Snowflake Positioning vs AWS, GCP, Azure

Amazon European Logistics Buildout

AWS: Competitive Positioning & Shift from Horizontal to Vertical Selling

Amazon Advertising 

Disclaimer: This interview is for informational purposes only and should not be relied upon as a basis for investment decisions. In Practise is an independent publisher and all opinions expressed by guests are solely their own opinions and do not reflect the opinion of In Practise.

{audio:0:01:58} How does an ISV program works at AWS.

{audio:0:03:35} AWS realized, about three years ago, that we were behind in engaging with our ISV partners. They built this newer team; it's not as new anymore because there are many of us across the globe. Our remit is co-sell, meaning how do we help our ISV partners and as you know, there are thousands of them, growing daily. 

{audio:0:04:18} My role is to help these partners engage with our AWS sales teams so that we can co-sell together. That carries a lot of different meanings and facets with it. Sometimes a Snowflake may already have a footprint with a Northwestern Mutual, and the AWS rep works with Snowflake because Snowflake can get us into other lines of business within an enterprise client. More often than not, AWS has the footprint and helps the Snowflakes. I believe GCP and Azure have similar functions to a certain extent. I think Microsoft, even dating back to the 90s, always had a strong partneresque type program. 

{audio:0:05:38} The traditional ISM, ISV Success Manager, is regionalized, meaning they focus on a certain country or area. If a sales rep says, I need a data warehousing partner; I'm considering Amazon native Redshift versus Snowflake; how can you help me and get me connected with the Snowflake folks to talk about these things? Our ISV partners are large ones like the Snowflakes and the CloudStrikes and Datadogs of the world.

{audio:0:07:36} What's struck me from the outside is the attitude toward partners changing, which was more and more accentuated over the past 12 months. I heard that there'd been a reorg where you merged the market space with your organization. 

{audio:0:08:36} Amazon, at a senior level, has realized the value of ISV partners in growing our business. It comes down to the cliché, but the one plus one makes three. They've got the data to say that as our partners grow in AWS, so does AWS. So to give you a stat, when I started two years ago, our what we call partner-attached was at roughly 10-15%. That means for every AWS workload revenue that came in, 10-15% of that business was attached to some type of partner, whether an ISV or an SI reseller. As they brought in new leadership to AWS to focus on partners, that goal for this year is over 25%. 

{audio:0:09:58} Your other point was Marketplace being folded in from being a separate transactional entity into the partner fold. What we call APN, the Amazon Partner Network, is the ACE system (APN Customer Engagements) within a partner sales force, our sales force, just to manage partner deals. APN and Marketplace are now part of the same realm, so to speak, within Amazon. To your point, there has been a shift over the last three to six months of that folding in with the APN and the partner org.

{audio:0:10:51} What changed here in terms of a business model? My understanding was always that in the Marketplace, you would get a cut. I've heard several figures, but let's say 15%, 20% of the partners' revenues. I'm trying to understand why that shift and what has changed.

{audio:0:11:16} I don't believe any fee structure has changed. What's changed is more emphasis on helping the partners sell more through the Marketplace. Marketplace is a cash-generating machine for AWS, not just because of the listing fees and the percentage that AWS takes of Marketplace, but once you get a client, whether greenfield, SMB, or enterprise, into Marketplace, it makes procurement that much easier. It makes programs that much easier, and it makes a client stickier. That's the value of Marketplace.

{audio:0:11:59} Let’s take the specific example of the call today, Snowflake. You were their partner. Snowflake was selling directly or is co-selling through you and is also selling through Marketplace. Do I understand that correctly?

{audio:0:12:25} Yes, they can also bypass Marketplace altogether and sell directly to a customer with or without AWS.

{audio:0:12:36} What are the incentives for Snowflake? I guess they get more, but they have less reach than AWS?

{audio:0:12:54} Correct. Snowflake is definitely in our top five as far as ISVs and revenue that they bring in, so they are very dominant. I think your question is, why would Snowflake even co-sell with AWS if it costs them more to go through the Marketplace, or what benefits are they reaping by partnering?

{audio:0:13:15} I guess it's your reach and the sheer number of clients.

