We turned a $2M services agency into a $7M SaaS product

The hardest stretch was January through April 2024. The SaaS spinout had no revenue. The agency, which was funding the spinout, had a slow first quarter — a major client paused work for two months. We had four months of combined runway between the two entities. I spent a Sunday in February staring at our shared bank balance and trying to figure out which payroll cycle would be the one we couldn't make.

We didn't miss any payroll. But the experience cured me of any romantic notion that running an agency-funded spinout was easier than fundraising. It is in some ways harder, because the agency clients can't be told you're cash-strapped or they'll renegotiate.

§Two engineers asked to move

The second hard stretch was a cofounder argument in summer 2024 about resource flow. Two of our best agency engineers — people who'd been with us since 2018 — wanted to move to the SaaS team. The agency couldn't afford to lose them. We argued for a week. The eventual answer was: we'd let them move only if we could hire and train two replacements at the agency first, on a six-month overlap. Both engineers waited the six months. Both made the move. The agency survived; the SaaS gained two senior engineers who fundamentally changed our product velocity.

The third hard moment was emotional. By late 2024 it was clear the SaaS was going to outgrow the agency. My cofounder, who'd built the agency from scratch, started to feel sidelined. We had several painful conversations about identity, about what each of us wanted the next decade to look like. We eventually restructured equity in both entities so that he held more of the agency and I held more of Drift, with cross-ownership that protected us both.

The unlock that turned Drift from a converted-agency-clients SaaS into a real product company was a single piece of public content. In June 2024 I published a long technical post titled "Why your design system should be code, not Figma." It argued, with specific examples drawn from our seven agency engagements, that most companies were maintaining design systems in two places — Figma and code — and the bidirectional drift was costing them roughly 20% of frontend engineering hours.

The post got 84,000 reads in two weeks, with a long thread on Hacker News and a top spot on Frontend Focus. It generated about 600 inbound demo requests over the following ninety days. Drift's customer base diversified away from "former agency clients" to "people who read the post and decided to try us." That diversification was the moment we became a real SaaS company, not an agency annex.

We wrote three more long posts of similar length and depth over the next year. Two of them produced similar results. Owned content remains our largest source of inbound, beating paid ads, SEO, and partnerships combined.

§Killing the agency without killing the company

The twelve months after we made the hard call to wind down agency services were the most uncomfortable period in the company's history. We let go of seven people — five client-services managers and two senior creatives — who had been with us since 2020. Revenue dipped 31% in Q2 2025 before SaaS ARR caught up in Q4. By March 2026 we hit $7.1M ARR on the product, with services revenue intentionally capped at 12% of total (and serving only as customer-success hours for the largest accounts).

We made two product hires that decided the outcome. Devika, a former Figma product manager, joined as Head of Product in early 2025 and rebuilt our roadmap process around weekly customer interviews. Cole, an ex-Webflow growth engineer, ran our self-serve funnel and pulled trial-to-paid conversion from 8.4% to 21.7% inside six months. Net revenue retention sat at 128% by year end. We lowered our average deal size deliberately — from $48k to $11k ACV — because the smaller customers churned at a quarter of the rate. The "agency-to-product" transition rhymes with the developer-tools indie story in this cohort, where the founder learned that craftsmanship in docs and onboarding does the work a sales team would have done in a higher-ACV motion.

1. 1
   You cannot build a SaaS on the side of a billable services team.

   The economics of attention always favor the client work. Spin the SaaS out into a separate entity with separate people. Don't compromise.

2. 2
   The seventh time you build something for a client, that's the product.

   Pattern-recognize on what you keep recreating. The repeated custom build is the spec.

3. 3
   Agency clients are a great wedge of first customers, but a bad sample of the market.

   Plan to diversify away from them by month nine. If you don't, you've built a captive consultancy with a SaaS skin.

4. 4
   Cross-equity protects cofounder relationships through transitions.

   When one entity takes off and the other doesn't, the cofounder of the slower one needs structural protection. Set it up before it's needed.

5. 5
   A long, specific, technical post can do what $200k of paid ads cannot.

   It is the highest-leverage marketing channel for B2B SaaS founders who can write. Write the post.

What this looked like in practice.

The thing nobody tells you about *agency-to-SaaS transitions* is that the agency cash flow is the trap. Every quarter you choose another agency engagement to keep the lights on, the SaaS roadmap slips a month. We had to set a calendar deadline — June 30, 2025 — and stop signing new agency contracts before we built the product muscle.

The thing nobody tells you about *firing senior people you love* is that they will, eventually, thank you. Three of the seven we let go are now product leads at SaaS companies that bought Drift Studio. Two of those purchases were directly because of those alumni.

The thing nobody tells you about *lowering ACV deliberately* is that your investors will read it as decline and your unit economics will show the opposite. We had to write a long memo explaining churn-adjusted LTV before our board accepted the move.

The thing nobody tells you about *hiring from Figma and Webflow* is that those people have built self-serve funnels at a scale most B2B founders have not seen. The pattern transfer is enormous.

The thing nobody tells you about *services revenue cap* is that the cap has to be a written rule, not a soft preference. Every quarter, sales pushes to break it. We hold the line and the SaaS multiple holds with us.