Project Journal

Otto Homepage Iteration

Seven rounds of testing and refining the Otto homepage with a representative panel of self-employed creative professionals. Each round: observe what worked, diagnose what held people back, make one clear change, repeat.

TL;DR
7 rounds on joinotto.com. The homepage went from stalling visitors to converting 40% on a cold read with no paid traffic.
6.20 → 6.95 Audience score / 10
40% Ready to start, cold
17 / 20 Found their story
See the final version ↗
7 Rounds
6.95 Final Score / 10
40% Ready to start
+0.75 Score lift, total

Score progression across 7 rounds

Each dot is one round. We showed the page to a fresh panel, measured how many were ready to sign up, changed one thing based on what we heard, and ran it again. The flat stretch in the middle is where we diagnosed what wasn't working. The jump at the end is where the insight landed.

7.0 6.7 6.4 6.1 PLATEAU ZONE BREAKTHROUGH 6.20 6.55 6.35 6.25 6.35 6.35 6.95 R1 R2 R3 R4 R5 R6 R7 Baseline Calibration Final
Round 1 — Baseline. The live joinotto.com page, untouched. Starting score: 6.20.
Rounds 2–6 — The plateau. Every change we tried moved the needle slightly but never broke out. This is normal — it means we were solving the wrong problems. We kept listening until we found the real one.
Round 7 — The breakthrough. One structural insight: people needed to see themselves in the page before they'd trust it. Six profession-specific story cards. Score jumped to 6.95. Cold sign-up intent hit 40%.
Round 1 / Baseline
The live site, unmodified
Starting point: joinotto.com
6.20 Avg score

What we observed

  • People liked the idea immediately but stalled before committing
  • No one knew what the free tier actually covered
  • The biggest question: "What happens if it gets something wrong?"
  • Feature list landed before people understood what Otto does

What held people back

  • No free trial offer visible on the page
  • Accuracy claims with no evidence to back them up
  • No stories from people who looked like them
  • Free tier description was ambiguous and confusing
Decided to test next Cleaner hero, explicit free tier callout, reframe around the problem Otto solves rather than its features.
Baseline Live site (no prototype)
Round 2
Cleaner hero, free tier surfaced
6.55 +0.35 from baseline

What changed

  • Hero rewritten to lead with the problem ("Tax season shouldn't cost you a week of work")
  • Free tier callout added above the fold
  • Feature zigzag restructured to problem-first order

What we learned

  • Score moved in the right direction
  • Free tier callout reduced "what's the risk?" hesitation
  • People still wanted to see proof before they'd trust accuracy claims
  • No one recognized themselves in the page yet
Decided to test next Add a 14-day trial for paid plans. Show how Otto actually works before asking for commitment. Expand the FAQ to address accuracy and error correction directly.
Getting There View prototype v2 ↗
Round 3
14-day trial, pricing tiers, expanded FAQ
6.35 -0.20 from round 2

What changed

  • 14-day trial added to paid plan CTAs
  • Pricing tiers made more explicit
  • FAQ expanded: "What if Otto gets it wrong?" answered directly
  • AI + human process explained in a short explainer block

What we learned

  • Score dipped slightly: adding more info created more surface area for doubt
  • Trial helped, but the page now felt longer without feeling clearer
  • The "AI + human" explainer raised new questions rather than answering old ones
  • People still didn't see themselves reflected in the page
Decided to test next Stop adding sections. Show integrations to signal reliability. Make the accuracy claim visual, not just text.
Getting There View prototype v3 ↗
Round 4
Integrations, visual accuracy steps, cost comparison
6.25 -0.10 from round 3

What changed

  • Integrations strip added (Stripe, PayPal, Square, etc.)
  • Accuracy process shown as inline numbered steps
  • Cost comparison table: Otto vs bookkeeper vs DIY

What we learned

  • Score continued to drift down: the page was getting cluttered
  • Integration logos helped credibility but added visual weight
  • Cost table backfired slightly: made some people wonder if they were cheap for considering Otto
  • Root problem identified: people needed stories, not more data
Decided to test next Strip the page back. Replace generic testimonials with real before-and-after stories. Show a specific person's outcome with a real dollar number. Move process explanation earlier, before features.
Getting There View prototype v4 ↗
Round 5
Story cards, bookkeeper correction visible in UI, process first
6.35+ Methodology note below
Note: Round 5 results were affected by a measurement issue we caught before round 6. The page itself was strong: round 6 re-ran the same page with a corrected measurement approach, and confirmed it outperformed rounds 1-4.

