with driving efficiency savings as well. So I’m kind of contradicting myself in a way. In some ways, the cloud spend and magnitude isn’t a problem, but I’m seeing that some of the vendors in this space aren’t really helping the situation.

Kevin Petrie:
It’s interesting, there was some research, the FinOps Foundation has done some interesting research on this discipline of FinOps, which I view as really the same as cloud cost governance. They found that about 30% of respondents said the biggest challenge is forecasting their cloud costs for a given initiative. And then an equal amount, about 30% said the next biggest challenge is getting engineers to act on that information to start to factor and cause what they do. So I think that there’s a people and a process and a technology problem here.

Mark Cusack:
Yeah, that makes sense. And I would add actually just a quick one again. Leaving aside the magnitude of the cloud spend, that unpredictability starts to become a real problem. I’ve had lots of conversations around that about predicting your spend the next month.

Heather Brodbeck:
So what are maybe some best practices around that then? So when you have those conversations and you’re hearing that feedback, what just maybe some… I mean we’ve all been in those discussions. Like you said, Mark, with on-prem, obviously there’s cost management and everything, but it’s much more in your control. So maybe what are some differences and best practices around cost management for cloud?

Mark Cusack:
I can chime in. First of all, particularly from a technical basis, for example. So obviously when you’re moving your analytics workflows to the cloud, you start to think about the compute and storage requirements. I think today a lot of the focus is on the spend against these two particular line items. How much is the compute that I’m going to throw at my analytics going to cost, or how much is the storage overhead going to be?

But there are a few things that can actually be real gotchas at the end of the month, and you have to start to factor in things like cloud egress charges, the amount it’s going to cost to move data from one cloud to another, or within one cloud service provider from one region to another, and that can actually mount up.

One of those other kind of hidden parts of your cloud bill at the end of every month is also things like API spend. Many folks don’t really understand that when you’re making a lot of cloud service provider API calls against certain services, you are racking up these background of API spend. I’ve seen some bills where 20% of the monthly bill can be accounted for in these almost below the service API calls. So I think you just really need to understand exactly what you are spending and trace it all the way back to the business value that you want to address, and you really have to get into the detail to really understand some of this.

Kevin Petrie:
Yeah, Mark, I think that’s very well put, and agree that there are a lot of facets to this. From a cloud resource perspective, consumption of applications, storage, network and compute, there are a lot of tools that organizations can use to figure out how to spend those precious compute cycles in particular efficiently. They can use observability tools to optimize, to schedule workloads at the right time. They can use forecasting tools may be offered by a cloud platform or others to help figure out how they can meet their SLAs and avoid unnecessary spending.

But there’s also the simple reality that people cost a lot of money, and so you want to make use of people’s time, and that’s where getting back to the fundamentals really helps. So many folks that we work with are still focused on basic data consolidation, basic data governance practices.

We talked to a grocery chain based in the United States, and they said that moving to the cloud actually made some of their silos worse, because they didn’t do anything from a process perspective to figure out how to handle it. So there are a lot of things that if you get back to basic fundamentals, you can handle and operate more efficiently in a cloud environment.

Mark Cusack:
I think that’s a really good point about the staffing side of things as well. And I think there’s a skill gap you have to account for as well when you’re breaking down these costs of moving to the cloud. Because often, many, especially larger enterprises, they don’t necessarily move everything to the cloud in one go and burn the ships from their on-prem environments.

So they end up with two technical stacks in place; one in their own data center, and then a hybrid arrangement in the cloud. And so you end up needing to rescale and train staff to operate two different technology stacks as well. And so there are additional overheads that you have to consider when you are planning that kind of migration of your workloads from on-prem into the cloud.

Kevin Petrie:
Definitely agree.

Heather Brodbeck:
So maybe to add onto that a little bit. So Kevin, you mentioned an example of a retailer that had challenges; any other examples or maybe drivers of any experiences you all have had with customers around challenges they’ve had with cost and maybe what the drivers are? I know we’ve talked obviously about different pricing models, around consumption-based versus subscription, things like that. So any examples you guys have that would give more insight to some of the drivers.