{audio:0:13:20} Correct. I can’t state Snowflake’s stats, but I can tell you Datadog, which is also a top partner, I want to say they have 5,000 customers and AWS has 200,000. It's some outrageous stat, but it's that stat that says we need AWS to expand our footprint and help us get net new logos. For any of these ISVs, the one consistent thing is net new logos, and they see AWS as a machine to help them get net new logos and expand with current clients. 

{audio:0:14:19} I’m trying to understand the perspective of AWS co-selling with Snowflake. You have your product that is competing with Snowflake. Snowflake is expanding in its products, but initially, Redshift was the main competitor to Snowflake. In the beginning, the way I read it, AWS was competing and trying to attach their own product rather than Snowflake, and this changed. 

{audio:0:15:19} Yes; it’s really interesting you say that. If I look back at a year and a half ago, when I was supporting enterprise sales teams, it was a rub. Honestly, I would get on co-sell calls between a Snowflake sales rep and an AWS sales rep, and they were both extremely guarded. Snowflake felt that AWS was just trying to get information so they could sell around them and sell Redshift.

{audio:0:15:52} Conversely, while trying to sell more Redshift, AWS felt like Snowflake was the competitor. I believe what has evolved, and Snowflake has continued to grow substantially with AWS, is that Snowflake does data warehousing and everything that comes with it extremely well. All the bells, whistles, and much more functionality and support than Amazon could ever give to Redshift. I think of it as Redshift is more of the blank open slate if you want to build on it. Snowflake is in the build versus buy model. I believe that AWS has recognized that over the past year and has changed their tune on selling. I had one rep put it to me, if you can’t beat ‘em, join ‘em, and this is a really good rep who gets it. He said sure, I want to sell Redshift, but I don’t want to alienate my relationship with Snowflake because if they win the business, I need them.

{audio:0:17:05} Could you speak to why Snowflake won the business? Superior performance? Easier to manage?

{audio:0:17:14} Yes. Easier to manage and cost-effective; they handle tons of data, whether on-prem, in the cloud, or shifting; the storage, the management of the data, the whole ecosystem, and support. They are a pay-for-what-you-use model and have a strong sales team in the field. When I talk to clients, I hear good things about Snowflake. I think they've made migrations easier. Redshift is a native product within Amazon. Sure, it's a great product, but you're not going to get the kind of support you would with a Snowflake one, honestly.

{audio:0:18:13} You would say Snowflake is strong in migrating? That’s how they win their clients?

{audio:0:18:22} Helping clients migrate to the cloud, yes.

{audio:0:18:26} And AWS is probably stronger with companies starting natively on the cloud.

{audio:0:18:33} Definitely. So we have what's called our DNB, our digital native business. Those are companies that never started on-prem; they are in the cloud. Some of the younger startups are starting in the cloud. So yes, that is a sweet spot for AWS. However, with so many large enterprise companies in the ecosystem that have been around for years, the Procters and so on, they know they need to transition from on-prem to the cloud. Many of them do it in somewhat of a hybrid manner. It's not like you flip the switch overnight. So we work very closely with those clients on their migration efforts. It's a side note, but there are many programs in AWS. Our migration acceleration program is an example of funding available with a partner like a Snowflake to help that client migrate to the cloud. So yes. It's a yes and yes. AWS has a foothold in both.

{audio:0:19:47} Would you think the increased collaboration with Snowflake is to win against Microsoft, who is stronger historically with enterprise clients because they own those relationships from prior times and in their on-prems' infrastructure?

{audio:0:20:06} Yes, I think that’s part of it. The stronger we can be with a partner like a Snowflake, helps us compete against Microsoft, as GCP is very up-and-coming and aggressive in the ecosphere. I wouldn't say it's the major reason; it's a part of it. We could have said two years ago that Snowflake is a competitor; let's not partner. Still, we took the reverse approach because every instance that Snowflake sells has a benefit to AWS, and Snowflake was chosen as one of these what we call top 25 ISVs that are given extra resources from AWS to help with co-sell.

{audio:0:21:02} What are the incentives for a rep? Is he better off selling his own product, or was that the case before, and now it has changed? Could you talk about the incentives of the reps at AWS?

