Outskill runs skill-based online courses and launches 4-6 new programs every month. Each launch comes with a landing page, and the growth team needs to test multiple variants to find the highest-converting design.
The old process: create separate ad sets pointing to different landing pages, wait for statistical significance, then reallocate budget. This took 2-3 weeks per test and cost $14K+ in wasted ad spend on underperforming variants.
The A/B Testing Setup
With getowl, Outskill now runs link-level A/B tests:
- Single ad link shared across all channels
- Weighted splits: 50/50 to start, then adjusted as data comes in
- Real-time analytics show which variant converts — not just which gets clicks
- PostHog integration pipes every click event into their existing analytics stack
The Results
- 28% more sign-ups in the first quarter after switching
- Testing cycles shortened from 2-3 weeks to 4-5 days
- $14K/month saved in ad spend previously wasted on underperforming variants
- The team manages 3,200+ active links across 12 members
How They Run Tests
- Create a single getowl link with two weighted variants (50/50)
- Monitor click-through and conversion data for 48-72 hours
- Shift weight to 80/20 favoring the winner
- After confirmation, set to 100% on the winning variant
- Archive and document results for future launches
Aamir, Head of Growth at Outskill: "We used to need a data engineer to set up proper A/B tests on our links. Now any marketer on the team can run an experiment in 2 minutes."