3, 2, 1 … Countdown to Retention

I started my career in the Product team at XO Group Inc., - the parent company of TheKnot, TheNest, and TheBump. At the time, TheKnot was the premiere wedding brand with a national magazine published quarterly, additional regional magazines, a website with the top Google-ranked content for most wedding topics, a local marketing place, a website builder - and the top rated wedding planning app in the Apple App Store.

We always marveled at how highly-ranked the app was - as a product team we knew there were a lot of things we wanted the app to do that it didn’t do yet, but we thought our good brand name was enough to keep the app at the top of the App Store list. And then, March of 2016* we saw retention greatly increase. No major feature had been released, no major marketing initiative. But no matter how we sliced the numbers, the cohorts starting in February had higher 7-day and 30-day retention (we cohorts by the date the app was installed; retention by when the user opened the app again).

[MixPanel Retention Chart]

The app was joint-developed by several product squads and each vied to take credit for the increased retention. But when we looked at retention data on a per feature level, nothing showed the same increase - not the wedding vendor marketplace, not the messaging inbox, not the inspiration boards. After many conversations, it finally occurred to someone that we had gone through some product debt with that release as well, features that were intended with the initial launch of the app but were deemed inessential to launch but should be added at some point, mostly low engineering efforts that polished the app more than increased it’s functionality. One such feature was a countdown to your wedding date on the homepage of the app. It took practically no time to engineer, it was an after thought. No one believed that this dinky little feature could have such an outsized impact on the app’s metrics.

But it did.

Gratefully, we had automatically tracked clicks in app, so we were able to redo the retention analysis based on interacting with the countdown timer (when you clicked it it cycled through the number of months, weeks, days until your wedding) as we saw the increase in retention. Digging further, we saw the frequency of clicks was off the chart, so high as to be shocking. We started to think that we had busted some of the tracking.

One of my teammates was planning her wedding, and when she saw the analysis it instantly made sense to her. While you’re planning your wedding, that countdown time was everything - and she knew it first hand. With her permission, we looked at her app activity, and it showed exactly what we had seen in aggregate - that day after day she would open the app and click to see the number of months until her wedding, the number of weeks until her wedding, the number of days until her wedding.

[Mixpanel Person View]

So we knew that the retention in app had increased drastically and the product feature that caused the increase - the next steps was figuring out how to leverage this to grow the app in other ways. We were also tracking when a user took screenshots of the app, and unsurprisingly, at this point, people were frequently screenshooting the countdown timer. On a hunch, I threw a screenshot I had taken into google’s reverse image search - hundreds of similar screenshots were returned, posted by brides- and grooms-to-be on various social platforms.

[Google Image Search View]

With that in mind, we added social functionality to the countdown timer that would automatically share it to the social platform of the users’s choice - with a tracked link to download the app for any of their engaged friends who might see it and click.

This was an inspirational experience for me. It’s rare for a product analysis to be so clear and lead to actionable growth - and when it it happens it like a rocket taking off. 3, 2, 1 … blast off!

*It might have been 2015, this is long enough go that some of the details escape me. The conclusions and takeaway are sincere, even if some of the details have been massaged by time.