{audio:0:21:22} Certainly. A rep has an overall quota goal, and several ways of relinquishing that quota exist. The number one way is to sell more AWS services. But what has been enacted – and I think this is important in getting to what you're saying – as partners grow within AWS, they achieve certain statuses such as ISV accelerate, which means you’re part of a partner program, and AWS reps receive SPIFs for co-selling with a partner. If the partner brings a deal to AWS, and it's launched and sold to the client, then the AWS rep receives a 1% SPIF of their variable compensation for that deal. That's just cash in their pocket. If AWS brings the opportunity to Snowflake and the deal closes, the rep receives a 3% SPIF against their variable comp. It doesn't matter if the deal is $50,000 or half a million; the SPIF is against a variable compensation number that the rep has as far as their compensation, their base plus variable compensation model. 

{audio:0:23:07} No worries at all. It’s confusing. Once a partner is in ISV Accelerate, they have to do certain partner things with AWS to get there, and there are probably about 1,000 ISV partners in ISV Accelerate. So if I'm Snowflake and I bring a deal with AWS with, let's say, Discover, and I work with the AWS rep on Discover, and we close the deal, the AWS rep gets a 1% SPIF for closing that deal. And that 1% is not based on the value of the deal. The deal could be $50,000 or half a million. The 1% is based on that rep's variable compensation as far as their overall compensation package. So to put it into perspective, if a rep has $100,000 variable comp, for that one deal they sell in, 1% of that $100,000 or $1,000 is cash in their pocket, which is capped out at 25%. So most good reps will cap out on that compensation towards that.

{audio:0:24:33} Conversely, if AWS brought the deal to Snowflake and it closes, that SPIF is 3%. So this SPIF is just one of the incentives for an AWS seller to co-sell with a partner. They can't get that SPIF just by selling AWS services; it's an incentive to work with partners, and that's one factor. 

{audio:0:25:01} That's good. I got it now.

{audio:0:25:04} That's been around for a while. It's a nice incentive, but AWS sellers, good big enterprise sellers, will say that selling with partners just takes too much time. I'm worried about selling my EDP into my client, the big deals. The second piece enacted that is showing traction – especially this year more than last year – is a compensation called SRRP. That is the SaaS Revenue Recognition Program, which means nothing, but I will explain it. Any ISV partner that's part of the SaaS Revenue Recognition Program – and there are about 120 of them now, so you can see where AWS has been doubling down, so to speak, on top partners – if they are deemed part of SaaS Rev Rec, 30% of the TCVs, of the total cost value of that deal goes against a rep's compensation. So it adds up pretty darn quickly. 

{audio:0:26:17} If you think about Snowflake again, if Snowflake and Discover sell a million-dollar deal – say it's a $3 million, three-year deal for easy math – into an enterprise client, that rep gets 30% of that or a $300,000 against their quota amortized over the three years. So that would be $100,000 a year; call it $8,000 a month amortized over the lifetime of that deal against their quota. At first, it wasn't a huge incentive, but if you start thinking about selling more and more of these partner deals, it turns into the gift that keeps on giving against your quota. 

{audio:0:27:14} It's become so much more material to the point where I can get on calls with reps, and they will ask a partner, are you part of our SaaS Revenue Recognition Program. You can see in 18 months that this program has gained a lot of traction with our ISVs and sales teams. I think most sales folks at AWS will tell you that the SaaS Revenue Recognition Program and the 30% commission towards quota is a bigger incentive to work with ISVs than the SPIF, which is a change. So two programs are cash incentives towards quota for those partners.

{audio:0:28:00} The third piece is this year, AWS reps and inside sales reps, and anybody tied to an account, has an MBO objective about so many partner-attached wins. It's probably in the neighborhood of 15 to 20, and it's just one of their, call it, 12 goals that their performance is evaluated against. They need to have X number of partner-attached goals. That's another incentive to work with partners. Many of my good reps will say, let's knock off these 20 goals before six months because I want to turn that one green, so I don't have to worry about it. That's another incentive to work with partners. The last one I'll discuss is Marketplace. Selling through Marketplace allows clients to relieve their EDP, so as more and more clients sign up for EDPs. Are you familiar with the EDPs?

