Enterprise architecture heats up to meet changing needs

Enterprise architecture practitioners and tools continue to shift focus to address the changing needs of businesses facing financial pressures, regulatory and compliance mandates and, lately, the coronavirus pandemic.

In this interview with TechTarget, Gordon Barnett, a principal analyst at Forrester Research, discusses the latest trends, tools and challenges emerging in enterprise architecture (EA) -- including a transition to agile practices and outcome-based approaches and a collaborative ecosystem of experts rather than a single dictatorial group.

In what ways have enterprise architecture practices evolved over the past 15 years, and what are the latest trends?

Gordon Barnett: When CIOs first wanted to create an enterprise architecture practice, they nearly always set up only a technical architecture team. The CIO was interested in proving to management that he could run IT like a business, so he looked at the technology landscape and costs and efficiencies. He was trying to portray the architecture as a portfolio of technology assets to enable the business to do things. Today, 50% of Forrester clients still operate in that era. 

Gordon Barnett

The second wave of EA maturity was where the architecture was perceived as a portfolio of business capabilities. A capability is a configuration of assets -- human capital assets, process assets, data assets, technology assets. You can add intangibles like [intellectual property] and finance assets. In this area, you started to see the EA team grow. You had information architects, process architects, business architects. The question these people were answering is: Are we investing in the right capabilities? In current research, 40% to 45% of Forrester clients operate in that area.

Where we are today -- because of all the digital transformation and the customer obsession -- is outcome-based architecture. You're now configuring and optimizing the business capabilities to deliver an outcome. This has introduced a whole breed of new architects. You start hearing people calling themselves a 'people architect,' a digital architect, a platform architect, an experience architect, an organization architect. About 3% to 5% of our clients are in that space at the moment.

Beyond outcome-based architecture, what other trends are you seeing emerge in enterprise architecture?

Barnett: We're seeing a move to build an ecosystem of subject matter experts rather a single fixed team of smart people. As you mature, you realize the best way to do enterprise architecture is to have a platform where more people can get together, discuss things and take many architectural views on the problem statement. They agree on a way forward, rather than having a dictatorial architecture group. To a large extent, people in financial services and insurance are ahead of the curve and more innovative in this approach.

The next one is the role of EA has changed because of people being more agile. In traditional EA, all of the frameworks -- [Federal Enterprise Architecture Framework] FEAF, [National Overall Reference Architecture] NORA, [The Open Group Architecture Framework] TOGAF -- tell you to have an architecture review board. Now you have trusted governance rather than an architecture review board. And you tend to have EA providing services, whereas EA used to be in project teams and architecting projects.

What are the greatest difficulties organizations face in meeting their enterprise architecture needs?

Barnett: Skills is definitely one of the biggest challenges at the moment. Most people are making the decision to expand their EA, or start an EA if they haven't had one, and they just move in people from one box to another. Just because you can code software doesn't mean you can think like an architect. If you are a systems engineer, you know the processes and systems, but it doesn't mean you can do capability modeling and things like that.

One of the biggest barriers to EAs moving forward is ROI.

When it comes to tools, one of the biggest barriers to EAs moving forward is ROI. The reason it's hard to come up with an ROI is because people don't do activity-based accounting. They don't identify how long they spend doing all of their tasks. If they had that information, they could say, 'I can save this amount of money if I automate these things.'

The other big barrier is that people on the business side are now tech-savvy, and they question the need for EA. They don't want EAs telling them to use certain technology. A lot of the business [leaders] are now thinking, 'IT is just a cost center. I want [IT] to be an order taker.' So, a lot of EAs are being constrained in the value they can deliver because the businesses don't want them involved in collaborating with the business.

Do most organizations use enterprise architecture tools?

Barnett: In January, 68% of Forrester clients said they do not have an EA tool. They nearly all start off with Microsoft products -- Visio, Excel spreadsheets, Word documents, maybe PowerPoint. And they're always struggling because there is no assurance that what is in the document actually reflects what is being put into production. Nobody goes back to change the project documentation when the environment changes. Therefore, you have point-in-time views.

They later move to something like open source, or they build a database themselves. Then they move to an EA tool, but the EA tools are not cheap. The average probably starts off at $30,000 a year. It's hard to build an ROI on that. This is why many of the EA vendors are now a lot broader than EA. They've gone into portfolio management strategy or application portfolio management.

What's the current state of enterprise architecture tools?

Barnett: There are 54 vendors trying to sell to enterprise architects. This is an enormous market. The main players are only looking at the big organizations, because they believe they have the funds. As you move down, you tend to find that small and medium-sized organizations are using open source drawing tools rather than commercial software packages. But, if you really want to be a mature EA practice, you need a commercial off-the-shelf package. Today, the enterprise architecture management suites support only some of the things that EA needs. Therefore, you need an ecosystem of products to be able to do your job.

The vendor market is always changing. There's lots of consolidation through mergers and acquisitions and lots of new players coming onto the market. They're all taking it from different angles. Some might try and do all of it, but they will be strong in one area. There might be one set of EA tools that are big on portfolio change and capability modeling and another that is strong on technology asset management or strategy development.

Years ago, it was all about modeling and capturing data. Modeling is largely automated now. A lot of the professional EA tools are using AI, natural language, image recognition, speech, analytics, etc. to help with data collection. So, EAs can spend more time on the architecture rather than data entry.

How effective are the enterprise architecture tools?

Barnett: Every year when we do our surveys, we ask what people like and don't like. The biggest complaint about tools is usability. You've got to remember that these tools are built by architects for architects. So, if you're not an architect, and you've not been trained to think like an architect, these are not the tools for you. But the big EA tools are becoming more and more user-friendly.

What you need in your tool is the ability to see the relationships between building blocks, the dependencies, and understand where the pain points are. It's important to be able to run algorithms against your models to identify the weak spots in your architecture. Without the tool, you have gut feel. To make an informed decision, you really need to have the science behind it. The value of EA tools is in the insights they give.

What drove the spike in the number of EA tools?

Barnett: Let's go back to the last financial crunch. As soon as disaster hits, everybody wants to know where they can cut costs. They don't have anything that brings everything together and shows them what is causing the cost. It's simple to say, 'Oh, let's cut 30% of people and 30% of technology.' That doesn't mean you can support the business. You need better insights. So, that was the primary driver.

Just before that, there was a focus on best practices -- things like Sarbanes-Oxley, that actually insisted that you have an enterprise architecture, and the insistence by the Office of the CIO in the U.S. government that every agency comply with the Federal Enterprise Architecture Framework. The Saudis have NORA. Then you have people getting together behind The Open Group Architecture Framework.

Even now, with COVID, things have changed. People have said, 'We need to know how we change our architecture to support remote access and all of the consequences of that, like the security risks.' This is why enterprise architecture has grown. You need to be able to have some representation of the impact if you move technology assets. If you don't have the tool, with up-to-date, reliable information, you have 20 people in a room trying to whiteboard and work it out.

How has the coronavirus pandemic impacted enterprise architects and enterprise architecture?

Barnett: They focus on surviving, adapting or growing. Depending on the company's strategy, EA will revise its mandate. For surviving companies, the EA focus will be on optimizing costs and efficiencies based on current technology assets.

Adapting company's EAs will focus on reprioritizing investments -- for example, speeding up the move to cloud. They are looking to adapt to the environment, maybe introduce collaboration technology and so on. Growing companies will invest more in innovation so that the company can be better positioned when the markets pick up. This may mean EA focusing on automation, AI and other emerging technologies.

The interview was edited and condensed for length and clarity.