Product Designer · Sendible · 2022–2026
Rebuilding a SaaS funnel converting five times below benchmark
Homepage, pricing, demo and sign-up. Eighteen experiments over four years.
Overview
- Role
- Product Designer — design and experimentation across the website funnel
- Timeframe
- 2022–2026
- Team
- Head of marketing · head of content marketing · growth marketing · external development agency. Copy, SEO and campaign strategy sat with colleagues; design and experiment execution sat with me.
- Scope
- Homepage, pricing, demo, sign-up, product pages, blog
Sendible’s homepage converted at 0.78% to its primary call to action. The B2B SaaS benchmark is 3–7%. Over four years I designed and ran eighteen experiments across the funnel alongside our marketing and content leads, built the scoring framework the team used to decide what to test, and learned the same lesson repeatedly: the metric that’s easiest to move is rarely the metric that matters.
The problem
The homepage drew 25,000–30,000 views a month and converted 0.78% of them to a click on its primary CTA. Site-wide, sessions converted to trials at 0.96%. Against a 3–7% benchmark, this wasn’t a rounding error — it was the largest single lever in acquisition.
The diagnostic was worse than the headline.
99.22% of visitors scrolled past the hero without taking any action. The interactive product demo — the most technically ambitious module on the page — recorded one click across 9,515 impressions. The customer logo strip recorded two.
But visitors who did click a CTA converted to trial at 20.82%, and those who submitted a form directly converted at 35.5%.
The page wasn’t failing to convert people who engaged with it. It was failing to get anyone to engage.
Then the finding that reframed the entire programme. The best-converting element on the page was a customer story module at 46% conversion — sitting at scroll position nine, where fewer than 15% of visitors ever arrived.
We had been optimising the top of a page whose most persuasive content was buried at the bottom.
How we decided what to test
With one designer, one copywriter and an external agency, we could run perhaps one meaningful test a month. Prioritisation was the work.
I built a scoring framework on impact, confidence and ease. The confidence scale is the part that mattered:
5 — quantitative evidence, and we’ve won with this pattern before
4 — quantitative evidence; this fix is our best hypothesis
3 — qualitative evidence
2 — best-practice argument only
1 — opinion, including leadership’s
Scoring confidence explicitly changed how the team argued. A director’s hunch and a junior’s hunch both scored 1. Disagreements stopped being about seniority and became about evidence.
Every test then went through the same template: hypothesis, data rationale with the source report cited, primary metric, secondary metrics, and guardrail metrics — the things we agreed we would not damage in pursuit of the primary.
I also wrote the growth-driven design process the team worked to, covering discovery, design, development, launch and a three-day post-launch review. It outlived every individual project it was written for.
What I learned, three times over
Engagement is not conversion
The interactive demo generated real engagement above the fold and produced no signups.
Rather than removing it, in February 2025 I moved it below the fold and wrote the hypothesis down: a lower position would produce fewer interactions but more meaningful ones, from people who had already read enough to be interested.
By mid-2026 it was recording one click per 9,515 impressions. I removed it in the July restructure.
Two tests, eighteen months apart, one conclusion. The module was interesting to build and interesting to use, and it never once produced a customer.
Conversion rate is not impact
The RE/MAX customer story converted at 46% — the highest rate of any module on the page — and sat where under 15% of visitors reached.
The fix wasn’t to improve it. It was to move it. A module converting at 46% for 15% of traffic is worth less than a mediocre one everybody sees, and no amount of optimising the module itself would have changed that.
More trials is not more revenue
Two experiments, the same lesson.
Defaulting the pricing page to monthly billing produced a better customer conversion rate — 6.76% against annual’s 4.48%. Annual produced $2,111 in MRR against $1,480, and $132 average revenue per account against $78. The variant that converted worse made 43% more money.
Then, in July 2026, the homepage hero test.
The experiment that decided nothing, told properly
We tested a hero variant leading on “unlimited users” against control, running 9–20 July 2026.
