How we 3.8x'd testimonial completion rates by removing every login screen
We rebuilt our customer testimonial flow around a single browser link — no app, no signup. Completion rates went from 12% to 46% in eight weeks. Here's the playbook.
TL;DR
Removing the login wall from your testimonial request flow is the single highest-leverage change you can make. In our cohort of 230 SaaS teams, completion rates jumped from a median of 12% to 46% — a 3.8x lift — within eight weeks.
Eighteen months ago, our testimonial pipeline at Proofly looked like every other SaaS team's pipeline: send an email asking for a quote, hope the customer replies, follow up three times, give up. The hit rate hovered around 4%, and the quotes we got back were polite, generic, and unusable.
We knew video would be more persuasive — every landing page test confirmed it. The blocker wasn't appetite. It was friction. So we ran an experiment: rebuild the flow around the assumption that every step beyond "click and record" loses 30% of customers. This is what we shipped, what broke, and what worked.
The five-step flow we shipped (and why every step matters)#
The new flow has a hard ceiling: five steps from link click to upload finished. We arrived at five by working backwards from a 60-second target — anything more than that and completion rates fell off a cliff. Here's the sequence:
- Open link. No tracking interstitial, no cookie banner, no marketing copy. The page is the recorder.
- Grant camera and mic. Native browser prompt, nothing custom.
- Read three prompts. Pre-written by the founder, one screen, no scroll.
- Record. One button. Tap to start, tap to stop.
- Upload + consent. Auto-uploads in the background while the consent checkbox loads.
Steps 1–3 are silent — the customer doesn't have to type anything. Step 5 is the only place a form appears, and it's a single checkbox.
The numbers, before and after#
We measured 230 teams across 11 weeks. Each team ran the legacy flow for two weeks, then switched to the no-login flow for nine. The lift was consistent across team size, ARR band, and industry vertical.
| Metric | Legacy flow | No-login flow | Lift |
|---|---|---|---|
| Median completion rate | 12% | 46% | +283% |
| Time from send to completion | 4.2 days | 18 hours | −82% |
| Audio-only opt-ins | n/a | 18% | new |
| Day-30 reuse on landing pages | 31% | 67% | +116% |
The biggest surprise was the time-to-completion drop. We expected the same number of customers to complete, just faster. Instead, the speed itself drove more completions — customers who would have abandoned the email entirely on day three never got the chance, because they finished on day one.
What we removed (and what we'd never remove again)#
Cutting the login was the obvious win. Less obvious: we removed a thank-you screen with a CTA after recording. Customers were closing the tab before reading it, so we replaced it with a plain "Saved. You can close this." line. Engagement on the post-recording email — the one we send 24 hours later asking for a tweet — went up 22%.
The one thing we tried to remove and immediately put back: the consent checkbox. We tested making consent implicit ("by submitting you agree…"). Two customers asked us to take videos down within a week. The friction of one extra checkbox is worth a hundred of those conversations.
The shortest path to a great testimonial is the one that respects your customer's three free minutes. Everything else is a tax.
Implementation notes for your own stack#
If you're not using Proofly, you can still build a version of this in a weekend. The pieces:
- A public route that takes a slug and renders the recorder — no auth required.
- The browser's
MediaRecorderAPI — supported in Chrome, Safari 14.1+, Firefox, Edge. - An upload endpoint that accepts a chunked POST and stores to S3 or R2.
- A consent record stored alongside the upload (timestamp + IP + user agent + the surfaces named).
The hard part isn't recording — it's resuming uploads on flaky Wi-Fi. We use a chunked upload with exponential backoff. About 8% of recordings on mobile drop their connection mid-upload; with retries, only 0.4% fail to land.
What to ship this week#
If you want to copy the playbook, here's the smallest version that works:
- Pick your three best customers. Send them a single link, no email template, just "Would you mind recording 45 seconds about [specific outcome]?"
- Use Proofly, Testimonial.to, or roll your own with
MediaRecorder. Whatever you ship, make sure it's one click to start recording. - Publish the first three on your homepage above the fold within 48 hours. Don't wait for ten — three is plenty.
- Measure the conversion lift on the page and decide whether to keep going. (You will.)
The TL;DR for your slides: friction is the variable, and login is the largest unit of friction in the flow. Cut it, and everything else gets easier.
Frequently asked
Quick answers
How long should a video testimonial actually be?+
Aim for 45 to 90 seconds. Anything shorter feels promotional; anything longer rarely gets watched in full. The sweet spot in our data is 62 seconds, which produces the highest replay rate on landing pages and the most-quoted clips for case studies.
Do I need a release form before publishing a customer's video?+
Yes — but it doesn't need to be a separate DocuSign. A timestamped, IP-stamped checkbox in the recording flow that names the publishing surfaces (website, social, ads) is legally sufficient in the US, EU, and UK. Proofly stores this consent automatically with every submission.
What if customers don't want to be on camera?+
Offer audio-only as a fallback in the same link. About 18% of completed testimonials in our network are audio-only, and they convert nearly as well as video on landing pages — likely because the bar to visual polish is removed.