The Economics of Cloud with Esteban Kolsky

The first part of Logikcull’s interview with Esteban Kolsky addressed common misperceptions about the cloud, which Kolsky defined, strictly, as a three-tiered architecture divided between infrastructure, platform and software.

“It’s the three elements you use when you create an application: the presentation layer, the logic layer and the underlying network and database,” Kolsky, a former Gartner analyst and widely respected technology consultant, said.

Here we dive — into the weeds, at times — into the evolution of cloud computing, the economic advantages and challenges its purveyors face, and how companies like Salesforce are pioneering customized “industry clouds.”

Also, we answer the question: What the heck is industry cloud?

Logikcull: What was driving the success of Salesforce in the early days? It was the lower total cost of ownership and the ease of use.

Esteban Kolsky: Yes. And it was the fact that they quote unquote were “renting” out the software, because the customer didn’t pay for everything — they didn’t own it. So the customer didn’t have to buy servers. They didn’t have to buy software in-house and host it on their systems. The customer was distributing the cost over three years.

At the end of the day, it was cheaper at the beginning, but as it started growing, Salesforce started charging the same as its competitors. Three or four years in, price-wise, Salesforce was the same as having an on-premise solution — because, with an on-premise solution, the customer makes the first investment, but then after that, it only pays a 15-20 percent yearly maintenance fee. With a hosted solution, you take the same amount you would have paid over three years and divide it evenly over 36 months, and at the end of the day, the customer has nothing. The customer doesn’t own anything.

Logikcull: Well if that’s true, what’s driving enterprise cloud adoption then?

Kolsky: True cloud architecture is adopted because of reduced cost. If you only pay for what you use, you don’t have to buy stuff you don’t need! [laughing] That’s the easiest way to say it. So if I have two customers running on my application, I pay for that. If I have a hundred, I pay for that. I’m only paying for what I use. The resources are distributed and leveraged in much better ways. One server running a platform that is used by 20 different clients, it’s the same thing you would have used before with an on-premise or hosted model, but now it can be scaled infinitely — so, for the customer, the more you use, the less that you pay. Look at the pricing models for Amazon (Web Services), and Google, and IBM and even Oracle cloud applications. The more you use, the less you pay because it’s cheaper for me (the cloud provider) to have more things running on the same hardware.

“True cloud architecture is adopted to reduce cost.”

Logikcull: And because Salesforce is not “true” cloud, it can’t leverage those economics of true cloud. Is that what you’re saying? To summarize what you just said: true cloud is lower cost because you only pay for what you’re using. And if Salesforce really isn’t architected that way, then Salesforce isn’t able to offer those economic advantages?

Kolsky: Well, so Salesforce wasn’t architected like that — the “true” cloud — originally. But they’ve spent the last six years re-architecting to get closer to that model. Now the platform is more independent. They’ve added a data abstraction layer which has given them the scalability they need for data management. And they’ve broken the software layer away from the rest of the system, and changed the database, which will put them on a three-tier “true” cloud model. They’re not there yet, but they are very, very close. But not the closest that you have. Still, today, if you’re a third-party provider that wants to leverage the Salesforce platform, you can do that. You just pay for access to the platform, and then you deliver that as part of your solution. Just look at something like FinancialForce.

Logikcull: And if Salesforce didn’t move to that architecture, they wouldn’t be able to compete with other offerings down the road.

Kolsky: That’s right. But Salesforce is the only one from the four major vendors — Microsoft, SAP, Oracle, and Salesforce. They’re the only one that has done the re-architect. Microsoft is the closest second one. But Oracle and SAP, they’re nowhere near.

Logikcull: So what about private clouds and on-prem technology? Are they making a comeback even amid all this cloud adoption?

Kolsky: Well I mentioned this earlier, but if you’re a CIO who has been tasked with moving to the cloud, the biggest question is, how do I move my legacy data to the cloud? And the easiest way to do that is to build a quote-unquote private cloud, which is essentially a privately hosted solution that leverages some of the cloud’s architecture technology, but not all of it. The private cloud is mostly a browser-delivered solution to a legacy application. So you need it as a stepping stone. Private cloud, hybrid cloud, mixture of public and private: they’re all stepping stones to get to a true cloud architecture where everything is open and public.

