Was it technical issues, SPAM problems or user behavior causing our registrations for a new Web initiative to only reach 75%? See what we discovered…
A few weeks ago, we completed a soft launch for SHOP ‘n SAVE’s major 2010 social media initiative — Our Kitchen Stories, a community portal focused on the region’s ethnic food heritage [Learn more about OKS at http://bit.ly/b6Oj1b]. The launch struggled, but after testing, and altering, and testing, and altering — we were still unsure whether it was user behavior or technical issues causing the problem…
Using SnS’s Facebook following and free VIP tickets to a local taping of Top Chef, the PR team began heavily promoting the site to this smaller audience segment — and we had good numbers of folks initiating the registration. With the emphasis on initiating… However, our ratio of initial registrations to fully completed account signups was only around 60%.
Our Kitchen Stories, a social recipe site for Shop 'n Save customers and the community featuring the region's cuisine heritage.
We knew the site was working. It had been tested thoroughly over a four-week period with a diverse group of early users. The registration process was similar to most membership Web sites 1.) User completes form 2.) Email is sent to them to confirm 3.) User clicks on email to confirm 4.) Account registration complete. Yet, lower than expect numbers of users were making it all the way to step #4.
Was it issues with the email confirmation? We focused there first. We ran the email through a SPAM checker and received a score of only 1.6. Even with that low number, we saw room for improvement and adjusted some of the header information. We then conducted two informal tests:
1.) A team member set up 10 emails accounts (gmail, hotmail, aol, etc.), completed 10 registrations, and followed the results. Four (4) confirmation emails in the Inbox, 4 in a SPAM folder, and 2 never made it to the user. 80% success rate – not bad. And the two that didn’t make were university email accounts–we weren’t surprised.
2.) I also had 10 professional friends complete the process – again, an 80% result. This time 2 in the inbox, 6 in the SPAM filters (most friends were coming from corporate accounts so this too was not a surprise) and 2 never got it. Interesting enough, those 2 were both on a Mac.
Not scientific results, but enough to move forward with. We determined that is was an “email deliverability issue” and that 80% was an acceptable ratio. To improve delivery, we posted six different messages in key locations (such as immediately above the “Register” button) explaining the registration process, to look for the confirmation email in a SPAM folder, and instructions for whitelisting. We sat back and relaxed – we thought the issue was solved… Not just yet.
A week later, looking at the results, we had an even more depressing 40% ratio of completed registrations! It dropped. How? Was it technical issues? No, the site was sending the emails out. Was it SPAM issues? No, it was a variety email providers that we knew the email could get through too. Then, what was left? User behavior.
Despite a prominent message, users failed to look for the email in the SPAM folder or whitelist the Web site.
And even a second message did not increase the registration success rates above 75%.
While not scientific, there is only one final explanation: Due to the promotional, contest-oriented nature of our soft launch, users were only taking the first step of the registration process, not bothering to look for and confirm the account email, and assuming that this qualified them for the tickets to Top Chef. There was little incentive for the user to take the time to complete the four-step registration process.
Our Solution? Reduce it from 4 steps to 1 step. While this opens a host of other issues including data accuracy, it helps us ensure that the users who DO want to establish an active account, DO get the opportunity to do so. Today, the site uses a one-page registration form with a captcha check to protect the system from attacks (although captcha itself adds a level of complexity for the user).
The Results? First, the contest is over, so these registrations are not a result of consumers simply trying to win something. Of the last 15 registrations, 4 of them never “logged in” to the account. So, we are still at a 73% account registration success rate.
Now, if only to explain for the user behavior? If a contest is not in play, why would you register for a Web site, and then not even log in once? It feels like the logic of my teenage daughter …
Next Step? Email survey to those who register but don’t log in? We could just be leading ourselves down the same path…
Why do I hear Meatloaf? Two Out of Three Ain’t Bad …
Work completed as Chief Digital Strategist at GatesmanMarmionDrake+Dave
Like following your stuff….One other reason is the population of Pittsburgh, use of the internet and level of sophistication. People who are interested in ethnic food, tend to be a little bit older, their parents and grandparents cooked that way. Now they want to experience it. So they want to get the information without too much hassle.
The key to keeping them is to ensure they make a meal – like it and try it again and use the website as a source and inspiration to try it again.
matt
Thanks Matt. I never really thought about Pittsburgh as a bit more “technology” backward .. but it’s possible. As a region we are typically 2 years behind just about every other trend. Hope you’re doing well! And keep in touch.