About
An embedded testing team, run by the person doing the work.
repeatGROWTH is not an agency with an account layer between you and the testing. It is a small team that plugs into yours.
The founder
19 years. Same question every time.
In high school I was into Photoshop, Illustrator and graphic design, and making animations in Flash. But what I wanted to make was films. Film felt like the ultimate multimedia project: story, music, dialogue and performance all working together to move an audience. An art requirement put me in front of HTML instead, and I started building sites on CSULB school servers and Geocities about whatever I found interesting.
My first job out of that was running landing page work for a software company. We tested messaging, currency, language, urgency and trust, and watched the conversion rate move. Film tries to influence an audience. Testing tells you whether it worked. That was the end of guessing for me.
At DMV.org I optimized affiliate advertising, onsite search and the overall experience, then built the company's first testing and optimization team. With heavy traffic, a lot of affiliates and full revenue tracking, we ran more than 60 tests a month and added millions in annual revenue year over year.
What we built there became the process we still use. Analytics shows what customers did, but not what they were thinking, so we pair the numbers with watching real people use the site. That tells us why a page isn't working and what's worth testing. From there we write the hypothesis, design and build the test, and let the results decide. Every test, win or lose, teaches us something about the customer that feeds the next one.
When an algorithm change hit the site, I started repeatGROWTH, and former coworkers who already knew the process and shared the passion behind it came with me. We never added an account layer. The work succeeded because we were inside the team, not reporting to it, and that's still how we work.
Shane Hale, Founder
How we work
4 things we will not do.
Call a test early
Not at 90 percent, not because the line looks good on a Friday.
Hide the losses
They are published with the wins, on the same page, with the same detail.
Take work we cannot read
If the traffic cannot reach a clear result, we say so before you pay.
Put an account layer in between
You talk to the person running the test.
By the numbers
