The last two businesses I’ve started have been online consumer-focused marketplaces. The first business, DealsGoRound, was a marketplace where consumers could resell their unused Groupon, Living Social, etc. deals to other people. It was essentially StubHub for daily deals. The second marketplace is Betterfly, a place where consumers can find online and local instructors/coaches for things like learning guitar, speaking Spanish or improving their cooking skills. In both cases, consumers came to the website and were presented with a list of items to purchase, either a deal voucher in the case of DGR or a package of hourly lessons in the case of Betterfly.
As the person in charge of the product, it is my responsibility to monitor user behavior and make changes which increase the actions we want to see and reduce those that we don’t. “Actions” can mean all sorts of things from purchasing an item to sharing something on Facebook. For the sake of this post, I’ll focus on getting users to make a purchase.
While reviewing user activity with the Betterfly team, one of our engineers mentioned that he thought our instructor’s hourly rates could be perceived as expensive and thus be scaring buyers off. The great thing about this hypothesis was that it was easily testable. So we decided to run a simple A/B test. Using Mixpanel to track user activity and some simple code to handle the variation in what we displayed on the page, we set up a test where 50% of users saw an instructor’s price in the search results and subsequently on the instructor’s profile page (this was the control group) and the other 50% didn’t see the pricing until the very last step of purchasing after they had already selected the number of lessons they wanted and registered for an account (this was the test group).
We ran a couple thousand people through this test and the results were fairly clear.
Step 1 – Search Results
When viewing search results, the test group which did not see any prices had 25% more consumers clicked into an instructor’s profile to view more information about them.
Step 2 – Instructor Profile
Once on the instructor’s profile page, 3x the number of consumers in the test group clicked the “Book Now” (purchase) button when prices were not displayed.
Step 3 – Complete a Purchase
Finally, when it really mattered, 2.5x the number of consumers in the test group completed a purchase when the price was not displayed until after they moved into the purchase process and registered for an account.
In the end, the control group converted roughly 0.6% of users, 4 out of 683, into buyers. The test group came in at nearly double, 1.1% or 10 out of 905. The conversion rates still leave a lot of room for improvement; however, that shouldn’t minimize the fact that we were able to double our success rate with a single simple change.
It’s also worth noting that this data isn’t necessarily transferable across all marketplaces. Amazon and Ebay still show prices and I’m willing to bet that they’ve done a little testing on their end. Perhaps commodities like products on Amazon or Ebay tend to do better with a price being shown right away, but with services which tend to be unique or personal, up front pricing is a distraction. In our case, the data speaks for itself. By continuing these kinds of simple testing efforts and identifying the winners along the way, it shouldn’t be long before our conversion rate is something we can brag about.