Diary of a growth team 2: Moving the needle

“Work a little, move the needle a lot. Case study: iOS push notifications opt-in”

Read this to learn how to identify “quick-win” opportunities in your product to move the needle effectively and significantly. Today’s case study focuses on push notifications, showing how a simple change multiplied our opt-in conversion rate by three and increased retention by 40%.

How it happened

We had been playing around with push notifications on iOS for a while, disappointingly to little effect. To make the matter worse, we had never prioritized finding out exactly how many of our users actually had push notifications enabled in the first place.

One day, thanks to a chat with Ivailo Jordanov who provided some reference benchmarks, we concluded this was a situation that demanded a closer look.

For starters, there is a fundamental difference between the way Android and iOS handle push notification permissions:

• On Android, push notifications are enabled by default, so ~98% of our users actually have them enabled.
• On iOS, each app needs to explicitly ask for permission. The user is presented with this well-known message box:

iOS system popup asking user permission to enable push notifications

% of users who activated push notifications in the previous 100 days

Digging into our analytics platform we discovered that only 10% of our iOS users activated push notifications. This figure looked so low that we were certain we were facing a low hanging fruit. Most likely, even a small change could have yielded a great result.

Engineers reviewed the code responsible for the push notification opt-in. To our dismay, they found out that the user was actually asked to enable push notifications only after running the app three times and sending at least four clips. Our intention behind this strategy, which was compliant with best practices, was to let the user understand the value of the app before asking for permission to send push notifications. Thanks to our analytics investigation we could clearly state that the bar was set way too high, as most users simply were never asked if they wanted to enable push notifications.

Lesson learned: if push notifications are key to your retention strategy, do not wait too long to ask your users to enable them.

The next step was to look for a way to move that needle fast. A rather old but still completely relevant article on TechCrunch (link) explains there are several best practices surrounding push notifications opt-in.

Our first version of push notification “blitzkrieg”

The simplest and probably most brutal strategy depicted in the article, appropriately named “Blitzkrieg”, consists of asking the user straight away for permission on the very first execution of the app. This approach leads around 30% of users to enable push notifications.

Considering we were sitting at an unsatisfactory 10%, this plan sounded like a great gain in exchange for very little work.

So we designed a “blitzkrieg” enable push notification priming pop-up, depicted here, and showed it immediately as the user opens the app. If the user pressed Ok we would activate push notifications, if they pressed Not Now we would try again in the future (for more on priming and why it makes a lot of sense, see for example this optimezely article).

Opt-in rate climbed exactly as expected, from 10% to 30%. This, combined with the retention campaigns we kicked off (see the article here), increased our day 2 iOS retention by over 40%.

These were two hours of work that really paid off!

Lesson learned: improving a lot with little work

Since then we applied the same framework several times with very good success. You can try this at home, to identify opportunities for quick wins in your product:

1. Model your product. Lean analytics (link) offers great examples if you don’t want to come up with one from scratch
2. Define and measure key metrics for your apps
3. Benchmark your metrics with other apps in your industry (if you are not sure how to start, see appendix for a beginners’ list)
4. Once you identify an area where your product has a significant gap with the industry benchmark, you are almost certainly doing something simple wrong. Simply search the web for best practices on how to handle that can move the needle quickly
5. The hard part is done. Now implement, wait some time (often a day is enough), check your metrics, celebrate!

Appendix: How to benchmark your app

Good data online is hard to find, but here is a partial list to get you started:

• The Lean Analytics book has a good starting set
• Quora often offers great insight from insiders, hard to find elsewhere
• Various benchmarks are published by a few players, such as these ones by Localytics and Adjust. Aggregate reports are also available, like this one from mest.

Thanks to the Emoticast team who made this possible. Special thanks to Lorenzo Aurea and Michele Battelli for their invaluable help.

The journey continues, because growth is never done…