The Business Runs on What Nobody Wrote Down

The diligence covered everything countable. Contracts and renewal dates, revenue by client, margin by service line, the tool stack, the org chart, three years of financials. Nobody pulled a ticket from eighteen months ago to see whether the fix was written down anywhere.

That is the part of the company that decides whether it can absorb anything. A new technician. A bad quarter. An acquisition. The Friday resignation of one person who has been there since the beginning.

So here is the argument in plain words. You bought a business whose real operating system is what its people remember. It works, in the sense that tickets close and clients renew. It does not scale, because every new hire and every acquired company has to be trained by someone whose hours you are already selling to a client.

The industry admits the gap in an odd place

Kaseya asked 1,061 MSPs last November what makes winning new customers hardest. Difficulty creating and keeping consistent client documentation climbed from 10 percent to 17 percent in a year.

Look at where that answer showed up. Not in a question about internal tidiness. In the sales question. When a prospect asks how you will run their environment, a provider whose method lives in four people’s heads has nothing to hand over. The undocumented shop is slower on Tuesday and it also loses the deal on Thursday.

Automation stopped part way, and everybody is comfortable

The same survey asked how much of the workload MSPs had automated over the past two years. Nine percent added none at all. Most of the rest added a little: 55 percent automated up to a quarter of the work. One percent are anywhere near fully automated.

Then the number that explains the other ones. Fifty-nine percent say they are satisfied with the automation they have.

Satisfaction is the right feeling if the company stays exactly this size. It is the wrong feeling for the plan you underwrote. A shop that automates a quarter of the work and calls it done has decided, without saying so, that growth comes from adding people. That is the plan you are paying a multiple to change.

What automation actually moved

The same MSPs were asked which numbers improved as a result of automating. First response time led at 35 percent. Then customer satisfaction, technician efficiency, fewer errors, less burnout. Profit margin came near the bottom at 18 percent, and billing accuracy sat below that.

That ordering is worth an hour of your next operating review. Written-down, automated process buys capacity first and margin later. It shows up as work absorbed without hiring, and only after that as a better P&L. If the value creation plan books the savings in year one, the plan is wrong about the mechanism, not just the timing.

The knowledge has legs

Service Leadership, the benchmarking arm ConnectWise owns, reports that technicians with one to three years of experience leave at three times the rate of technicians with eight or more.

Put that next to an undocumented shop and the problem states itself. The senior people are load bearing. The junior people, the ones who would eventually take that load, leave before they get there. Every departure in the first group takes a piece of the operating system with it, and every departure in the second one wastes the training you paid for. Meanwhile the company pays a retention premium to keep a filing system that happens to be a person.

Two ways to run a platform, and no way to price either

Ask what the acquirers do about this and the honest answer is that they disagree.

Evergreen Services Group, one of the largest buyers in the market, has said plainly that it does not push its companies onto shared systems. Its head of acquisitions put it this way to a trade outlet: “We’re not mandating solutions, tech stack, things like that.” New Charter Technologies makes a similar promise to sellers, and says so on its own site.

Others integrate. Ntiva moved an acquired company onto one shared professional services platform, and the consultant who ran that project describes roughly sixty people working on it daily for about a year. Magna5 appears in its software vendor’s own case study putting nine acquired units on one platform in as little as two months.

Same task, two months against a year, and no one publishes what either cost. There is no timeline here you could underwrite. There is not even a benchmark for the underlying question: nobody publishes what share of an MSP’s work is documented, and nobody publishes tickets per technician or time to resolve for this industry either. You cannot buy that number. You have to measure it inside the company you already own.

The work, starting this quarter

Take the ten accounts that produce the most revenue and test them. Could a technician who has never touched that client run a week on what is written down today. Not in principle. Actually, on a Tuesday, with the person who normally handles it unreachable.

Count the automated work honestly before you count the money. Ask what percentage of tickets close without a human touching them, and compare that to what the plan assumed when it promised margin from efficiency.

Decide whether you are standardizing or leaving them alone, and write the decision down. Both models are running at scale in this market. Drifting between them is not a third model, it is an expensive absence of one.

Put documentation in the purchase agreement, not on the wish list. The window when a seller will still write things down closes the day their attention moves to what comes after the deal.

You bought recurring revenue, a client list, and a set of habits nobody has ever described out loud. The first two transfer on the closing date. The third one is the reason the plan slips. One question for the next operating review costs nothing to ask: if the best technician resigns on Friday, which clients feel it on Monday, and what leaves the building with him.