For large enterprises laden with technology debt, disparate systems, and siloed data, it’s important to have a strong and effective enterprise architecture function to guide the organization to a better technology and data fabric. But many attempts at enterprise architecture fail due to a variety of factors. In my conversation with Claus Torp Jensen, chief technology officer (CTO) at CVS Health, I discussed lessons and ideas on how to build an effective architectural function.
A: It’s a good question. We have to revisit and question the notion of enterprise architecture because I think historically it’s been about setting standards and direction. I found in my CTO role—and even before that as a chief architect—that architecture needs to be much more. The slogan for my team is that “we turn vision into action.” We are the only place where different visions and needs and different IT programs come together. We are the only place that has a broad view which turns insight into actions.
A: I think you have to accept the fact that you cannot stay forward-looking only. One of the things we discussed internally in my team is that doing what is right is not always the right thing to do. Let me explain. There’s an idealistic view of what you should do and there’s a practical view of what you can do here and now. Historically, enterprise architecture is heavily focused on what is right, and they have not always managed to balance the different forces at play in a given situation or decision.
If you look at many people’s definition of governance, it all boils down to control. I have a different philosophy—I believe that governance is about helping the right people make the right decisions at the right time for the right reasons based on the right information. If you take that approach to governance, you get a very different result.
In the spirit of turning vision into action, it is our responsibility that the company we work for becomes holistically integrated, that we can holistically understand the consumers that we serve and ultimately enable our business partners to fundamentally transform our products. That is what we are here for and why you need a strong architecture function in a modern organization. It’s not because of standards, it’s because you have to make a difference to the business.
A: I did not create an enterprise architecture function to be part of the planning life cycle only. If you want to do what I described above, you need to go beyond that. So, there are a few things. I embarked on a five-year culture and talent transformation journey. If you take the Forrester spectrum of “strategize > influence > translate > shape > execute,” we are squarely in the middle part. You can’t focus on everything. In today’s day and age, it’s tough to drive a top-down process. I chose a middle road in how we interact with the organization.
It involves three things:
A: I don’t think you can solve that through organizational design. I think it has to be done through a combination of organizational thinking, talent, and culture evolution. The architecture team can make decisions and provide guidance, but for them to stick and be sustained, a variety of other factors come into play. You have to understand the business imperatives and strategy, you have to generate trust, you have to influence, you have to communicate well. You need to know how to interact with executives, you need to know when to talk and when not to talk and just listen.
All those skills have nothing to do with your technical acumen. You can be the greatest architect in the universe, but if you do not understand people, you will never be as good as someone who may be less sharp than you technically but is better at understanding and interacting with people.
A: I believe that the majority of the architecture resources and mindshare should go towards shaping and executing. In my CTO organization we have six focus areas:
It’s a mistake to think of these in terms of distinct functions and departments. They need to work in synergy, complementing each other depending on the priority and nature of the question at hand. You have to think of them as different dimensions of the same problem.
For example, the architecture forensics team is a powerful team with a combination of such skills as business architecture, technology architecture, data architecture and security architecture that can be dropped in to diagnose and solve gnarly problems. Finding these skillsets in a single individual is hard. You have to think in terms of stitching together a team with an effective amalgam of such skills.
As another example, the technology research and innovation activities track about 300 emerging technologies and capabilities. The experimentation part is easy. The hard part is making sure that you have actionable insights from these experimentations that allow you to decide what makes sense for the business at large. Ultimately, you want to provide advice on how these emerging technologies can help transform the company.
A: It’s an interesting question as to whether it’s the organization that loses patience or whether it’s the patience and stamina of the leaders who take on these missions that wears thin. My conjecture is that it’s the latter. People take on these missions without appreciating or understanding that you’re making a multi-year, often five-year commitment on this change. If you’re not willing to stake your career on it and stay true to the five-year journey, then you shouldn’t start it.
A: This problem is only addressable one way: Through developing a very patient and very deliberate plan, and then spending time, political capital, and resources to get to a higher quality data fabric for the company. It is hard. It requires ensuring authoritative systems of record are established, it’s about improving data conformance and integration. You need a strong business sponsor to back you.
I have often seen chief data officers get into roles and assume that data will magically get there. It does not. It is through a thoughtful partnership between the architecture function and CDO equivalents that we can improve the data fabric of companies. Bad data means bad understanding, which leads to bad direction in terms of driving differentiation through technology transformation.
A: I go back to my earlier comment that doing what is right is not always the right thing to do. You have to train the team to understand what “good enough” means. It’s not a term that we spend enough time on. You should have a North Star for the architecture and then enough programs and projects to get you there.
But what’s in between? That’s not well thought-through. There has to be a solution-direction statement for major capabilities and programs. Most companies don’t have that. We often select a set of platforms for business capabilities and then declare them global for the company. We do not delineate solution design from integration design.
We can only get good data designs and data coherency through a thoughtful data strategy that gets executed as a common thread across programs and projects, as part of the day job.
Author: Krishna Cheriath