Kevin Petrie:
We spoke with a firm that deals with rental cars, and nationally well-known chains – or global, really. They have a set of data, that they’ve held various functions. Marketing, sales, operations make use of the data for various use cases; and they do have a problem with runaway queries where folks come in not having been educated about the implications of the queries that they’re running, and that can balloon costs.

We spoke with another – this is a retailer – that was going to use the ability to throttle runaway costs and inform users about the compute cost implications of their decisions. They’ve used that as a primary evaluation criteria for various cloud analytics tools. So those are two cases where cost is sort of looming large in the decisions that organizations want to make.

Mark Cusack:
Yeah, I’m having an increasing number of conversations where cloud spend is a real concern. And again, I always come back down to data warehousing. I’m sorry, that’s what I do. These are the kind of conversations I get involved with across a range of different verticals.

I know that one business who’s in the utilities space was telling me about the sticker shock they experienced when they were doing a cloud data warehouse proof of concept. They go through the proof of concept, they think they understand the total cost of ownership of this thing; but when it gets into production, they find to reach the level of performance and meet their quality of service they wanted, they were throwing more and more capacity at this thing. And so it was just blowing their ROI calculations that they put together early on out of the water because they couldn’t meet their need.

I know this one particular head of architecture was telling me that it was going to be a very, very difficult conversation with the CFO around what the differences were between expectation and reality there. I could go on and on, but another company in the insurance space was telling me, a CTO there was telling me he was evaluating another cloud data warehouse; and in fact, they came to talk to us because they weren’t going to go back to this previous one before because it almost bankrupted their business from the spend they had with this previous. So just two examples of where particularly now people are getting very cost-focused and cost-conscious on their spend.

Heather Brodbeck:
Great. Like you said, we could probably go on and on. Kevin, you look like you have eight more examples in your head that you could talk through.

But let’s move on. So we’ve heard a lot about the pain points and drivers and all of that. So I think probably the audience on the call’s looking to understand some thoughts that you all have, or some things you’ve seen of how do we actually get control of that? Partially, what’s the remedy for getting more accurate forecast in place for spend?

So maybe, Kevin, first to you, so do companies have to sacrifice price for performance? How do we get cost certainty around all of this and also still be able to reap the benefits of the cloud?

Kevin Petrie:
Yeah, I think there are a number of techniques that organizations can use, both technically and from a people and process perspective. One is to use observability or other types of tools to look across their environments. They might find a bunch of spare old servers on premise that can handle a predictable workload adequately, meet performance SLAs, and so you don’t always need to jump to the cloud. So that’s one factor.

I think another is that the more you use cycles effectively, it improves both price and performance; and there are a lot of cases in which organizations are coming in willy-nilly because they’re democratized data users that don’t understand the implications of what they want to do, and so they spin up sandboxes and do various things. They might be duplicating other people’s work, they might be doing it at peak hours when there are higher charges. And so education becomes a really big component to help people understand this is the likely cost implication of what you’re doing, and we’re going to hold you accountable for that. And that’s a lot of what the cloud cost governance and ops gets to.

Mark Cusack:
Okay. Now I think that makes a lot of sense, and I think there’s actually an added dimension around this that you need to add into it as well. Not only do you need the kind of monitoring and forecasting around your spend in place as well, but you’ve got to look at ultimately what you’re spending your money on to deliver your analytics here. You’ve got to look at the platform and the efficiency of that thing as well.

And again, going to something I said earlier on, I think the data warehousing space for the last 30 years really focused on driving massive efficiency and throughput in a fixed footprint, and I think a lot of the data warehousing space in the cloud is completely forgotten of that idea. You can expand that beyond just data warehousing, frankly, into other areas of analytics as well.

Ideas like workload management, which are critical for on-prem systems, seem to be widely forgotten in cloud data warehousing today. And so, one of the things we focused on is taking all of the great ideas of efficiency that you drive from on-prem, and then marrying that with elasticity and separated compute and storage in the cloud.

