SaaS Customer-Journey & Usability Review — Conversion Audit for an Early-Stage B2B Messaging Product
Context
An early-stage B2B SaaS product — customer messaging for small, owner-operated businesses (cafés, salons, repair shops) in a bilingual EU market. The founder had built something genuinely opinionated: AI drafts every reply, but a human must approve each message before it sends. No autopilot. That’s the product’s core promise.
The ask was simple and honest: walk through the product as a first-time visitor would — from landing on the site to the moment of deciding whether to sign up — and give candid, expert feedback on what’s clear, what’s confusing, and where the value proposition lands or doesn’t.
The catch: the target buyer is a non-technical small-business owner with ten free minutes between customers. Not a PM. Not a developer. Someone who fears exactly one thing — a robot talking to their customers in their name.
My approach
A review like this fails if it’s just “I clicked around and here are my opinions.” So I structured it.
Three personas, three lenses — not three parallel reviews. I built three realistic small-operator personas and gave each one a job:
- The primary persona (a salon owner on a laptop) carried the full journey — landing, demo, exploration, the sign-up decision. Her fear: losing control of what gets sent.
- A mobile-first persona (a car-detailing studio owner who never opens a laptop) stress-tested the same journey on a phone — surfacing where the mobile experience degrades.
- A language-trust persona (a restaurant owner serving customers in a language she doesn’t fully read) existed for one sharp question: can you confidently approve a reply you can’t fully verify?
One controlled variable per persona. When the mobile journey broke where the desktop one didn’t, the difference was attributable — not a vibe.
Method stack: first-impression (5-second) test · Nielsen’s 10 usability heuristics · stage-by-stage journey mapping with emotional states · activation-funnel analysis · GDPR/ePrivacy cookie check · severity-rated findings (P1 critical / P2 important / P3 minor), each with a concrete recommendation.
Two deliverables, two perspectives. A 15-minute think-aloud Loom recorded in persona — the lived journey, hesitations included — and a written report from the PM perspective: findings, severity, evidence, recommendations. The video shows what a real visitor feels; the report shows what to do about it.
What I found
Eleven findings: 2 critical, 3 important, 6 minor — plus a set of small polish notes.
Without exposing the client’s specifics, the pattern:
- A critical context gap in the core daily-use flow. The screen where owners make their most frequent decision was missing the one piece of information needed to make it — forcing a workaround that, at volume, would push busy owners toward approving without reading. That quietly breaks the product’s central promise.
- Conversion blockers at the decision moment. Two competing calls-to-action at the exact point of sign-up, neither explaining what happens next — and the product’s single best selling point (you inherit a ready-made, editable workspace instead of building from scratch) was never stated where the decision happens.
- A compliance gap that contradicted the product’s own positioning. The product sells on EU data governance; the site’s cookie handling didn’t match that promise. Not a legal catastrophe — a brand contradiction, found exactly where a technically literate buyer would look.
- The product’s strongest ideas were its least visible ones. The mandatory human-approval model, the real (not simulated) trial workspace, the founder’s opinionated philosophy about response-time ergonomics — all genuine differentiators, all nearly invisible on the journey to sign-up. My headline recommendation wasn’t “fix what’s broken.” It was: make your best thinking visible at the moments that decide trust.
Every finding came with evidence (screenshots, timings — the “aha” moment was reachable in under two minutes, which is genuinely good) and a specific, lightweight fix. Where I couldn’t verify something from the outside, the report says so explicitly — an honest boundary beats a confident guess.
Outcome
- The client shipped a fix for one finding before the report was even delivered — I’d flagged it early, they acted on it mid-review.
- Full scope delivered in 3 business days: written report, prioritised fix list, think-aloud Loom, and answers to the client’s questionnaire.
- The engagement closed with 5.0 feedback:
“Working with Kate was really an eye opener for us in terms of how to look at the product. We changed so many things after her review and detailed report she made us! She is highly recommended and we look forward working with her again.”
“We changed so many things” is the point of a review like this. Not a list of complaints — a plan the founder actually executed.
Where this work lives
This case exists as the artifact on Upwork – SaaS Customer-Journey & Usability Review — B2B Messaging Product (EU) .
Delivered a 15-min think-aloud Loom + written report with 11 severity-rated findings and a prioritised fix list — in 3 business days.
The client shipped one fix before the report was even delivered, and closed with 5.0★: “We changed so many things after her review.”
Why this works
Most usability feedback fails in one of two ways: it’s either a vague vibe check (“feels confusing”) or a 40-item nitpick list that buries the three things that matter. The fix for both is the same — structure, severity, and honesty about what was and wasn’t verified.
I catch expensive product mistakes before they become development waste. This is what that looks like in practice.
Need a candid, structured review of your product’s journey — before your users find the problems for you? I’m open to contract and fractional PM/PO engagements.
Leave a Reply