Enterprise Architecture Management Suites

Forrester published The Forrester Wave: Enterprise Architecture Management Suites, Q4 2024, naming its Leaders.

Within weeks, three of those Leaders became one company. Bizzdesign, MEGA International, and Alfabet, the Software AG product, combined under the Bizzdesign brand at the start of 2025.

Forrester wrote about it directly, noting that all three were Leaders in its recent evaluation and running a simulation based on its own Wave scores which concluded that the combined platform would, on paper, outplay every other competitor in the market.

An analyst firm modelling what happens when three of its own Leaders merge is not something that comes up often. For anyone who built a shortlist from that evaluation, three options collapsed into one before the procurement finished.

What these suites do, and the credibility problem

An enterprise architecture management suite holds a model of the organisation: the business capabilities, the applications supporting them, the technologies underneath, the data flowing between them, and the projects changing all of it.

The value proposition is decision support. Which applications duplicate each other. What breaks if this system is retired. Which capabilities have no adequate technology support. What the estate will cost in three years. Where the technical debt is concentrated.

Enterprise architecture as a discipline has spent decades fighting a credibility problem, and the tooling inherited it. The caricature is a team producing elaborate diagrams that nobody outside the team reads, describing a target state that never arrives, using a modelling notation that requires training to interpret.

The caricature is unfair and it is not baseless, and Forrester named the problem precisely in the subtitle of its 2017 evaluation: vendor capability was increasing while stakeholder experience remained an issue.

That sentence captures the category's central difficulty. The tools got better at modelling. They did not get better at being useful to the executives, product managers, and engineers who were supposed to consume the output, and a model nobody reads is documentation rather than decision support.

Seven evaluations, and a collapsing criteria set

Forrester has scored this market repeatedly since 2011, and the criteria count tells the maturity story unusually clearly.

The Q2 2011 Wave, authored by Henry Peyret with Tim DeGennaro, evaluated ten vendors against eighty nine criteria. Mega International, Troux Technologies, Software AG, and alfabet led, with Forrester noting the market was still being defined and that several vendors including Metastorm, IBM, Casewise, and Avolution were transitioning from earlier generation EA tools.

The Q3 2015 edition, authored by Gordon Barnett, evaluated ten against forty two criteria: Avolution, BiZZdesign, BOC Group, IBM, Mega International, QualiWare, Orbus Software, Software AG, Sparx Systems, and Troux Technologies.

The Q2 2017 edition covered ten against forty criteria, adding Dragon1, Future Tech Systems, Planview, and UNICOM.

The Q1 2019 edition expanded to twelve, introducing Ardoq, Atoll Technologies, LeanIX, and QPR Software.

The Q1 2021 edition covered twelve against twenty eight criteria, with erwin and Adaptive joining.

The Q1 2023 edition covered twelve, including ValueBlue as a Contender with the maximum score in the roadmap criterion.

Eighty nine criteria to twenty eight over a decade. That is a market where most of what these products do stopped differentiating them, and the evaluation narrowed to where the differences actually are.

Forrester's framing in 2019 identified what those were: as traditional EA use cases became commoditised, the vendors offering use cases vital to transformation would lead, specifically strategic performance management, technology asset performance, operational insights, and people architecture modelling.

Modelling the estate became table stakes. Connecting the model to decisions about strategy, cost, and organisation became the differentiator.

The consolidation, and what it means

Forrester predicted continued acquisition activity in a 2024 landscape report, and the market delivered.

The 2025 three-way combination is the most dramatic instance, but the pattern runs through the whole lineage. Troux was absorbed into Planview. LeanIX was acquired by SAP. erwin passed through Quest. Several 2011 and 2015 names no longer exist independently.

Forrester's analysis of the three-way merger identified the logic clearly: the products were largely complementary, the geographic coverage did not overlap heavily, and the combined research and development capacity would be substantial.

That is the acquirer's case and it is sound. The buyer's position is more complicated, and worth stating plainly.

Complementary products means three code bases, three data models, and three customer bases. Integration takes years, during which roadmaps slow while engineering capacity goes into consolidation. Customers of two of the three products eventually face a migration to whichever one survives as the strategic platform.

There is also a competitive consequence. Forrester's own simulation concluded the combined entity would outplay every competitor on paper. A market where one vendor is that far ahead is a market where pricing discipline weakens, and buyers renewing in three years may find the negotiating dynamic has changed.

None of this argues against the combined product, which may well be excellent. It argues for asking specific questions during any evaluation in the next two years: which platform is strategic, what the migration path looks like for the others, what is committed contractually, and what happens to your licence terms at renewal.

Why the discipline is having a moment

Enterprise architecture has gone through cycles of relevance, and several forces are currently pushing it back up the agenda.

Technical debt has become a board-level topic rather than an engineering complaint, and quantifying it requires exactly the estate model these tools hold.

Artificial intelligence deployment has made the estate question urgent again. An organisation deciding where to apply AI needs to know which processes exist, which systems support them, where the data lives, and what integration is realistic. Those are architecture questions, and organisations without a current model are answering them by guessing.

Regulation has added a harder driver. Operational resilience requirements in financial services, notably in Europe, require firms to identify critical business functions, map the technology supporting them, and demonstrate they understand the dependencies. That is an architecture repository requirement expressed as a supervisory obligation, and it converts a nice-to-have into an evidenced deliverable.

And the application portfolio question has sharpened as software costs rose. Knowing which applications overlap, which are unused, and which could be retired is a direct cost lever, and it is the use case that most reliably funds an EA tool.

The problem no tool solves

The recurring failure in this category is not capability. It is currency.

An architecture model is accurate on the day it is built and decays continuously afterwards. Applications get added, systems get retired, integrations get built, and organisations reorganise. If maintaining the model depends on people remembering to update it, it will be out of date within a year and useless within two.

This is why automated discovery matters more than modelling elegance. A tool that populates itself from configuration management databases, cloud accounts, identity providers, and software asset management is maintaining a model. A tool that requires architects to keep it current is maintaining a document.

Forrester's emphasis on technology asset performance and operational insights among the differentiating use cases points at the same thing from the value side. Those use cases require live data rather than a periodically refreshed diagram.

The related failure is scope. An architecture team that models everything produces something comprehensive and unusable. One that models the twenty capabilities and two hundred applications that actually matter produces something people consult. The second is harder, because it requires deciding what does not matter and defending that decision.

Which returns to the stakeholder experience problem Forrester named in 2017. The tools have improved considerably since then, with better visualisation, role-specific views, and integration into the systems where other people work. Whether the underlying practice has improved as much is a question the tooling cannot answer, and it is the one that determines whether the purchase produces anything.

Forrester does not endorse any vendor named here, and tier placement should not be read as a recommendation to buy.