And so at Yellowbrick, we think we found a way of doing that, where you can get a lot of the efficiency that you see on-prem delivered in the cloud. So I think observability management and monitoring is one thing, but just you need a way of running this stuff at a lot lower price point to start with too.

Heather Brodbeck:
You have more, Kevin?

Kevin Petrie:
I do. I was going to add that as we head into this uncertain 2023 economically, a lot of organizations I think are still using analytics and using it as a way to understand their environments better, but they’re probably going to be investing less in the more speculative, maybe more advanced analytics-oriented use cases, investing more in sort of bread and butter, blocking and tackling analytics projects. Those ones are going to have more predictable workloads and they can help you, I think, to build on Mark’s point, reduce some of the uncertainty about what you’re committing to.

Heather Brodbeck:
Okay, great. Thanks. So that’s a lot of information. Maybe you can make it a little more, I don’t know, I guess tangible for us. What are some of the steps that businesses or tech leads can take today to start to get more control around their costs, and what those forecasts are? And maybe we can throw back up that slide, Kevin, that you had up previously, and maybe do a little bit more deep dive in some of these steps so that our audience can take away some tangible things that they maybe can start putting into action now.

Kevin Petrie:
Okay. Yeah, I guess the first thing I’d point out is that we’re talking about multiple stake shelters here on the business side and on the IT side. So on the business side, you can have executive sponsors, business owners, data analysts, data scientists; and on the IT side, you’re going to have data engineers, cloud ops engineers, who are doing DevOps and IT ops on the cloud, and also data warehouse administrators. All these folks need to come together in a more rigorous FinOps or cloud cost governance program.

First, it involves forecasting what your costs are, let’s say over a given period of time, a month, more likely a year; and figuring out the… It’ll start with the business. What’s the business requirements and use cases? They’ll huddle with cloud ops engineers, data engineers to understand the likely implications of that.

And then as you implement a project, start to automate it or start to operate it, you want to make sure you’re monitoring what the KPIs are. There are a lot of data observability products that are increasingly focused on monitoring cost. So FinOps is becoming more of a focus, and that’s going to assist chargebacks so that you’re holding individual parties within the business accountable for the cost of what they’re doing. The key pieces of infrastructure, as I mentioned, that you’re consuming are cloud-oriented applications, storage, compute and network.

And through looking at this as an iterative life cycle, you can really start to build a culture of cost governance. We’ve talked a lot about data literacy in other circles. That’s very important as you get people more data-oriented and data-driven with their decision making, you want to make them equally aware of the cost implications of what they’re doing. You want to make sure that you have a common platform to collaborate and also hold folks accountable.

As I said, I think there are specific things that can be done too, and I know Mark can elaborate on this too; which is looking at: Do you really want an all consumption-based model? Do you want to reserve some resources for those predictable workloads? Those can be ways to reduce some of the variables in this lifecycle.

Heather Brodbeck:
Great, thanks Mark. Can you add to that?

Mark Cusack:
I talked a little bit earlier about workload management being really important when you’re trying to get the most capacity and most throughput out of any analytical system. But I think what is becoming more important around this kind of cost governance side of things is workload analytics. It’s actually analytics on analytics, effectively; trying to understand how you are using your spend and how you are delivering value out of that. And there are some simple things that you want to be able to do.

If you want to look at across your workloads, across your business and understand, well, what’s effectively a fixed capacity workload here? Are there workloads here that are running 24/7, they’re run the business round the globe kind of workloads that are well-characterized in terms of the compute and storage that they use, for example.

And what are those variable workloads? Perhaps they’re ones that are come up at the end of each month for end-of-month closing, or end-of-quarter closing operations, or coming up obviously towards Black Friday. There’s obviously a big spike in cloud compute consumption at this time of year as well. But given that picture of workload, understanding what’s better off on subscription-based pricing where you’re going to operate those at a discount versus on-demand pricing, which is typically much, much higher in terms of compute per second costs here. So you need a degree of sophistication route to really truly understanding your analytics and workloads.