{audio:0:29:09} Okay, I’m trying to remember what it stands for, but it's an agreement between AWS and a client on hitting a certain amount of revenue with AWS. It's like any company.

{audio:0:29:39} An annual agreement. Going through the Marketplace, you can retire Marketplace revenue towards their EDP, so that's an incentive to the client. It's also an incentive to the AWS rep because they want their clients to retire their EDPs. So it varies from 50% of the deal's value towards EDP, or in some cases, if it's not new, it can retire up to 100% of the deal's value through Marketplace towards their EDP goals. So that's an incentive for reps to push more Marketplace, but most reps will only push Marketplace if their client wants Marketplace. It's a transactional function and easier from procurement, etc., but there are incentives to go through Marketplace. There are also Marketplace private offer promotions and so on, that ISVs can run to incent clients to buy ISV products through Marketplace.

{audio:0:30:49} Those four pillars make sense. Now you have incentives to push through those four levers. How was it before when a rep was using or co-selling Snowflake?

{audio:0:31:09} The difference now versus then is a couple of things. Number one, the AWS reps are much better educated and informed on ISVs than they were, call it, two years ago. It was a necessary evil in the business but wasn't a focus. PDMs, partner development managers, have been around for a while, the ISM side of the house is still fairly new, and part of that job is to educate AWS reps on the value of selling with partners. Part of it is the education, and part of it is seeing its benefits in their business and coming around to believing in the value partners bring to the business. So education, seeing results, and the newer incentives that are out there for partners. 

{audio:0:32:14} I go back to that education. There are more partner program educations now than there ever have been. Immersion days, partner talks, you name it. An AWS rep could spend half their week sitting on calls being educated about partners. They don't, but there is much more enablement and support in AWS for partners than before to get on the radar with these AWS reps. The summits are back. I didn't think about this before, but the world starting to open up again also adds a factor. These AWS summits, there was one in Atlanta and there's one coming up in New York, they are client summits, but they bring everybody out. Partners sponsor and support them; partners line up happy hours, dinners, etc. Some of this coming back to real life is helping sellers and partners and the whole ecosphere of the partnerships. 

{audio:0:33:25} How would you compare and contrast Databricks and Snowflake?

{audio:0:33:45} I can't say because I'm not close enough to either of them. I don't manage either of those partnerships. I know Snowflake better from the deals I've worked on with Snowflake, from the data warehousing side and their footprint. I would say from my perception, they’re more ubiquitous, and they're larger. Snowflake just has a larger footprint and what seems to be a growing and stronger foothold. Databricks is part of that top 25 – my peer who works on them knows more – but I don't get the sense that they’re as strong of a co-sell partner with AWS as Snowflake is. And I don't have any data to share as far as why. My understanding is that there are just some barriers there. Some of it may be competitive, they might have a strong foothold with Microsoft if I’m thinking about it right, but they’re just not as large. I don’t hear about them as much.

{audio:0:35:01} When you say top 25, I imagine Datadog and Snowflake are in the top 25. What else is an ISV? Would you also count Salesforce and ServiceNow as ISVs? 

{audio:0:35:12} No, ServiceNow is not. They fit more probably on the reseller consultancy side of the house. We don't work with ServiceNow as an ISV. They're a large force in the ecosystem, but they're not considered an ISV, so they're not part of that top 25.

{audio:0:35:42} Salesforce is not. Let me caveat that. Parts of Salesforce are, so Tableau is. Let me also caveat this top 25 because I don't want to mislead you. The top 25 chosen and deemed to have these added co-sell resources were not chosen necessarily because of their revenue size with AWS. They were also chosen based on the strategy they can bring to selling with AWS. For example, I have a very small CDP, customer demand platform, that barks up the marketing tree versus the IT tree, so they're not nearly big enough. Still, they were deemed part of this top 25 because they can help AWS expand to new lines of business and vice versa.

{audio:0:36:53} If we think about competition, the way I see them and I would like to have your view is in the emerging state of things like now, I think we're in the very early days of the cloud, I'm sure you think the same on that. If Snowflake wins, AWS wins. There is more storage, and more compute, so it's beneficial to both. However, on the data layer, the control gravitates over time to a Snowflake, which is why I understood the competitive reaction of AWS in the beginning. In my research, I also heard that AWS was revamping Redshift, changing the architecture because apparently Redshift has an architectural disadvantage versus Snowflake. Has this been discussed? As a long-term threat to losing out on the data layer or becoming commoditized to some extent?

