How SnapLogic's cloud integration bet paid off - with CEO Gaurav Dhillon

Gaurav Dhillon is no stranger to the pitfalls of integration. As a former CEO of Informatica Corporation, he has been trying to ease the enterprise integration challenge since he co-founded Informatica in 1993, back when the closest thing to a "Enterprise Service Bus" was down the street at the bus stop.

Fast forward to 2009, when Dhillon joined SnapLogic as CEO. From the get-go, Dhillon had huge decisions to make on SnapLogic's direction - decisions which may seem like no-brainers now, but were big gambles during a time when tech heavyweights dominated the integration landscape, and "big data" hadn't even started its bumpy ride on the hype cycle. Last week, I had a chance to talk in-depth with Dhillon about the chances SnapLogic took on cloud integration, and the drivers behind their triple digit revenue growth.

Two big bets: one on hybrid, one on analytics

Jon Reed: Tell us about those big bets that you made.

Gaurav Dhillon: The first big bet that we made was that one platform - a hybrid platform that runs on-premise and in the cloud - can tackle both app and data integration. That was a big bet. There's multiple, multi-billion dollar entities in the market - SAP, Oracle, Informatica, Cisco, and so on. But we thought that the back of all the modern web infrastructure, we could build a new company that is a superlative for where the world is going to, A, with the cloud, and B, with big data.

Reed: So what was the risky part of that bet - going for a hybrid landscape, combining apps and data, or both?

Dhillon: Both. We are a new product for an existing market. Enterprises have had all sorts of applications and, therefore, all kinds of integrations, ESB, ETL -  for a long time. When you come into this market, you have to meet the table stakes requirements. The integration has to be rugged, reliable, and secure. Mission-critical processes are at stake - whether it's shipments from iRobot, or Uber onboarding their contractors - all these things are on the line, using our products.

So, our products have to be robust, secure, and available, in an architecture that is metering what people are doing.  It was a bet on the fact that the enterprise would become hybrid. In our case, you have a very declarative style, a self-service model of integration.

Reed: You don't even like the word integration all that much, do you?

Dhillon: Integration is such a '90s word. As your colleague Phil Wainewright has said, it's really about connections. What we're doing is providing connective tissue for the enterprise as it goes into the cloud. Pretty much everything but financials is going, going, gone - whether it's human capital management with Workday or service management with ServiceNow. Certainly the front office with CRM with Salesforce, etc. All these clouds are taking over front office functions, and some back-office ones too.

Reed: Then there was bet number 2, where you went against the ESB and ETL grain.

Dhillon: The principal bet was hybrid. The secondary bet was being able to tackle the connections between applications and analytics. Every business that has over 1,000 employees is struggling with this problem. When we made these investments four years ago, people looked at us like, "Really? Is there going to be a future where there are things beyond ETL, and ESB is on the decline?" But we thought the future belonged to micro-services and lighter connections.

When I joined as CEO, those were the questions even people I'd worked with for a long time were asking me. Now, they are not those questions anymore. Now, it's pretty clear that a lightweight, web services model is the right model. The old heavyweight ESB feels like, "Why did we do that?"

Are vendors falling short on SaaS integration?

Reed: There is increasing debate about the costs and complexity of SaaS integration, now that companies have multiple cloud systems to integrate. What are your views?

Dhillon: I would say there are two eras of SaaS. There was an era of SaaS where all you really had was one significant enterprise play - often Salesforce. There weren't four horsemen of SaaS. Yeah, there was NetSuite, but at the time NetSuite was very much a midmarket cloud solution. Then, you have the emergence of Workday, and ServiceNow, and a plethora of others - Concur, SuccessFactors, Cornerstone...

Now, all of a sudden, you've got a true hybrid portfolio, and you have a real headache. In phase one, SaaS integration was pretty easy. It was like a Salesforce aftermarket, and point-to-point integration got the job done.  Now, not only do you have multi0point integration needs, you have what we call massively multipoint, MMP. You've got hundreds of endpoints in the enterprise. In consumer-facing businesses, marketing along has a cloud ecosystem unto itself. This is a pretty big deal. If you look ahead to 2020, I believe we are looking at the end of point-to-point integration.

Reed: When you talk with cloud vendors, just about all of them will tell you they've got great integration services, can integrate with this or that. They offer all these reassurances to customers  - but aren't they ultimately falling a bit short?

Dhillon:  With point-to-point integration, most of the cloud vendors can deliver what they say, but in the multi-point scenarios, they really don't have it. For example, you might have a great out-of-the-box integration between your cloud CRM system and LinkedIn - but that's not the only two systems you need to integrate. How do you  wire up the rest of your sentiment analysis? If you buy an analytics cloud, you are acknowledging that there is a multi-point problem, and, therefore, partners like SnapLogic have a role to play in solving that problem.

Reed: When you're talking about multi-point, are you referring to the need to integrate with different applications, or you are referring to a different approach to integration completely?

Dhillon: I am referring to a more complicated point of integration. The issues with multi-point integration are exponential. Your headaches are a square of the endpoints. With the proliferation of data sources and cloud systems, those end points are multiplying rapidly. So how do you manage this hairball? You can connect two systems using Java or Python, but when you're connecting 17 points, you are looking at some square of that - a very large number.

All of a sudden, you're writing an integration platform by mistake, and you shouldn't be. This is when the need for a third party integration platform goes from "nice to have" to "must-have," and this is why we are seeing the heavy growth in our business. The need for a data integration platform is now no longer optional. Therefore, people are evaluating products like ours and buying them.

Adobe - doing cloud integration at scale

Reed: Give us a customer example.

Dhillon: In the middle of the year, we started doing big, strategic enterprise solutions - Adobe was one of them. Before they became a cloud company, Adobe had all sorts of middleware. They had Informatica, webMethods, and what have you. But their business was changing, and that's never easy. Their IT people described it as an "asteroid belt of SaaS" that they had to sail through. They are a cloud-first company now; everything is in the cloud.

So how do you integrate that with the on-premise work that you do, and deal with the analytics that are changing away from traditional data warehousing, and so on? We went through a rigorous evaluation, where Adobe looked at the problems that they had and where they wanted to go.

Adobe picked our product to be an enterprise backbone for app and data integration - an Integration Platform as a Service, or iPaaS. They are in the process of upgrading, or in some cases, outright replacing their legacy middleware technology with SnapLogic. Now they have one platform for bringing in new SaaS, and for connecting the SaaS they already have. In addition, like everybody else, they have huge investments in analytics, the Adobe Marketing Cloud, etc. There's a number of initiatives around data lakes in Hadoop. Those are exciting projects where we are helping them connect as well.

For us, it's really been an exemplar of an enterprise going to the cloud and managing a business model change. But it's still a hybrid enterprise; they have on-premise financials and an on-premise data center. That combination is the sweet spot for SnapLogic. For us, it's a success that we are extremely proud of, because we are doing it on a very large scale at a multi-billion dollar corporation, and for a leading light in the world of cloud.

Next week, in part two, I challenge Dhillon on why we need another acronym (iPaaS). We also get into the problem of data gravity, and how enterprises should tackle cloud integration projects.

Image credits: Dhillon photo provided by SnapLogic. Feature image - © kelly marken - Fotolia.com

Disclosure: Diginomica has no financial relationship with SnapLogic; I was approached by SnagLogic PR and though the topic was compelling. Salesforce, SAP, Workday, Oracle and NetSuite are all diginomica premier partners as of this writing.