At day eleven I posted this to the team:
Across 8,300 views and 55 contacts, A is at 0.70% page-view-to-contact and B at 0.62%. Last week’s early lead for A has largely dissolved. Honest position: at this traffic level and effect size, this test is unlikely to reach statistical significance in any reasonable timeframe. So rather than waiting on the conversion rate alone, we’ll decide by the end of the week what the contacts are actually worth. A page that converts slightly less but brings in better-fit contacts is the one we want.
We then segmented the contacts by the plan tier each visitor selected before signing up.
| Plan tier | Control (A) | A % | Variant (B) | B % |
|---|---|---|---|---|
| Elite | 2 | 4.3% | 0 | 0.0% |
| Premium | 2 | 4.3% | 2 | 5.4% |
| Plus | 28 | 59.6% | 24 | 64.9% |
| Core | 8 | 17.0% | 4 | 10.8% |
| No value | 7 | 14.9% | 7 | 18.9% |
| Total | 47 | 37 |
My closing note:
B did shift the mix the way we hoped — a higher share toward Plus, and entry-level Core signups cut roughly in half. That’s a real signal that the “Unlimited Users” message lands with multi-user teams. Two things went the other way. A captured 2 Elite-tier trials and B captured none, and at the top of our pricing a couple of accounts matter more than a percentage point of mix. A also brought in more contacts overall. B attracted a slightly better-shaped audience, but A attracted more of them, including the biggest ones.
We kept control.
The test didn’t produce a winner. It produced a better question: were we optimising for the number of trials, or the value of them?
What shipped
July 2024 — homepage redesign. Launched as an A/B test on 23 July and stopped early on 2 August because the new design was clearly ahead. By October, bounce rate was down 39% and page views up 26%, with search impressions up 22% and average keyword position improved from 66 to 62.
October 2024 — product pages. Bounce on the Analytics page fell to zero; the Scheduler page fell from 84% to 46.6%.
December 2022 — pricing. Demos booked from the pricing page doubled, from two a week to four.
July 2024 — pricing. Scale-tier trials rose from 2 to 11.
July 2026 — homepage restructure. Six changes drawn from a 33-idea audit: removed the interactive demo and the customer logo strip, relocated social proof above the fold, added an objection-handling block, and rewrote CTA language for lower commitment. Shipped 24 July 2026. I left the business before the results matured, and I’m not going to claim numbers I never saw.
The wall: the sign-up page
Both optimised entry points — homepage and pricing — funnel into the same sign-up page. It abandons 79.54% of the people who reach it. 6,902 clicks, 20.46% completion.
My hypothesis was that visitors arriving from pricing had already chosen a plan, and the sign-up page didn’t acknowledge it — so a panel confirming the selected plan should reduce second-guessing at the final step.
Two things stopped it.
Technically, our CMS’s dynamic content couldn’t execute inside the iframe the sign-up page ran in. Our development partner proposed a workaround rewriting the iframe source from URL parameters, which was viable.
Organisationally, the sign-up page belonged to product, not marketing. Marketing owned the conversion target, and didn’t own the page that determined it.
That’s the part I’d handle differently. I made the technical case and I made it to the wrong room. What the constraint deserved was a shared metric between marketing and product, agreed before the work started rather than discovered at the end of it.
What didn't work
In mid-2024 I designed two replacements for the demo booking page and tested them against the existing one.
| Visits | Demos booked | Conversion | |
|---|---|---|---|
| Control | 147 | 30 | 20.4% |
| Variant 1 | 155 | 21 | 13.5% |
| Variant 2 | 111 | 17 | 15.3% |
Both of mine lost to a page I’d considered dated. We reverted.
The sample was small enough that the result wasn’t conclusive, but the direction was consistent across both variants, and I had no hypothesis for why the existing page was outperforming — which was itself the finding. I’d redesigned it on the assumption that it looked old, not on evidence that it converted badly.
What I'd do next
At 25,000–30,000 monthly views and sub-1% conversion, most tests on this page cannot reach significance in a sensible window. The answer isn’t more tests — it’s fewer, bigger swings, judged on contact quality rather than conversion rate, with plan-tier segmentation built into the analysis from the start rather than reached for when the headline number disappoints.