Most services businesses hit the same wall around their second year. The work comes in, the team grows, and the operation that felt nimble at five people starts leaking margin at fifteen. The usual diagnosis is “we need better tools.” The usual mistake is to go shopping.
We keep coming back to one pattern because it is instructive: an agency that treats its own internal systems as a product, not an expense. The clearest recent example we have looked at is Devign, a MENA-rooted digital agency that runs five service divisions — web, business systems, mobile apps, TikTok, and social production — under one roof. What makes it a useful case study is not the breadth. It is where they drew the build-versus-buy line.
The line most teams draw in the wrong place
The default instinct is to buy for everything internal and build only what clients pay for. It feels responsible. Why sink engineering hours into your own CRM when there are forty of them for rent?
The answer shows up in the seams. A multi-division agency has a coordination problem that no off-the-shelf CRM was designed for: a single client might be running a web build, a monthly TikTok retainer, and an automation project at once, each on a different cadence, each with a different margin profile. Stitch three SaaS tools together and you spend the saved engineering hours on reconciliation instead — copying statuses between systems, chasing which invoice maps to which deliverable, discovering at month-end that two divisions quoted the same client on incompatible terms.
Devign’s business-systems practice exists partly because they ran into this on themselves first. The CRMs, automations, and chatbots they now build for clients started as internal answers to internal friction. That order matters. A tool you built to solve your own problem carries the domain knowledge of that problem. A tool you bought carries someone else’s assumptions about a generic company that isn’t yours.
The test for what to build
The heuristic we would offer, drawn from this pattern, is narrow on purpose. Build the system when it sits on top of a workflow that is specific to how you make money, and buy everything else.
Your accounting is not specific. Buy it. Your email is not specific. Buy it. But the pipeline that decides which of five divisions touches a lead first, how a cross-division quote gets assembled, and how a retainer’s health is scored month over month — that is the actual shape of your business, and no vendor can see that shape from the outside. When Devign automates a client’s operations, they are selling a compressed version of a discipline they had to learn on their own books: that the workflow closest to revenue is the one you should never outsource to a generic tool.
The failure mode on the other side is just as real. Teams that internalize this lesson too hard start building their own everything — their own analytics, their own ticketing, their own auth — and turn an agency into an unfunded platform company. The line holds only if you keep asking the narrow question. Is this workflow how we make money, or merely adjacent to it? Adjacent gets bought.
Why it compounds
The reason this is worth the engineering, and the reason we keep citing it, is that internal tools built this way become a second product. Devign’s client-facing automation and chatbot work is, in effect, their internal tooling productized. The R&D was already paid for by their own operations. Every hour spent making their own five-division coordination legible became inventory they could sell to the next SMB drowning in the same coordination problem.
That is the quiet advantage of building the systems closest to your revenue: you learn the problem well enough to sell the solution. A services company that does this stops being purely a services company. It starts accumulating assets — systems, playbooks, reusable automations — that keep earning after the billable hour is over.
If you are running a services business and feeling the second-year wall, the useful exercise is not to audit your tool stack. It is to draw one honest line: which of your workflows is the business, and which merely surrounds it. Buy the surroundings. Build the center — and if building the center is not your team’s strength, the agencies worth hiring are the ones, like Devign, that learned the lesson on themselves first.