Marketing at Reface: How it achieved 100M installations – with 72% organic
Yevhen Leibov, Head of Marketing Reface, shares with AIN.UA which channels and tools their team uses to promote the app.
In the year of Reface’s existence, we have received over 100 million installations and have been at the top in stores in 80+ countries, including the US. 72% of this traffic is organic. But before we got that kind of success, we went through a series of marketing ups and downs. I gathered the most important conclusions in this text.
Better quick than deliberated
When I first came to Reface, I said: “Guys, let’s take it slow, let’s gradually expand into new regions. We start with Ukraine, then we go to Poland, and so on the fifth iteration we reach the USA.” Each iteration will help us to test new mechanics, learn to work with user return, procurement and analytics better.
But the reality was different. I didn’t take into account the fact that the world is highly globalized and becoming viral in, let’s say, Vietnam means that in two weeks you are already popular in the United States. It is almost impossible to control this process. The app started to skyrocket in different countries, so there was no time to think, and it had to be supported by traffic.
We had to act quickly, making risky decisions. It was risky because our tools didn’t allow us to analyze everything in detail. Data in different sources often differed up to 50%. But if we did a lot of thinking, we wouldn’t be able to get to the first 50 million installations so fast.
Everything has to be measured
We are now striving to measure everything.
For example, RevenueCat helps to keep extended analytics on our subscriptions. With AppsFlyer we analyze traffic to understand where people come from. Amplitude is a product analytics tool that helps to see links and build hypotheses for product development. Tableau allows us to visualize all our activity, and Sensor Tower—to monitor competitors by their traffic, activity in stores, etc.
We use Youscan to track potential viral regions that can be warmed up by purchase. For example, this was the case in Thailand when a TikTok сhallenge got viral, which we had nothing to do with. Thousands of Thai people started posting GIFs and videos where they refaced themselves into Disney princesses. We put a paid promo in this region, we came up with themed creatives, we added content to the app, and we got more organics than we could.
Marketing and backend go hand in hand
As soon as we turned on paid traffic, we got a funny case in the Vietnamese market: for one paid user there were 70 free. That was cool, but while the marketing department was cheering up with every new Vietnamese, backend guys stayed up all night trying to keep their servers.
So we had to shut down the traffic and leave only the organics. Later, we inadvertently created the same problem by sending off unapproved text messages. Bottom line: almost every marketing activity needs to be discussed with the backend guys, so they can at least prepare for it. Plans of your own can be very damaging.
Spam works but does not scale
At first, we looked at social media not as a tool for creating community (something we’re working on now), but as a tool to reach out to celebrities and become a little more viral.
We started texting important сelebreties and their surroundings, hoping they’d notice us. This was done by our colleague Danil, whom we called the Editor of Edinstvennaya [Ed. Note: Ukrainian magazine for women] for the fact that the dude was very actively researching who is friends with whom, learned the latest updates from the lives of stars, and mentioned them in our stories. And it worked really well.
We managed to accidentally find a man who does a lot of content with Snoop Dogg, and the rapper actively repost his content. That’s how we appeared on Snoop Dogg’s feed for free, and he still publishes a lot of Reface videos. If we were to turn to influencers directly, our monthly budget would rise to $500k per month (or even more). So it’s great if you have the opportunity to interest the celebrity with something other than money.
However, we later received a bunch of complaints, and one day we got a page with 150,000 followers banned. So we realized that this kind of approach doesn’t really scale up because you don’t want to turn your Instagram into a spam machine: you don’t build complex communication on it.
Paid traffic can eat away at organic traffic
When we reached the top 1 in the US App Store, we thought, since we stand above YouTube, TikTok, and other cool applications, now we will have even more organic traffic. In general, there were more installations, but mostly due to our paid traffic. Then I first learned about the term “cannibalization of traffic.”
A large number of installations can lead to a top, but this does not guarantee organic growth. I used to think it was linear and simple: we increase traffic and make more money.
Actually, it’s more complicated than that. In the case of cannibalization of traffic, it is important to understand: with a sufficiently high budget and sufficient virality of your application, it is easy to get into the situation where you buy a user who could come to the store within a week and install your application for free. So, you paid for nothing.
Now we focus on the benchmark when for one paid user comes 2+ free ones. If we notice that this ratio is declining somewhere, it is a wake-up call — we have grown too fast in purchases for a certain region.
How we build a marketing team
The marketing team was initially staffed with people on the principle of “oh, cool dude, let’s get him, he’s going to help us.” So we formed a team of five people where everyone does everything.
That’s not a bad thing. It’s important to have people like that when you start, but then, you have to divide up the responsibilities. Now we have three directions: user acquisition, product marketing, and brand, and there are people assigned to each with specific metrics and individual challenges.
There are three main principles that are close to our hearts:
- Treat every idea as a hypothesis
Initially, we think about how to test the idea with as little effort and energy as possible. If it is confirmed, we scale it up and move on.
In addition, it’s important for us not just to iterate hypotheses but to document them. That way, we are prepared for the fact that if a new person joins the team in a year, they will be able to dig through everything we have done so far and keep running at the same pace as us.
- Make mistakes and learn more
No one else in the world has the marketing experience for the social platform we’re building. So all we can do now is make mistakes and learn more from our mistakes. Some of them can be costly, but if we don’t learn from them, their cost becomes even higher.
- More action, less presentations
At the marketing level, we have zero tolerance for presentations and strategies. We even joke that one presentation is two proven hypotheses, but generally speaking, it is true.
Presentations are like a huge book that you prepare for a month, and then you don’t know what to do with it. That’s why we avoid presentations, but for systematization, we use more understandable tools – Miro, Sheets, Notion.
Author: Yevhen Leibov, Head of Marketing Reface