The Dual Landing Page
One page for two audiences, twice the apply starts
Client
Glassdoor
Role
UX and visual design, from hackathon concept to A/B test
Team
PM, advertising team, B2B web team, engineering
Timeline
2019–2020 · Hackathon idea to tested page
Overview
Glassdoor’s search ads on terms like “Glassdoor” sent every click to the employer page for posting jobs, even though many of those visitors were job seekers. I designed a “choose your own adventure” page that let each visitor pick their path in one click. It started as a Hack 14 project, won, and became a real A/B test against the existing page. It doubled apply starts and lowered the cost of each new customer.
2x
apply starts
+64%
job alert sign-ups
10–15%
lower cost per acquisition
A/B test on paid search traffic: the dual page vs. the existing employer page. Bounce rate also fell 40–70%.
01
Problem
Job seekers who clicked Glassdoor’s employer ads landed on a page that wasn’t for them, with no way back to Glassdoor.
Every click from these ads went to /post-job. Job seekers bounced, the consumer side lost visitors it had paid for, and the high bounce rate hurt the ads’ quality score, which raised the cost of every click.
Send each visitor to the right place. Let job seekers and employers choose their path in one step.
Lower the cost of paid search. Cut bounce and raise ad quality scores.
Keep it testable. Make copy, icons and buttons easy to swap so the ads team could keep iterating.
02
Two audiences, one page
Start at the hackathon. For Hack 14, our team mocked the idea up as a split page: job seekers on one side, employers on the other. Winning gave it the momentum to become a real project.
Balance two audiences. Each side has the same structure: an icon, a question (“Hiring?” and “Job hunting?”), three benefits and one button. In the final round the employer side got the emphasis: a filled Post Jobs button against an outlined Find Jobs, since posting jobs was the ads’ primary offer.
One clear action per side. Post Jobs for Free for employers, Find Jobs for job seekers, with a sign-in link for people who already have an account.
Back every claim. Each employer benefit, like “Employers read 50% fewer resumes,” carries a footnote to its source, so the page reads as proof rather than a pitch.
Remove the exits. The header holds only the Glassdoor logo, linking back to glassdoor.com, and a Need help? link, so visitors choose a path, go home or get help, with no full navigation to wander into.
Built for testing. I explored four layouts, from a full-bleed split to side-by-side cards, and kept copy, icons and buttons easy to swap. The page stays out of search results and carries ad tracking through every link.
Four layouts for the dual page, from a full-bleed split (A) to an employer card given more room (D).
On phones the two cards stack, and each button runs the full width of its card.
Redlines for engineering: spacing, type styles, button sizes and icon dimensions for the desktop page.
03
Outcome
The page did what the ads team asked. Employer sign-ups held steady while each one cost 10–15% less, so the ads stopped paying for clicks that went nowhere. Job seekers, who used to bounce, finally had somewhere to go: apply starts doubled and job alert sign-ups rose 64%. Bounce rate fell 40–70%.
04
What I learned
The simplest fix can be the biggest. The page asked one question, and that was enough to change the numbers.
Hackathons are a back door to the roadmap. A hackathon mockup got a real test, because it was easy to say yes to.
Serve the visitor and the metrics follow. Helping job seekers who weren’t our target audience also made the employer ads cheaper.
