50k Installs and 1-Year On: NHS Go

NHS Go was launched a year ago this month, and in the time since has racked up over 50,000 installs across iOS and Android devices. SOHO Managing Partner, who oversaw development of the brand and delivery of the app, discusses some of the lessons learned during this time.

Whilst originally brought in to “create the app and brand”, subsequent to launching the application, we were asked to apply our experience in delivering targeted online advertising campaigns to the project.

Over an initial 30-day period, we developed a strategy for delivering incredibly cost-effective installs of the app (as much as an order-of-magnitude cheaper than previously-run campaigns). However, in the process, we came to appreciate some important things.

Addressing health inequalities

One critical realisation we made early on was that the breakdown of users finding our app organically was not spread evenly across our broad target demographic of under-25s, but concentrated amongst young people with higher-education backgrounds, or who come from higher-income families.

To reach those most in need, we introduced a hyper-targeted approach to social-ad retargeting that specifically sought out young people who (1) would be unlikely to hear about the app otherwise, and (2) who would be most in need of its care.

Cost-effectiveness is a complex measure

A typical key-performance indicator (KPI) of an ad-campaign might be the cost-per install (CPI) of an ad-set. Indeed, our initial approach was to obtain the lowest CPI possible, to maximise installs of the app amongst our target demographic (under-25s in the London region). From this, we quickly ascertained a few things.

1. If we were serious about minimising CPIs as much as possible, we would have to be extremely particular about the platforms we targeted.
   • Medium specificity: Of Google AdWords, Twitter, Facebook, Instagram, and Apple Search Ads tested in the initial advertising period, the top-performing platform (Instagram) yielded results 32-36% better than the next-best performing platform (Facebook).
   • Platform specificity: social advertising targeted at iOS devices performed 10-20% better on a CPI measure than their Android equivalents.
2. Targeting so particularly with our advertising ran a significant risk of being socially-exclusionary. Were lower-income and less-educated young people as likely to use Instagram and iOS devices? Whilst we had a strong inclination, we really wanted to go one step further than anecdotal evidence, and began looking at platform usage data overlaid atop of geographies ranked by household income.
   • Twitter analysis of platforms: Analysis of several million tweets showed that users of Twitter for Android predominantly live in poorer regions and areas than their Twitter for iOS counterparts. This is reinforced by external studies showing iOS users tend to holiday in more expensive locations, and that the iPhone has a greater market-share in wealthier US states.
   • Not preaching to the choir: If we accept that children from lower-income households may be more in need of the sorts of support provided by the NHS Go application, then there would seem to be a very real risk that optimising solely for CPI might result in “preaching to the choir”: only putting the app in-front of people for whom it would deliver diminished benefits, all for the sake of better vanity metrics.
   • Ignoring exclusionary variables in our targeting: as a result of this early realisation, and conversations with the Healthy London Partnership team (representing all 32 of London’s Clinical Commissioning Groups), SOHO ran an inclusionary advertising campaign across mediums and platforms, specifically seeking to optimise adverts without restricting targeting parameters on these bases (and others which we determined might be reflective of socioeconomic background).

Long session-durations aren’t always great

Very often, long periods spent in-app are indicative of good, engaging content. When looking at feedback obtained from people who had been using the app for an extended period, however, we found that the reality was more mixed than this simplistic assumption would imply.

Many users were indeed amongst the most positive reviewers of our app. Others, though, were stuck. A small number couldn’t find the information they required. A larger group struggled to take in some of the more confusing medical information in the app.

We concluded that certain key content must be more easily-digestible in-app, and reframed for young people. Improving the ways users engage with the app forms a large part of our vision for year two.

Conclusion

One year in, and with fifty-thousand users, we are now in a much better position to test hypotheses and gain insights into what works, and what doesn’t, when communicating with young people about their health.

Having proven the app’s utility, we’re now making much-needed improvements to the way information is delivered and communicated within it, as well as how it is marketed to young people generally.

We hope to bring additional functionality into the app in the coming months and years, and are lucky to be working with a truly visionary team within the NHS who act upon evidence, and look to emerging technology for new ways of solving old problems.

If your organisation has a digital challenge in need of a thoughtful, novel, evidence-based strategy, then contact us via the form on our website, or email SOHO Managing Partner, David Wilkinson, at [email protected]