Heather Brodbeck:
Great. Maybe just one more, I guess, question for me or something to expand on a little bit. I know we’ve talked about when you were talking about cloud governance and that framework that you showed, and you’re referencing the FinOps team; that’s something I think you have somewhat new around how cloud cost is managed. I think maybe you could expand a little bit on how that group plays a critical role in the management of cost.

I think for my learnings, they see more horizontally across a given organization; and part of their job is to actually lower cost overall, while working with the teams that have all these requirements that are trying to get things done that they’ve been mandated to do. So how does that FinOps team maybe play a critical role in cost management?

Because I’m sure some of the folks on the line that’s not really a group they maybe have today, or maybe they kind of do and they don’t call it that. But as I’ve been thinking about my own past experience and leading implementations of data warehouses, I think having a team that saw more horizontally across the horizon of all the costs would’ve been probably beneficial and informative for me as one of the decision makers.

Kevin Petrie:
Yeah, it’s a great point. I’m aware that we’re in a world of proliferating ops and programs and cross-functional teams and matrices, so the last thing you want to do is add another layer of bureaucracy that ties up time.

But the way you can view this is that you’ve got a business owner that has analysts on his or her team, they want to achieve a certain goal. And when they do that, as they start to scope that, what they should also be doing is talking to a finance manager who says, “This is your budget.” Now IT’s going to consume this budget on your behalf, but you need to understand what the cost implications – as I said before – are of the project that you’re scoping.

So it’s really kind of huddling with that finance manager and making sure that they’re apprised and you have their sign-off from the get-go and throughout the process. And then periodic checks across those stakeholders that I mentioned; data engineers, analytics, data scientists, data analysts, and cloud ops engineers. I think that’s the way to do it, maybe on a monthly periodic basis.

Mark Cusack:
Yeah, I’d agree with that. And I think we’re in a current cycle in this industry in general of a rush to decentralize everything. Decentralized technology, decentralized organization. You see the rise of organizational paradigms like data mesh come around, where we want to get almost anarchic in the way that we organize our data strategy and throw everything out to lines of business.

But unfortunately, there are some things that you want to centralize, and I think most people reasonably expect that you have to centralize things like data governance. You need to have the same common standards that apply across all of your different lines of business, and we probably say the same about security. These shouldn’t be siloed. And I think the same applies to FinOps governance as well.

I think you need some level of at least FinOps like some standards that are essentially agreed on that you can apply and monitor across every line of business. So I think there are some things that are great to decentralize, but I think there are standards like FinOps and data governance that probably should be centralized.

Heather Brodbeck:
Yeah.

Kevin Petrie:
Yeah. Agreed.

Heather Brodbeck:
Okay, thanks. Well, we’re almost at time. Maybe just one last question. So if you can look into your crystal balls, what is the future of data warehousing, cloud on-prem, whatever, what do you think projecting forward, say, five years from now, what is the footprint in the cloud for an average across industries and organizations? Is this going to take off all the way to 90, 95% or do you see, based on some of the cost challenges in here, is there going to be maybe a dip and then rise back up? Just what are your thoughts on that?

Kevin Petrie:
Mark, why don’t you go first?

Mark Cusack:
Yeah, I’ll go first. Well, it remains to be seen where we’ll go. Obviously, we’re studying that kind of up cycle of flipping from more say on-prem analytics into cloud. That’s continuing. Now, it remains to be seen if at some point that will be somewhat of a trend to repatriation from cloud to on-prem, who knows.

But I do want to put my CTO hat on around where I see data warehousing going in the next five years, very briefly. And the first thing to say is data warehouses aren’t going away. They’ve been around for a long time, and what we’re actually seeing is more and more new use cases that are getting applied to them in the advanced analytics machine learning space as well. Because at the end of the day, you often end up having to process and manage and analyze structured data and a database or data warehouse is still the best place to do that.

