Case study: the CRM a 40-person tech company actually needed

According to Retool's 2026 Build vs. Buy Report, 35% of enterprise teams have already replaced at least one SaaS tool with a custom build, and 78% expect to build more of their own tools this year. Newsweek's coverage frames it as a quiet shift away from packaged software toward tools that fit the business.

We've been watching this happen one client at a time. Here's a recent one.

The situation

A B2B software company, roughly 40 people, seven-figure ARR, selling into mid-market operations teams. They had a well-known SaaS CRM and had been on it for three years. On paper, everything worked. In practice, three things were true:

- Sales reps kept a parallel spreadsheet because the CRM didn't model their deals well (multi-stakeholder, long procurement cycles, usage-based pricing). - Their RevOps lead spent about a day a week reconciling the CRM against the spreadsheet, Stripe, and a product analytics tool. - Forecasting meetings ran on gut feel because nobody trusted the pipeline numbers.

They weren't looking to replace the CRM. They asked us to build "a dashboard" that pulled the systems together.

What we found in discovery

The dashboard wasn't the problem. The CRM's data model was. It was built for a generic sales motion, and this team's motion wasn't generic. Every workaround, every custom field, every Zapier automation was a patch over that mismatch.

We counted 14 fields the team actively used and 60+ they ignored. Two integrations were paid for and unused. The reps hated logging calls, so they didn't.

What we built

A purpose-built pipeline tool. Not a CRM replacement in the marketing sense. A CRM replacement in the practical sense: it did the six things this team actually needed, and nothing else.

- Deal records shaped around their sales motion, including a stakeholder map and a procurement-stage tracker. - Two-way sync with Stripe and their product analytics, so usage and revenue lived next to the deal. - A forecasting view that showed weighted pipeline against historical close rates by stage, not by rep guess. - Call logging via a simple email-forward workflow. Reps stopped complaining.

We run it on managed infrastructure. They don't think about hosting, backups, or uptime. They open the app and it works.

What changed

Six months in: the RevOps lead got roughly a day a week back. Forecast accuracy against actual close numbers moved from "we don't really track it" to within about 12% over the last two quarters. They canceled the old CRM and two adjacent tools. Net software spend went down. That wasn't the goal, but it's a nice side effect.

The takeaway

Most teams don't need a better CRM. They need a tool shaped like their business. If your team is keeping a spreadsheet next to a SaaS tool, that spreadsheet is telling you something. It's worth listening to.