From Zero to 35K Users
at $0 Current Paid Acquisition
I sold Jobsolv as a hands-on service before building the software—approximately $58K+ in founder-reported services revenue—then vetted 40+ dev agencies, directed a 27-person roster across the build, and reached a founder-reported 35K+ users with current paid user-acquisition spend at $0 in July 2026.
The Non-Scalable Thing,
Done First
Jobsolv is an AI-powered recruitment platform — ATS-ready resume tailoring and automated applications for professionals chasing $100K+ roles. But it didn't start as software. It started as me, doing the work by hand for paying clients, in a market crowded with venture-funded competitors burning cash on paid acquisition.
No outside funding, and the July 2026 snapshot recorded $0 current paid user-acquisition spend after earlier paid tests were stopped. The operating question was whether validation discipline, procurement judgment, and earned distribution could support the next stage.
Service → Vetting → Product →
Distribution
Validation discipline
Before writing a line of software, I ran Jobsolv as a white-glove service: 26 clients at $2,000–$3,000 each and approximately $58K+ in founder-reported services revenue. Referrals were one recorded acquisition source. Just as important, the service surfaced product requirements a survey might have missed: some clients wanted to approve every application, others wanted full autopilot, and many needed recalibration on what the market pays. The selected cohort informed the SaaS; it did not prove the later product would reproduce the service outcomes.
Procurement judgment
Only after the service proved demand did I invest in software — and I treated hiring a dev shop like a procurement analysis, not a leap of faith. I interviewed 40+ agencies, collected quotes, and required itemized, feature-by-feature time and cost estimates instead of one inflated bundle. The red flags repeated: sales closers who couldn't answer technical questions while their engineers sat silent; 'what's your budget?' pricing that mirrors your number back at you; well-reviewed shops that quietly de-prioritize smaller clients when bandwidth runs out. Sitting through those interviews taught me more about buying software than any framework.
Zero-to-one product leadership
I directed a 27-person cross-functional roster across the build, with up to 22 active in a week—a development agency, a separate UX/UI design agency, and specialist freelancers. I wrote PRDs, feature documentation, and acceptance criteria, ran standups, and reviewed delivery against the spec. The product combined ATS-ready instant resume tailoring with auto-apply based on needs observed during the service phase. First-party records show over 20K applications processed; that is usage, not a placement-rate claim.
Distribution over budget
Growth used SaaS directories and AI aggregators, useful Reddit contributions in career communities, SEO/AEO/GEO content for traditional and AI-assisted discovery, university and career-coach partnerships, a Product Hunt launch, and press releases syndicated through EINPresswire. In Jobsolv's first-party acquisition data, some channels produced attention without corresponding signups. We kept the channels associated with recorded acquisition and cut the ones that only produced applause — even the ones I'd been sure would work.
PRISM Applied
at Startup Scale
The same evidence discipline I developed in-house—applied to my own bootstrapped product after early paid acquisition tests were stopped.
Behavioral Diagnosis
First-party journey analysis suggested that trust signals and perceived effort reduction were stronger hypotheses than feature-count messaging, so I prioritized them for testing; the available evidence does not establish universal conversion drivers.
Impact Prioritization
Prioritized experiments by lifetime-value impact. Founder analytics associated first-session resume tailoring with stronger retention, so it became an activation hypothesis. The cohort window and denominator are not published, so the page does not present a reusable lift benchmark.
Hypothesis & Execution
Ran 50+ growth experiments across onboarding, activation, and retention. Each test had a named behavioral mechanism and a predicted impact.
Revenue Measurement
Measured each experiment against acquisition, activation, and retention. Every readout ended with a clear scale, stop, or learn decision; no generic win-rate benchmark was used as proof.
Compound & Scale
Readouts that met their decision rules could move to rollout and follow-up measurement. Founder-reported acquisition records associated signups with referrals, SEO/content, directories, and product-led channels; they do not isolate each channel’s causal contribution.
"Every stage of Jobsolv was a bet I refused to make blind: I sold the service before I built the software, interviewed forty agencies before I hired one, and cut every growth channel that produced applause instead of signups — including the ones I was sure would work. Ego is the most expensive line item in a startup. Evidence is the cheapest."
See all case studies
Explore the enterprise experimentation work at NRG Energy, or read weekly breakdowns on Lean Experiments.
Make Growth Decisions
You Can Defend
Know what to test, when to trust the result, and what to do next. Practical decision guides for analysts, growth teams, and founders.
Opens Substack to confirm · Free · Unsubscribe anytime
Read Better Decisions
Practical guides on experimentation, analytics, commercial judgment, and using AI without outsourcing the decision.
Open Substack (opens in new tab)