But what I will see, if we look at purely data warehousing, I mean Moore’s Law is no longer true when you consider the CPU transistor growth right now. We’re not doubling every two years. It’s probably more like every 10 years. So if you ask me what to put your bets on in what data warehousing will look like in five years, I think you’ll see a lot more of FPGA usage, specialized hardware, but in the cloud to really start to drive even more performance and scale out of analytics platforms in the future. We’re starting to see that now. We’ve seen it for a long time, on-prem, and we’re going to see that more and more in the cloud. So specialized hardware in the cloud is the way forward.

Heather Brodbeck:
Okay. Kevin?

Kevin Petrie:
Great. I agree with Mark about the data warehouse. In fact, I would argue that the lakehouse concept is really a repackaged data warehouse, because you’ve got the consistent theme depending on who defines it has the consistent theme of SQL structures and query capabilities, on top of increasingly cloud-native object storage. So it’s kind of the data warehouse repackaged to handle multi-structured data in a lot of cases. So I think that’ll continue.

There’ll be a continued shift from on-prem to the cloud. There’ll be adoption of multi-cloud environments. So a continuation of trends that we have today. There will be a long tail of stuff, little stuff that remains on-premises, and so those factors will remain. But I think overall, there’ll be this push and desire by enterprises to kind of clawback data so that they’re not tied into one vendor, one platform, so that they can move data between clouds. And that’ll be a real back-and-forth in terms of the cloud providers, some of the big cloud platforms, not Yellowbrick, but trying to hold onto that data and enterprises trying to maintain portability.

Heather Brodbeck:
Okay, great. And we do have a couple questions here from our audience. I know we’re just a little bit over time, but we’ll hit one of them, and any of the remaining we’ll reach out on those. So this one was, what are the implicit costs of cloud analytics models besides the direct costs?

Mark Cusack:
I’ll chime in on that one first. I think one of the implicit costs that people forget, actually it comes back to people, cost and time. And if you consider advanced analytics and the machine learning life cycle where you prepare data, you train models and you score them. 80% of people’s time is around data engineering and preparing those models. It’s the cost of moving data around before you can even begin to process it. Choosing things like which machine learning model I want to apply, which algorithm, that’s the easy bit; but it’s all of that upfront data preparation and data moving and getting ready for advanced analytics where the implicit costs go, in my view.

Heather Brodbeck:
I think Kevin?

Kevin Petrie:
That’s well put. I think that’s well said. I would agree that and sort of build on the point that people cost a lot of money, and so you want to make sure you’re making use good use of people’s time. So a lot of that gets back to the fundamentals of governance, fewer copies of duplicative data, ensuring data quality, trying to reduce and automate that data wrangling process.

Heather Brodbeck:
All right. Well, thank you. I think we’ll wrap up here. Great conversation. Appreciate all the information. Anybody that’s on the line, if you have additional questions, I’m sure Kevin and Mark would be happy if you want to reach out to them directly on LinkedIn, happy to have a conversation. So yeah, this was great. Maybe we’ll do it again sometime.

Kevin Petrie:
I look forward to it.

Heather Brodbeck:
Thanks, guys. All right. Bye everyone.

Get the latest Yellowbrick News & Insights
Stories From the Field: The Devil Is In...
As a software engineer, a big part of my job...
Enabling Enterprise-scale Cloud Analytics Without Sky-high Costs
Righting the Wrongs of Cloud AnalyticsThe cloud has led to...
Yellowbrick Squares Up to the Data Warehousing Heavyweights
Following a knockout decision, we can announce that Yellowbrick is...
Book a Demo

Learn More About the Only Modern Data Warehouse for Hybrid Cloud

Faster
Run analytics 10 to 100x FASTER to achieve analytic insights that have never been possible.

Simpler to Manage
Configure, load and query billions of rows in minutes.

Economical
Shrink your data warehouse footprint by as much as 97% and save millions in operational and management costs.

Accessible Anywhere
Achieve high speed analytics in your data center or in any cloud.