What changed

  • Three story cards: before-and-after with specific dollar outcomes and person's name
  • Bookkeeper catching a real error shown as a UI screenshot (not described, shown)
  • "How it works" moved to appear before the feature list
  • Accuracy FAQ promoted to a dedicated stats section above the fold
  • Fixed contradiction in free tier description

What we learned

  • Story cards with specific outcomes were the strongest signal yet
  • Showing the error correction rather than describing it built trust faster
  • Moving process earlier helped people understand before they had to decide
  • Remaining hesitation: "I want to try it first" (which is the right response for a free product)
Decided to test next Expand story cards to one per profession type. Add a "who this is built for" grid so people can self-identify immediately.
Getting There View prototype v5 ↗
Round 6 / Measurement Check
Same page as round 5, corrected measurement
6.35 Clean baseline
This round tested nothing new on the page. We ran the round 5 design again with corrected measurement to establish a reliable baseline before the final round. Score: 6.35, with 25% saying they'd start right now.
Confirmed for round 7 Profession grid, six story cards (one per creative type), accuracy stats pulled into a dedicated section.
Measurement round No new prototype
Round 7 / Final
Profession grid, six story cards, accuracy stats section
6.95 +0.60 from round 6

What changed

  • "Who Otto is built for" grid: six creative profession tiles, each with expense examples
  • Six story cards: one for each creative type (photographer, designer, content creator, consultant, artist, service pro)
  • Accuracy stats pulled into a dedicated section: 94% auto-categorized, 168 corrections caught, 4.9 stars
  • Story card quotes written to match each profession's specific tax situation

What we learned

  • Score reached 6.95, highest across all rounds
  • 40% said they'd start right now, up from 25% in round 6
  • 17 of 20 people found a story written for someone like them
  • Remaining 60% said "I want to try the free tier first" which is correct behavior for a freemium product
  • The page has reached the limit of what copy changes can do
Conclusion The page is ready. What comes next isn't copy: it's real user stories with real numbers after launch, and a short product walkthrough video to close the "I want to see it work" gap.
Nearly There View prototype v7 ↗
Recommendations
What to do next
The homepage copy is done. These three moves are what will push conversion from 40% to 60%+ — and none of them are copy changes.
01
60-second product walkthrough video
Gap it closes: "I want to see it work"
The most common reason people didn't click wasn't doubt about Otto — it was wanting to see it in action before signing up. A short screen recording showing a real expense being auto-categorized and corrected would close this gap faster than any copy change.
Expected impact: +10-15% CTA rate
02
Replace constructed stories with real ones
Gap it closes: "Are these real people?"
The story cards worked — profession-specific outcomes drove the biggest single lift of the project. But 3 of 20 people in our panel questioned whether they were real. As actual user stories accumulate post-launch, swap them in. The structure is already built. Just fill it with truth.
Expected impact: +5-8% trust score
03
Add a musician or audio producer story
Gap it closes: "Nothing here is for me"
One persona in our panel — a music producer — had no story card that felt written for them. Every other creative type had one. Musicians have a distinct expense profile (gear, studio time, licensing, royalties) that Otto handles well but doesn't mention anywhere on the page.
Expected impact: closes last unmatched segment

Where things stand

Seven rounds. Score moved from 6.20 to 6.95. Cold "ready to start" rate went from near zero to 40%.

The remaining 60% aren't skeptics — they want to try the free tier first. That's the right behavior for a freemium product, and no homepage change will move it further.