{audio:0:38:15} I would agree with you; it's definitely been discussed. I feel like there was some type of shift, and again because I don't manage Snowflake, I can't say this for sure. In the past year, Snowflake and AWS leadership said we could work together for the sum of the whole and benefit both of us. The cloud is continuing to grow, and we can both win or choose to compete. I think the choice was let's work together, and I think those conversations are transparent and honest about how we will work together. But could we revamp Redshift and make it a stronger competitor to Snowflake? Absolutely. That happens all the time. At this stage, I think we've agreed that the sum of both is better, but could that shift down the road as we invest more in a competitive nature and capture more of that data on behalf of AWS versus Snowflake? That could shift as well. I watched when AWS brought in their latest healthcare initiative, and UnitedHealth Group was one of my clients. That made for a very difficult discussion. A tough sell in that AWS was creating something that would compete with UnitedHealth Group. But the evolution and innovation of AWS mean there will be times when we compete with our partners and clients.

{audio:0:40:16} The picture of Snowflake is very often that customers want to be multi-cloud, and that's the value that they bring, that basically, on the level where you're at, you're always on Snowflake. Still, you as a customer can decide where your data resides, and that's the value add the Snowflake brings; apparently, that's their pitch. Now I'm asking you, have you come across multi-cloud? How important is that to clients?

{audio:0:41:00} Yes, and in fact, it kind of surprised me. I would have thought, coming from Oracle, Google, and other data companies, that it would be much simpler for a client to put all their eggs in one basket. Procurement, systems, onboarding, all those things, wouldn't it be simpler to be with one cloud? Before AWS, I worked very closely with Procter & Gamble, the number one CPG advertiser. They were multiplatform, and I was told it was for negotiation purposes, which is why they didn’t go with one provider. Fast-forward to AWS, and we find the same things. It helps them have a pulse on all three different clouds, so that's one piece; the control and negotiation power versus just having all with AWS. 

{audio:0:42:14} I have had a couple of enterprise clients in the FinServ space that went all-in AWS and switched off Microsoft in that route, but I would say that's the exception versus the norm. There is another reason for multi-cloud for some clients; UnitedHealth Group is a perfect example that has had a stream of acquisitions over the last five years. Every one of those acquisitions came with its own baggage of whatever that might be; that's where in the case of UnitedHealthcare, they had a stronghold with some of the newer acquisitions that came in with AWS versus some of the legacy that was on Microsoft. That's the other piece of it, too, you've got a lot of mergers, M&A activity out there that also increases multi-cloud.

{audio:0:43:12} I'm not talking specifically about P&G, but if we take a company like P&G, what would the split be of their IT spend? Would it be like 60% AWS, 30% Microsoft, and 10% Google? How would that split be across clouds?

{audio:0:43:33} I can't speak specifically to P&G, but if I think about an enterprise-type client, I would say yes, your percentages are probably spot on. What I hear in the field, and I hear this more from my ISVs than I do actual clients, is that AWS is dominant. It's hard to get around not working with AWS, so they will have a stronger foothold. Microsoft has more of the legacy, but they're still up and coming as Azure becomes more mature. Then GCP is kind of like that little engine throwing a lot of money at the game and a lot of money at clients, especially around their Marketplace. They're developing that. So they're more up and coming.

{audio:0:44:42} From what you hear, you're in the industry, your talk to people, you’re at these conferences. How do you see this sector evolve five years from now, and what role does a company like Snowflake play in your view of the world?

{audio:0:44:59} I think Snowflake has such a stronghold where they compete in data warehousing. They have such a strong lead and are continuing to grow that they'll grow with it. I think the stat I heard when I first started is 4% of IT spend in the cloud – that was an Andy Jassy – and I think it's closer to 10% now. I believe the journey is still very early, so I think it's a tale of, to a certain extent, all tides rising between all three of those. GCP is not going to give up. I know because I worked at Google, and I know in talking with colleagues that the investments are going to the cloud side of things, to the point where sellers that were typically digital media are now jumping ship over to the cloud side because that's where the money is. I know that GCP is throwing money at it. 