“Private cloud, hybrid cloud, public-private mixes, they’re all just stepping stones to get to a true cloud architecture where everything is open and public.”

Logikcull: This move to 100 percent cloud, can the software giants of the past — IBM, Oracle, Microsoft — can they be as as profitable as they were with the on-premise model? And generate as much revenue?

Kolsky: Well, that’s the question. This is what’s effecting Salesforce. Salesforce is close to being able to deliver platform services, but they cannot do that because they have to get rid of the application revenue which is nice, guaranteed revenue — which is what the entire business model is based on. So, they would have to move away from the mindset that says “I’m going to be able to make this much money, because I have a contract that last three years” to “I don’t know how much I’m going to make because it’s up to people to use my solution,” which is how you price a pay-as-you-use, pay-as-you-go model.

Logikcull: In truly elastic fashion.

Kolsky: Yeah. If people don’t use my services and my platform, I don’t get paid. Now it’s like, I have a license and I need to use it, because it’s locked in — maybe through a browser or through a platform, but you’re locked in and you have to use it, right? As we move closer to cloud computing architecture adoption by organizations, including the infrastructure and the ability to build the applications and all those things, you’re going to see companies like Salesforce change their revenue models. But they’ve been trying to do this for a while, and they can’t get buy in from Wall Street and the public. If you’re a publicly traded company, you just can’t have inconsistent revenues like that.

It’s the same problem for everyone else — Oracle, IBM, Dell, HP, all these guys. They want to adopt cloud but they can’t get rid of the licenses and the persistent revenue they have. They need to be able to show that they still have it and maybe build a different stream and then slowly begin to replace what they have. It’s going to take a while.

“Oracle, IBM, Dell, HP, all these guys — they want to adopt cloud but they can’t get rid of the licenses and the persistent revenue they have.”

Logikcull: Let’s talk about the industry cloud… What the heck is it?

Kolsky: [Sighs deeply] I guess just advice provided by analysts? Salesforce decided to call every single application they have a cloud. They have a service cloud, and analytics cloud, and so forth, and one of those became an industry cloud. If you go back to the beginning of CRM, the beginning of ERP (enterprise resource planning) and look at how they evolved, these companies start with a basic solution in certain industries and then slowly find places where they require more focus — different integration levels that are not required anywhere else, that require different workflows, and compliance and liability and legal schemas, and require different ways to work. As you start evolving, you start building industry-based solutions for whatever you’re delivering.

So (in the context of Salesforce) the industry cloud is basically an application that’s based on a specific cloud of Salesforce. I don’t remember what exactly they started with, but healthcare was one of them, and financial services. They’re basically looking at what the clients are asking and delivering value for their specific industries in the form of a cloud application — and they’re calling that the “industry cloud.”

Logikcull: I don’t understand why that’s such a novel thing. It’s just a cloud deployed by vertical, right?

Kolsky: Yeah, it’s only a novel marketing element. What Salesforce does really well is market, so it’s a novel marketing element that nobody has really caught on to. Nobody has been delivering industry-focused applications on a cloud architecture or on a hosted model. So what Salesforce did was say, “If you’re in financial services, we’ll do Salesforce this way.” They’ve been doing this for non-profits for a while, and for government. Now they’re starting to go into other industries and making products for those industries because it’s easier to market a product than say, “You can use this (basic Salesforce) if you want to.”

“Industry cloud is only novel in the sense that it’s a novel marketing term.”

Logikcull: But even beyond Salesforce, just generally. Everyone’s doing this (industry cloud), right?

Kolsky: Yeah, everybody’s doing it. It’s the natural evolution. Once you have a solution that captures 80 percent of your industry, then you start customizing things to capture the last 20 percent.

As told to David Austin and Robert Hilson of Logikcull. David can be reached at Robert can be reached at