DevOps and Digital Quality of Service at Concur

Enterprise cloud customers expect great product functionality combined with a high level of service quality and excellent user experience.


Dec 03, 2016

Enterprise cloud customers expect great product functionality combined with a high level of service quality and excellent user experience. Historically, on-premise software suppliers focused only on functionality without paying much attention to user experience.

Ensuring performance, resiliency, availability, and user experience require the provider to understand how users consume their service.

Kevin Evans, Vice President of Cloud Services at Concur, explains how his organization uses DevOps to meet these challenges. He also discusses the role data in making sound decisions about the product and service.


Michael Krigsman: I'm Michael Krigsman from CXOTalk. I'm here at Future Stack '16, and thanks to New Relic for inviting me. And I'm talking with Kevin Evans, who is the VP of Cloud Services at Concur. Hey, Kevin.

Kevin Evans: Hey, how are you doing Michael?

Michael Krigsman: Good. Thank you so much.

Kevin Evans: Good.

Michael Krigsman: So, you're VP of Cloud Services at Concur.

Kevin Evans: Yes.

Michael Krigsman: What does Concur do?

Kevin Evans: So, Concur is the world's leading travel and expense management provider for enterprise customers. We have about 40,000 customers, and about 35 million users that use the service on a regular basis to help them book and manage travel; as well as create, submit, and manage expenses as the manager.

Michael Krigsman: So, are you building cloud at scale?

Kevin Evans: We are. We definitely are. The business has really grown substantially over the time that I've been there, which is about 12 years, and we're expecting that we'll continue that growth path.

Michael Krigsman: Now, I know that DevOps is also a key part of what you do. So tell us about that.

Kevin Evans: Yeah. So, for us, I think that the key message around DevOps is something that we actually call it "end-to-end," because it's really about giving responsibility to the teams, the product teams, the engineering groups - to have compete ownership of the services they're designing and building from start to finish. And that start to finish obviously starts with the capabilities and functionality of the software. But the finish is "How do we run the software? How well does it perform? How well does the service actually behave for the end-user?" And so, the end-to-end philosophy for us is really about giving the controls, but also the responsibility and accountability to each of those teams. [It's] about making sure that the service is delivered to the end-user the way the end-user expects it.

Michael Krigsman: So the end-user experience, then, is the reference point that your engineering teams use.

Kevin Evans: It is. It's extremely important for us to consider the end-user and the experience that they have with using the service in everything that they're doing.

Michael Krigsman: How is this different from traditional approaches to engineering?

Kevin Evans: So, I think if you were to look back at how we used to develop software at Concur, it was really about focusing on the product and the functionality, and the requirements were heavily weighted towards "How is the functionality going to work for the end-user?" And I think what you will look at today and the major difference with how we build and deliver software is it's really more just as much about the service as it is about the functionality. And so, how well is it going to perform? How resilient is it? How is the availability of the service going to come across for the end-user?

Michael Krigsman: What kind of data are you collecting, and how is the data the glue between the end-users and the engineering division, and the organization?

Kevin Evans: We collect a tremendous amount of data. Sometimes, we refer to ourselves as a data collection company that just happens to do travel and expense management because the amounts of data that we're collecting about how the service is used, how the user interaction is happening, what the experience is like for the end-user ... Coming back and narrowing that to when there are challenges: What did the infrasturcture look like? What's happening on the servers when those things are happening? What's happening with the code when the experience is bad? Or in many cases, what's happening when the user experience is good?

Michael Krigsman: So there's a drive to get inside, into the user from this data.

Kevin Evans: Yeah. So the only way that we have found that we can successfully make the right decisions from an engineering standpoint about how the servers are going to work is to be evaluating how the service is working in real-time. And so, that real-time aspect of the data, the fact that we're collecting so much information about user experience as they're using a service. And then we need to leverage that in our design process.

Michael Krigsman: What about the role of mindset and culture in all of this as well?

Kevin Evans: It's a real change from a traditional model where you're used to (and I think this is no longer [the case]) ─ I think it was the previous process that a lot of companies went through in developing software. You focus on the functionality. I think it's now about focusing on the service itself, and that goes to not just the development engineers who are building it, but from a product standpoint [of]: How are we delivering it? How do we upgrade it? How do we manage it? And that's user experience. It goes all the way back to the very first part of the very front part of the process.

Michael Krigsman: What are the challenges associated with making this kind of change?

Kevin Evans: Yeah, so there's a very complex set of systems that run the service and a very complex set of users that are from different variants, as far as where they're at in the world, what types of systems that they're using, whether they're using a tablet, whether they're using a browser. There's a lot of complex data that has to come together. And we need to coalesce all of that information, and make it meaningful to a broad set of users that have different requirements, as well as different experiences about working with data. So, it's really a challenge for us to make sure that we're presenting useful information that can be used to make decisions, as opposed to just presenting data.

Michael Krigsman: So figuring out the right type of data, what's actually going to be meaningful in terms of making engineering decisions that will impact the user positively.

Kevin Evans: Absolutely, absolutely. That's the key. Anyone can collect a lot of data. It's what you do with it, and how you turn it into information so you can actually make good, intelligent decisions from that data.

Michael Krigsman: So then, what advice do you have for others who might be doing the same kind of thing?

Kevin Evans: Yeah, so I think that the key is really about getting that mindset that software is only part of the equation. It's really about the entire service, and looking at that from a perspective of: How does all this data come together? And how do you make use of it?", and putting systems and processes in place so that data can come together and users of all different domain expertise, whether they be in the product side, or they be in the user experience side, or whether they be in the development engineering side can now actually put all of it together and make sense of it.

Michael Krigsman: Fantastic! We've been talking with Kevin Evans, who is the VP of Cloud Services for Concur. Kevin, thanks a lot!

Kevin Evans: Thank you!

Published Date: Dec 03, 2016

Author: Michael Krigsman

Episode ID: 401