Pega CEO Alan Trefler and his conflict with public cloud
- We got to sit down with Pega CEO Alan Trefler at the company’s annual user conference in Las Vegas, where it became clear that he has mixed emotions about the cloud.
Earlier this week we reported how Pega CEO Alan Trefler laid out how he believes that buyers are repeating the same mistakes they made in the past in the cloud, during his keynote at the company’s annual user conference in Las Vegas.
His argument largely centred around companies buying into the marketing buzzwords and creating siloed cloud islands, which don’t give a 360 degree view of the customer. Which is an argument I can buy into, as it’s about customers not thinking through their long-term strategies and not rethinking how their services could be designed. They’re just lifting and shifting to the cloud.
However, during a media sit-down with Trefler, we got a broader view of his conflicts with the cloud. And whilst Trefler is a smarter man than I, I’m less convinced about his view of where the world is headed. Or should I say, where buyers should be headed right now.
Trefler believes that customers want choice. He says that the cloud vendors out there aren’t selling buyers something flexible and variable, and instead are locking them into systems that leave them with no choice regarding data location. But ultimately Trefler believes that CIOs should be responsible for downtime and availability. And that businesses that are more sensitive to any downtime should be avoiding public cloud providers.
I can relate to some of these points more than others. For example, I do think that cloud promised a world free of vendor lock-in in the early days, which simply isn’t true and has proven to be unrealistic. I also think that pricing has proven to be less reliable than many customers would have hoped.
However, I take issue with Trefler’s fundamental view that the public cloud isn’t the delivery model for most modern digital services. I don’t view on-premise private clouds as having the same capabilities, simply because the scale and investment in the technology cannot be replicated in most private environments.
And whilst I would agree that his customers are probably telling him that they want private clouds, I’m increasingly convinced that this is at best a medium-term option.
Bizarre and fantastic
I should add here that Pega has absolutely invested in the cloud. It has re-architected its platform for cloud delivery and has a partnership with AWS to deliver services in the public cloud, if that’s what customers are looking for. And the company is experiencing strong growth, so you know, I could be completely off the mark here.
But I can’t help but feel that Trefler isn’t entirely convinced about the benefits of the public cloud. And whilst private clouds are a compromise, they still retain most of the problems on-premise architectures have delivered for decades.
I think that people use the cloud in all sorts of bizarre and fantastic ways, and not in a good sense, to describe a set of needs and desires, problems and solutions, that you need to talk about a little bit more specifically. People associate with the cloud a level of variable pricing, a level of freedom and a level of flexibility that honestly when you look at many of the cloud solutions are just empirically not there.
My belief is that a cloud architecture is absolutely critical and all of the Pega technology has been moved to run on a cloud architecture.
But I don’t think that the word cloud stuck on other marketing materials automatically creates the sort of placebo based panacea that a lot of people talk in terms of.
He’s correct there. Running into the cloud without thinking about how you can do things differently is just as bad as rebuilding the same thing on-premise. It should be about thinking about how you can either deliver services differently to your customer, or serve your organisation in a more frictionless way.
But Trefler is convinced by the hybrid model for most enterprises. He said:
We think that a hybrid model is appropriate for us and I actually believe a hybrid model is the most appropriate thing for most companies at scale. If I was 18 people in a shoebox I couldn’t afford to set up these sorts of things. But once you start getting companies of a certain scale, then the hybrid I think makes the most sense.
That’s the whole point though isn’t it? A company of 18 people could potentially disrupt and be competitive in a market against a company of scale *because* of the cloud.
Trefler disagrees. He said:
I think customers are very interested in some of the things that cloud promises, which is the speed of deployment, flexibility and variability, in terms of elasticity. I think customers are very interested in that and that’s why we offer a Pega cloud system, which has been proving very popular. But I also think that customers want tight integration with their systems, they want a real-time transactional environment where you can have confidence that if a debit is placed on one system that the credit will be placed on another, and other types of things that are in many cases better suited to on-premise technology.
What are the consequences?A lot of Trefler’s argument also rests on the consequence of failure. Where he sees reliability and accountability more suited to on-premise, in-house environments. He said:
A lot of it depends on the consequence of a failure. If the consequence of a failure is low, then being on the cloud can make staggering amounts of sense. That’s why a company like Netflix, which is a very well respected company, they’ve gone on to the cloud very materially because the risk of failure there is that I might have to restart my streaming session. That’s very different from from Philips (healthcare), where the consequence of a failure in that environment might actually be life or death.
And so when one is thinking of the cloud, one needs to think both of the type of cloud, because you can have clouds with low consequence of failure, but you can also have clouds like the Salesforce one that went down for 24 hours.
And by the way, the thing I just don’t get is how any CIO could tell their management, their board, after something like that happened that they had no option to move it somewhere else. This is frankly part of the cloud delusion, that somehow not just being locked into a piece of software, but being locked into a piece of software running in a certain place is okay. We wouldn’t have the nerve to tell our customers that they should have no choice but to be running our stuff in the place that we tell them. I think that’s kind of a bizarre thing. But it’s interesting, no one brings it up. Have you guys written that’s absurd? You guys know it’s absurd.
I actually don’t know it’s absurd. Especially given that there is plenty of research out there that proves that cloud providers are living up to their promise of 99.9%+ availability and that on-premise systems have historically, on average, suffered a greater deal of downtime, at a greater cost. Not only this, but I can think of plenty of examples of companies that left customers without access to their services for days, if not weeks, because their on-premise systems have failed them. These are things that are by and large avoided in the cloud.
But Trefler sees choice as the key requirement for customers. He said:
Ultimately I think that a level of optionality is something that in the world of high technology, where things change and people have different opinions, and you don’t know who is going to buy them, not having choice is a bit bizarre.
I think the fact that we offer our clients the choice is consistent with the fact that the systems that we put in tend to be very important to the businesses. I think for some systems running it in the cloud doesn’t matter because it’s probably not mission critical to the business. And I think there is a collection of systems in the middle that may become disappointed.
However, he did finish by saying:
This is not against cloud, cloud is tremendous. You’ll see us use it a lot for things we use in-house, for things we offer our clients. And for customers that run entirely on the cloud.
Not much more to add. I think that there are plenty of customers and companies out there who will disagree with me and who will agree with Trefler. And obviously there are companies out there that are regulated in certain regions to have their data on-premise.
But basically I have come across very few examples in recent months where I thought: “That would be better on premise”.