GTM Sauce

Rise Science · B2C · sleep and energy app

Test each ad in both web-funnel and app-install campaigns, separately

Ads that flopped for app installs sometimes win on web, and vice versa

Workedpaid-socialweb-to-appexperimentation

What they did

Most creatives are run in both web and app campaigns separately rather than making web-specific creative; creative is built around personas, hooks and value props (e.g. more creative for women over 50 because web skews that way). Web campaigns get full e-commerce-style signal without ATT/SKAN limits, so creative winners show up much faster than on app (where it can take 1-2 weeks). Running both also makes performance less correlated month to month.

What happened

More chances of finding a winner per concept; web and app performance move independently, smoothing dips

In their words

Why Web Onboarding Should Sell The Problem, Instead Of The Solution – Leon Sasson, Rise Science

Play the episode

Leon Sasson… at scale, that maybe the traditional ad marketing wasn't really able to reach them, who knows exactly why. And then you can imagine very quickly that the way that you talk to a 50-year-old is very different than the way we talk to a 30-year-old. Especially in the lens of sleep and health, the problems might be very, very different. Because the audience changes, then you're starting essentially with a whole new blank slate on what works there.

Leon SassonSo that's pretty high level, but it was pretty shocking that, "Hey, the audience is different." And that brings a corollary to the marketing, because the audience is different, then the creative that works can be very different on web than app. Not all of it, but some of it can be very different. Which then means that in a way, web funnel when you're running at scale on, say, Meta, or Google, or whatever other channel you might be running, you can end up where you have almost uncorrelated marketing performance between different types of campaigns. Everyone that markets might tell you, "Oh, this month performance is down, and everyone's freaking out." And when you run web and app separate, you realize, "Oh, maybe web is down this month, but app is up." Or the other way around. So you end up with actually a very different sort of cycle that, in a way, just makes the ecosystem a lot healthier, and more robust to flukes in performance. So that's a bit of the high level that we've learned on the website.

David BarnardAre there specific channels or even placements that you feel like do especially well on web that you're seeing a lot of success with? So on Meta, is it more the Facebook placements than the Instagram placements? Or is it Google versus Meta? Are you seeing any specific channels perform better than others?

Leon SassonWe find that generally speaking, they all work the same, because the web audience, at least for us, has leaned older. You tend to see more shift towards Facebook blue app than Instagram, but that's pretty high level, and that ebbs and flows all the time depending on creative. So on that side, there's not a lot of differences. But on the creative side, we often see creative that maybe wasn't working on app that now starts working on web, and the other way around. So you have more chances of winning when you can try something twice, and maybe it's going to work on one side and not the other, because the audience is different.

Leon SassonSo our job on the web is, we're getting people to realize the problem they have, which often is something that rhymes with, "Hey, you're exhausted." Or, "You have all these potential problems in your life, and you might not have realized that sleep is the reason why you have those problems. So we have this app that will help you take control, and feel a lot better, and unlock that part of your life." We had to unlearn a lot of the app onboarding …

Leon Sasson… things change. But one of the nice things about web is the algorithm that learns for delivery tends to be a lot... So the algorithm, the signal and the measurement is a lot more effective, because the platforms, Meta, Google, and TikTok, and whatever else you may be, they don't have to deal with the scan, ATT, privacy related constraints that, "Hey, data is delayed. The attribution is broken. You can only see 50% of the traffic instead of 100%."

Leon Sasson… an e-commerce, which is a lot more easier to pipe all the instrumentation into the system. So the learning system is faster, so it's a lot easier to run those campaigns, one. And then two, the algorithm tends to optimize a lot faster. So you tend to find the people a lot faster, and you tend to know if creative works a lot faster than on app, where you might not be able to really dive down, and you might need to wait a week or two to know if it's working. And depends on the budget, if it's too low budget, it might be longer. So on web, a lot of that iteration cycle tends to be much, much, much faster, which is a huge unlock for web. So it's not necessarily about testing strategy. We do test most creative across both app and web separately, but at the beginning, I believe we might not have been doing that. So it depends on testing strategy, and we're always iterating on that.

David BarnardI wanted to dive into a few tactical things since we are running short on time. Are you spinning up a bunch of different funnels? Are you matching the funnel to the creative? Is there a persona bent where the... Do a creative with a specific persona in mind, and then that leads to a very specific funnel with that persona in mind?

Get tactics like this every Monday

New tactics from the week's founder interviews, each linked to where it was said.

More from this episode

Tags: meta, google, tiktok, creative-testing, web-funnel