· Software
Six in Ten Technical Staff Already Built Software IT Never Approved. That Is the Real Build vs Buy Question.
Build versus buy assumes someone sits down and chooses. At most companies, someone already built it, and nobody decided anything.
Most build versus buy advice assumes you are standing at a fork in the road, weighing a purchase order against a sprint. Retool's 2026 Build vs. Buy Shift survey of 817 technical builders says that is not how most of this actually happens anymore. Sixty percent of respondents built software for work in the past year without IT's knowledge. A quarter of those did it regularly. Of that group, 31 percent gave the same reason: building it themselves was faster than waiting on IT to provision anything. Thirty-five percent of teams had already replaced at least one SaaS tool with something built in house, mostly workflow automations, internal admin tools, BI dashboards, and CRMs. Seventy-eight percent plan to build more in 2026.
That is a survey from a company that sells the tools people use to build these things, so read the enthusiasm for "build more" with that in mind. But the 60 percent figure cuts against Retool's own interest, not for it. A vendor pitching build-it-yourself tools has no reason to highlight that most of that building is happening in the dark. They did anyway, because the number is real and worth naming.
My position is that the build versus buy decision most businesses think they are making has already been made, piecemeal, by whichever employee got tired of waiting on a ticket. The actual question left on the table is not whether to allow this. It is whether anyone priced in what happens after the tool ships.
Here is the honest case for buying instead, because it is a real one. A SaaS vendor serving thousands of customers absorbs the patching, the security monitoring, and the infrastructure failover that a 40-person company would otherwise have to invent from scratch for a tool three people use. That is not a regulatory argument. It is a market one: specialization and scale genuinely beat a one-off build for anything that is not core to how you make money, and pretending otherwise just because building got cheaper is its own kind of overreach. Fifty-one percent of Retool's respondents shipped production software their teams actually use, and about half of those report saving six or more hours a week. That is a real number, not a marketing claim. It also says nothing about who keeps that tool running once the person who wrote it changes teams, which is the part the survey does not answer and most build-vs-buy pitches skip past.
Vendor guides love to claim maintenance eats 60 to 80 percent of a custom tool's lifetime cost. I could not trace that figure to an actual study, only to other vendor blogs repeating each other, so I am not going to hand it to you as fact. What I will say, from having run this comparison for clients, is that the maintenance bill is the one line every build estimate lowballs, and it is also the one SaaS pricing already has baked in, badly as that math has gotten this year.
So the test is not "is this cheaper to build." It is three questions, asked before anyone opens an editor:
- Is this workflow a commodity every company needs, or does it sit inside what actually makes you money? Commodity, buy it.
- Who owns this for the next three years, by name, with real hours budgeted for it, not just this sprint?
- Does IT or security know this exists before it ships, not after an audit finds it?
A tool that clears all three is a legitimate build. A tool that fails question two or three is not a build decision at all. It is a liability that happens to run, regardless of how little it cost to stand up.
This is the order our custom software team actually runs it in: differentiation first, ownership second, governance third, cost last, because a free build with no named owner costs more in year two than an honest SaaS renewal ever will. If someone at your company already shipped something useful that nobody above them knows about, that is worth finding out this month, not during your next security review. Talk to us if you want a second set of eyes on which column it actually belongs in.
Sources
References used in this article. Links also appear alongside the relevant claims.
Let's make it happen.
You don't need a polished brief. A couple of lines about where you want your company to go is plenty, and we'll come back with what we'd tackle first.
We scope the work around your goals and timeline before quoting anything, so you know exactly what you're getting.