The Midmarket EA Gap
There is a gap in how enterprise architecture is practiced, and it maps almost perfectly to company size. Large enterprises with thousands of employees have dedicated EA teams, established tooling, and a seat at the strategy table. Small companies do not need formal EA — their landscape is small enough to hold in one person's head.
In between sits the midmarket — organisations with 200 to 5,000 employees, somewhere between 50 and 300 applications, and exactly one person who is supposed to keep the portfolio under control. This is where the gap lives.
The Invisible Architect
In most midmarket organisations, the person responsible for enterprise architecture does not have that title. They are the IT Director, the Head of Infrastructure, the CTO, or occasionally an enterprise architect who also handles security reviews, vendor management, and half the project governance.
This person knows the landscape. They know which systems are critical, which ones overlap, and where the technical debt is worst. But they carry that knowledge in their head, in scattered spreadsheets, and in slide decks they update before the quarterly IT review.
The knowledge exists. The structure does not. And when that person leaves or moves to a different role, the institutional understanding of the portfolio leaves with them.
Why Enterprise Tools Do Not Solve This
The standard advice when a midmarket IT leader wants to get serious about EA is to evaluate the enterprise platforms — Ardoq, LeanIX, MEGA. These are serious tools with serious capabilities. They are also designed for a different context.
An enterprise EA platform assumes you have a team to implement it, a metamodel strategy before you start entering data, budget for a multi-month implementation, and an ongoing capacity to maintain the repository. In a midmarket context, none of those assumptions hold.
The result is predictable. A midmarket company buys an enterprise tool, spends three to six months on implementation with consulting support, gets a partially populated repository, and then watches it go stale because nobody has the time to maintain it.
The tool is not the problem. The mismatch between the tool and the context is the problem.
What the Midmarket Actually Needs
A midmarket EA capability needs to solve five specific problems, and nothing else.
Inventory: a structured, maintained list of applications with enough metadata to classify, score, and report. Not a comprehensive metamodel — a practical inventory that one person can keep current.
Classification: a way to sort the portfolio into actionable categories. TIME (Tolerate, Invest, Migrate, Eliminate) is the industry standard because it maps directly to decisions. But classification without strategic context is just colour-coding.
Capability alignment: the bridge between what the business needs and what IT provides. A capability map does not need to be exhaustive. A two-level map that connects business functions to the applications that support them is enough to start having useful conversations with leadership.
Health visibility: a dashboard or report that answers the questions leadership actually asks. How much do we spend? Where is the risk? What is duplicated? What should we retire?
Decision support: when someone asks "should we keep or replace this system?", the answer should be derivable from data in the tool — not from a two-week analysis.
Everything beyond these five — process modelling, data architecture, integration mapping, stakeholder views — is valuable but not foundational. You can add it later. You cannot justify it first.
The 60-Minute Test
There is a practical test for whether an EA tool fits the midmarket context. Can a single person go from zero to a usable portfolio view in under 60 minutes?
That means importing or entering the application list, getting basic classification in place, and seeing something that looks like an answer to the question "what does our landscape look like?"
If the answer is no — if the tool requires metamodel configuration, custom field setup, integration work, or consultant time before you can see your own data — it is not built for the midmarket.
The right tool for this context is opinionated. It makes structural decisions for you. It assumes you do not have time to configure and will never have a dedicated EA team. It trades flexibility for speed-to-value.
The Cost of Doing Nothing
The alternative to proper EA tooling is not chaos. Midmarket IT leaders are competent people. They manage portfolios with spreadsheets, slide decks, and institutional knowledge.
The problem is what happens at decision points. When the board asks about cloud migration readiness. When a merger requires landscape integration. When a new CTO arrives and wants to understand the portfolio in their first month. When the person who held it all in their head changes roles.
At those moments, the lack of structured, maintained portfolio data costs real time and real money. Not because the organisation is incompetent, but because the knowledge exists in a form that cannot be queried, shared, or built upon.
The gap is not a tool gap. It is a structure gap. And the right response is not a bigger tool — it is a tool that fits the reality of who will actually use it.
Atlas is built for organisations with one EA or IT leader, 50-300 applications, and no time for six-month implementations.