{audio:0:45:59} We just had a senior leader go over to GCP, so they're going to continue to grow, they've got unlimited pockets, and they're going to continue to grow and invest in that product. They're not going away. They've got an uphill battle mainly because AWS is just so far ahead, and the irony with how AWS stays ahead, is that AWS culture, unlike Oracle and some of these bigger companies, is very innovative. It gets back to the leadership principles, constantly iterating, improving, and evolving. The leadership won't let it become an Oracle where we overthink things. It’s evolving. In answer to your question, I don’t know the percentages of the three major cloud players as far as their percent of the business changes that much in the next five years. I don’t see AWS losing ground, but I don’t see GCP stopping either. 

{audio:0:47:09} You don’t see anyone else?

{audio:0:47:12} I hear little bits and pieces about IBM and Oracle, but not really, no. My purview is very North America-centric; you may have a different purview from a global perspective.

{audio:0:47:29} Yes, there will be regional differences, but I would argue North America is three to four years ahead in that industry.

{audio:0:47:37} Yes. I would agree. They want my support globally because they say at least a year ahead. They don't say three to four years ahead but at least a year ahead from an ISV perspective. 

{audio:0:47:55} To get a sense, let's support Snowflake; you said they would grow with it?

{audio:0:48:04} If you look at customers using Snowflake for an extended period, Capital One is the core example. I suspect you are thinking of this logo when you said going all-in on AWS. There are a few of these, and Capital One is one of them. If Capital One spends 100 with AWS, do you have a sense of how much they spend with Snowflake?

{audio:0:48:39} I think what you’re trying to get to is, with an enterprise client that spends with AWS, what ratio would a Snowflake pay?

{audio:0:48:51} Correct. I guess a company, like Netflix, the large clients they get. Hundreds of millions for AWS are based on more or less public information on the web. 

{audio:0:49:07} My question is, how large Snowflake becomes and what its footprint is because it's the second-largest bill item for the more mature Snowflake accounts. Where does it land?

{audio:0:49:22} Yes, I would be guessing, but I think it will vary. Let's say Snowflake is the data warehouse of choice of a customer primarily on AWS, and what's that spend ratio? I think you're trying to say that.

{audio:0:49:40} My ballpark would be 15% to 25%. But I might be high. I would need to go in and do a back-of-the-envelope analysis.

{audio:0:49:57} You still speak of Snowflake as this data warehouse company, which was their initial branding, but they're going more and more towards this vision of data cloud. I looked at their summit keynotes and stuff. They're releasing tons of stuff, transactional databases, they used to build only structured data, and now they're going towards semi-structured and non-structured data. It seems as if Snowflake is building a fourth cloud on top of the three clouds.

{audio:0:50:40} Is that something being discussed?

{audio:0:50:44} Meaning discussed with AWS?

{audio:0:50:50} I'm sure it is; I'm not privy to those conversations. I can say subjectively that my counterpart, who runs Snowflake, is at every one of those summits you're speaking about, meaning they're invited to those summits, so it's a strong partnership. He's ingrained in conversations around partnership and co-sell. He's not necessarily going to be sitting at the table with a product on how they evolved. Still, I can pretty much guarantee that Snowflake's product people and AWS product people are partnering on how we grow together and how the next evolution of whatever Snowflake is creating work within AWS or, for that matter, compete. It's just the nature of where it's going.

{audio:0:51:49} Very quickly related to that, I think AWS, from a leadership standpoint, has been very transparent that there will be times where we compete. It's just the nature of the ecosphere of cloud.

{audio:0:52:06} Initially, it was just Redshift, but over time, it would become some of the database products that you have; it would become some of the AI and ML workload services that you have that would be competitive with stuff like that, that’s why I’m asking. 

{audio:0:52:27} Yes, and if you think back to pre-AWS and just look at the Amazon machine, Amazon's retail model was built the same. If a seller came in with the latest dish soap that was off the charts, Amazon would recreate it as their own private label and squash that, which is not ideal from an enterprise standpoint, but it's the economics of how Amazon thinks and runs. So as much as we value those partners, Amazon will invest in growing Amazon. 

