Case study: five SaaS tools, one internal app, and a saner ops team

Retool's 2026 Build vs. Buy report says 35% of enterprises have already replaced at least one SaaS product with custom software, and Newsweek's coverage points to AI-assisted development as the reason the math finally works. That's the macro story. Here's a smaller one.

What the client had

A 40-person services company, roughly $12M in revenue, running client delivery across five tools:

- A project management SaaS for tasks and timelines - A separate time-tracking tool that fed their invoicing - A CRM they used mostly as a client directory - A shared spreadsheet for capacity planning - A form builder for intake and status updates

None of it talked to the others. Their ops lead spent about a day a week reconciling numbers between the time tracker and the PM tool because tasks lived in one system and hours lived in another. Every new hire needed five logins, five onboarding walkthroughs, and a printed cheat sheet for which tool was the source of truth for what.

Combined subscription cost was around $2,100 a month. Not the biggest line item. But the tax on people's time was the real problem.

What we built

One internal web app. Projects, clients, tasks, time entries, capacity, and intake forms in a single data model. Roughly 12 weeks from discovery to production, delivered in weekly increments they could actually use.

A few decisions worth naming:

- We kept accounting where it was. QuickBooks stayed. We built a one-way sync for approved invoices. No reason to rebuild what already worked. - We didn't build a CRM. They didn't need pipeline management. They needed a client directory with contact history. So that's what they got. Two screens, not twenty. - Capacity planning stayed simple. A weekly grid showing who's booked on what. No resource-leveling algorithm. The ops lead wanted to eyeball it, and eyeballing it was working.

Infrastructure runs on managed Postgres and a small container host. They don't touch it. We handle backups, updates, and monitoring.

What changed

The reconciliation day is gone. New hires get one login. The ops lead told us the thing she noticed most wasn't the time savings, it was that she stopped having the same three arguments every month about which number was right.

Subscription costs dropped to under $400/month for the remaining tools they actually use. The build paid for itself in about 14 months on software savings alone, and that's before counting the reclaimed time.

The part worth taking away

They didn't replace SaaS because SaaS is bad. They replaced it because five tools designed for five different companies didn't add up to one tool designed for theirs. The trend in the Retool data is real, but the reason to do this isn't the trend. It's whether your team is spending more time feeding the tools than doing the work.

If that sounds familiar, a discovery call is free. We'll tell you honestly whether a build makes sense, or whether you're two integrations away from being fine.