The problem
Pitchkit had four ways to create a pitch, but the creation page didn't make any of them feel easy. Methods were listed as flat text links with brief descriptions. Users had to read, compare, and decide, all while staring at a page that offered no sense of momentum.
The data told the story: 72% of users who reached the creation page started a pitch, but only 38% finished one. The mid-process abandonment was brutal, especially for URL generation (Pitchkit's fastest path) which only 12% of users even tried. Most defaulted to the blank editor, got stuck writing from scratch, and quit.
What I changed
1. URL generation became the hero
The page now opens with a single dominant card: paste your website URL, pick your funding stage, hit Generate. A lightning bolt icon, a clear input field, and a "Takes about 30 seconds / Creates all sections" footer. No ambiguity about what happens next.
The three alternative methods (story, guided, simple) sit below a "or choose another method" divider in a three-column grid. Visible but clearly secondary.
2. Animated progress messages replaced the loading spinner
The old experience: click Generate, stare at a spinner for 30 seconds, wonder if anything is happening.
The new experience: a sequence of status messages that rotate every 3 seconds:
- Analyzing your website...
- Extracting key information...
- Crafting your pitch narrative...
- Finalizing your sections...
Each message pairs with a spinning icon. Users see the system working through a visible pipeline instead of an opaque wait.
3. Card-based layout with color-coded methods
Each creation method became a distinct card with its own accent color, icon, and hover state:
- URL generation: primary blue, Zap icon
- Tell your story: orange, MessageSquare icon
- Guided builder: indigo, BookOpen icon
- Simple editor: slate, FileText icon
Cards respond to hover with border color shifts and shadows. Each has a "Get started" link with an animated arrow. The visual hierarchy is immediate. You know what each method is before reading a word.
4. Survey-driven "Recommended for you" badges
After completing the onboarding survey, the recommended method displays a sparkle badge: "Recommended for you." This carried over from the survey work but became far more effective with the new visual layout. The badge now sits on a card users can actually scan and compare, not a text link they might scroll past.
The results (first 6 weeks)
| Metric | Before | After | Change |
|---|---|---|---|
| URL generation adoption | 12% | 47% | +292% |
| First pitch completion rate | 38% | 58% | +53% |
| Median time to first completed pitch | 11 min | 4.5 min | -59% |
| Mid-creation abandonment | 34% | 19% | -44% |
| Users completing pitch in first session | 31% | 54% | +74% |
Completion rates by method:
| Method | Before | After |
|---|---|---|
| URL generation | 71% | 82% |
| Story mode | 34% | 41% |
| Guided builder | 29% | 44% |
| Simple editor | 42% | 46% |
URL generation had always been the highest-completing method. Most users just never found it.
Why it worked
1. Made the fastest path the most visible. URL generation produces a full pitch draft in under a minute. Burying it as one of four equal links was an expensive design mistake. Promoting it to hero status aligned the page hierarchy with actual user outcomes.
2. Progress messages turned waiting into engagement. 30 seconds of a loading spinner feels like a minute. 30 seconds of "Analyzing your website... Extracting key information... Crafting your pitch..." feels like 15 seconds. Users who saw the animated messages were 3.2x less likely to navigate away during generation compared to the old spinner.
3. Cards reduced comparison effort. The color-coded, icon-driven card layout let users pattern-match visually instead of reading descriptions. Eye-tracking sessions showed users scanning all four methods in under 2 seconds with the card layout vs. 6-8 seconds with the old text list.
4. Reduced paradox of choice. Four options still exist, but the visual hierarchy says: "This one is probably what you want. Here are alternatives if not." Combined with the survey recommendation badge, 64% of users went with their first click, up from 41% when all methods competed equally for attention.
The compounding effect
This revamp stacked with the onboarding survey I launched the same month. Users who completed the survey and saw the redesigned page had a 67% first-pitch completion rate, nearly double the pre-revamp baseline. Neither change alone would have achieved that. The survey provided intent; the page redesign provided momentum.
Both changes came out of running Skene's PLG skills on the Pitchkit codebase.
Lesson
The pitch creation methods didn't change. The AI behind them didn't change. What changed was removing the gap between "I want to create a pitch" and "I'm creating a pitch." The old page asked users to become experts in Pitchkit's feature set before they could start. The new page says: paste a URL, we'll handle the rest.
The fastest path to time to value was already in the product. It just wasn't the default.