Case Study: WorkingPoint
SEM Landing Pages
WorkingPoint is an online small business management solution offering
small business accounting, online invoicing and many other small
business needs.
We had a problem at WorkingPoint. Our product was hailed as "The Swiss Army Knife of Small Business Financial Software" and celebrated by the likes CNBC, PC Mag, BNet, Web CPA and MarketWatch. But all the praise in the world could not fix suffering conversion rates on our paid search campaigns. We just weren’t acquiring customers at the rate we wanted or needed.
Our early paid search campaigns at WorkingPoint consisted of buying traffic and sending them to either our homepage or one of three pages that described invoicing, accounting or financial reporting features. Conventional online marketing wisdom told us landing pages were the way to go, but we had no marketing department and none of our engineers had ever built one.
Everyone on our team had a different theory about the strategy that we should take to convert visits to sign ups. Our team, consisting of coders, engineers and our CTO (me!), were each advocating for different approaches. Rather than zero in on one strategy, based on who could quote the most articles or argue their case more persuasively or even had the fanciest title, we opted to act on a phrase that would later become akin to a manifesto for our team, “Our opinions don’t count, only data does”.
As a reader of this wiki, you're hopefully a believer in the principle of getting ideas in front of customers as quickly as possible. This is where the feedback loop comes into play.
The Basic Feedback Loop
- Build
- Measure
- Learn
- Repeat
The key is to loop as fast as possible. The more you loop, the more you improve.
- Build: within hours we launched three landing pages that would allow us to experiment with copy, layout, buttons and other design aspects.
- Measure: landing pages were equipped with tracking code from Google AdWords, Google Analytics and some additional internal measures to allow segmentation on free vs. premium.
- Learn: once statistical significance had been reached, we analyzed the data. Concise copy ruled and "Sign Up Now" was nowhere near the best call to action.
- Repeat: We used what we learned and cranked out another set of iterations in no time.
We built all 6 test versions of the different approaches suggested and pitted them head to head against each other in carefully controlled AB testing.
The competition quickly escalated, each of us betting on our versions with the winners cashing in on a bounty of free lunches and after work beverages.
During these head to head battles, we worked hard to ensure that we duplicated all conditions explicitly except for the variant including time, date and keywords to ensure the purity of the data.
Based on this feedback loop, the whole team then worked together to polish and optimize the winners through over 30 variations. Some of the most effective, refined features emerged as a result of collaboration, as the team came together to optimize the best performing options.
Over about six weeks, our learning and improvement was remarkable. The early wins weren't earth shattering, but with each trip through the loop we improved. We tried everything we could think of and almost anything that was suggested to us because we wanted to avoid local optimums. Unsuccessful branches were mercilessly culled no matter how much we believed in them, who their internal champions were or how much time it took to develop or design them. Overall, the conversion rates we achieved have made a few jaws drop. Rather than reading up on CRO or pontificating about feelings and intuition, we took up learning by doing and quickly became expert.
I'll wrap up by yelling at you as we yell (with a smile) at each other: "YOUR OPINION DOESN'T MATTER!!" Build, measure, learn and repeat. Get your work in front of customers because it's their opinion that does.
Comments (0)
You don't have permission to comment on this page.