{audio:0:53:11} It makes sense. To close off the last 10 minutes, I’d like to talk about customer behavior and churn because you were talking about stickiness in the beginning. Could we talk about the stickiness of AWS and clients? Is that something you measure, and could you compare how you think a Snowflake would do there?

{audio:0:53:44} And you’re talking about the stickiness of an AWS client staying with AWS, correct?

{audio:0:53:50} Yes, and how many clients leave AWS for good?

{audio:0:53:54} You know what? I would say not many. That's not to say it doesn't happen. Where it’s going to happen is where GCP – more so even than Azure – puts something in front of a client that is so lucrative that they can’t not leave; from an EDP, incentives, whatever, they "buy their way in." That's going to happen because it can happen. I don't see it happen very often. 

{audio:0:54:38} I think what Amazon has done with our enterprise clients, from the EDPs, the incentives, more and more go into these SCAs, so to speak, as incentives to stay sticky. If anything, those incentives are growing to keep these clients. So I would say retention-wise, if I think about ongoing clients, I know the EDP commits are growing; I know that just from what I hear from the reps. I've yet to hear of an EDP that shrunk year over year. 

{audio:0:55:20} Once a client, let's say a division of a big enterprise, is with AWS, have you experienced cases where churns often switch to another cloud?

{audio:0:55:46} I'm saying that I don't hear of that very often. But I want to put a caveat to that, where that does happen is where Amazon is a competitor. So you will see that in retail, you will see that in even healthcare, or where Amazon creates something that competes. That’s why Microsoft is so strong in retail because retail doesn’t want to support Amazon. 

{audio:0:56:19} Would you expect churn rates for a company like Snowflake on the data layer to be higher or lower than what you see on the cloud layer?

{audio:0:56:28} I would say a little bit higher than AWS only because I think Snowflake has other ISV competitors in that data layer, and I can't speak to it as much. I just say it from the perspective that the ISV ecosystem is growing so substantially that there will be competitors to Snowflake from an ISV partner standpoint versus a cloud. I would assume any ISV will have a higher churn than what we see on AWS regarding data cloud. 

{audio:0:57:09} You were talking about the increased consumption of existing customers over time. What drives that? Is it just the ongoing shift of workloads to the cloud and more divisions, applications, and data being stored? Why is that growing so aggressively? Because the growth rates are astounding.

{audio:0:57:44} Yes. I think it's what you just said. It is more workloads being shifted to the cloud. If you think of a typical enterprise client, they’re going to dip their toe in the water. They're going to move the workloads first that aren't going to kill the business if something goes wrong. So they migrate in pieces. Once they get that first layer that is not customer-facing and get the learnings from that, then they’re going to migrate the second layer, then they’re going to migrate the various pieces. There’s a system, a pattern, a plan, and a strategy for what migrates first. You never migrate everything altogether unless you’re cloud-native and start in the cloud; that's a different story. 

{audio:0:58:43} But if you think about the growth from 4% to 10% and ongoing to the cloud, these enterprise clients realize they have to get there, but it is a slow journey. So it’s continuing to evolve. I supported FinServ, and FinServ, especially because of the privacy and the security involved with FinServ. They were very cautious and very slow-moving. They all knew that they needed to get to the cloud, but it was not a rush. A lot of that, if you think about enterprise, these are IT people that have been in their careers for 30 or 40 years, and 30 or 40 years ago, cloud didn’t exist. So these are not people jumping up and down saying, I have to get there. Any leadership knows they have to get there, but nobody's going to risk client, breach, security, etc. That's why there are so many security-based ISVs; for that exact reason, it's securing the cloud.

{audio:0:59:53} You talked about these attach rates - what does that exactly mean?

{audio:1:00:18} It means partner attached. That's a good way of thinking about it. That means that of every dollar of AWS services sold, 30% of those have a partner attached to the deal. I would think of it more from account versus a revenue perspective.

{audio:1:00:49} It's more than that because you would take a cut on those revenues, so the 30% is a cut.

Snowflake, Redshift, & AWS Partner Program

ISV Partner Sales Lead at Amazon Web Services

inpractise.com/articles/snowflake-redshift-and-aws-partner-program