AppLovin is looking at software as its main cash cow.
Going forward, the company plans to prioritize its software platform and focus less on its apps business.
During its Q1 earnings call on Wednesday, AppLovin told investors it expects revenue from its software business, which includes AppDiscovery (a marketing platform), MAX (in-app bidding software) and Compass (analytics within MAX) to hit $2 billion by 2023.
Software is now responsible for 40% of AppLovin’s revenue compared with 14% when the company went public in April of last year.
Going hard on software
At the time of its IPO, AppLovin was yodeling from the rooftops about the first-party data advantage it got from its portfolio of apps. AppLovin publishes its own titles under Lion Studios and has stakes in (or has outright acquired) nearly 20 game studios around the world.
(AppLovin studio roll call: Ace Games, Athena Studio, Belka Games, Clipwire Games, Forever9, Geewa, Kryss, Leyi, Machine Zone, Magic Tavern, MagicAnt, Nuts Power – really, that’s what it’s called – PeopleFun, Redemption Games, ZenLife and Zeroo Gravity.)
Historically, AppLovin invested heavily in user acquisition (UA) for the apps produced by these studios so they could serve as a ready source of first-party data and audiences to feed its machine learning models and Axon, its homegrown recommendation and prediction engine.
Over time, the flywheel started to spin as more developers used its technology and AppLovin became less reliant on first-party data coming from the content side of the house.
“We’re seeing customers come online,” said CEO Adam Foroughi. “We’re serving more ads, getting a bigger feedback loop … and we’re seeing [our] machine learning continue to improve without a necessity for our own games to be fueling the data.”
Although apps are still the majority of AppLovin’s business, the software side is growing faster than expected, Foroughi said, noting that revenue in the first quarter is up 4x year-over-year and that software contributed more than 80% of the company’s net profit in Q1.
“For the last several quarters, we’ve been talking about how the apps business is not as strategic as it once was,” he said. “Given the success of our software platform, we will no longer run our games as a cost center.”
In other words, AppLovin is going to audit and restructure its apps business so the apps biz can operate more efficiently as its own stand-alone business unit. That could mean “operational changes and possible plans to sell or spin off some of the studios,” Foroughi said.
Going forward, AppLovin plans to operate the studios with “more profitable” spend on UA, a process that already started late in Q1.
“Traditionally, we were willing to spend more on new users, valuing the scale, audience and data as a justification,” Foroughi said. “This led to operating it around breakeven.”
But now that its machine learning algorithms aren’t as reliant on its own apps as first-party data fuel, AppLovin will prioritize its margins – and software has way higher margins.
The goal is to eventually operate the apps business at 20% higher EBITDA margins, which is more in line with typical game developers.
But you are what you eat, and that applies to machine learning models, too.
Unity Software reported earnings earlier this week. Its market cap was cut in half. So, no, it did not go well.
Unity, which has a suite of products and services for app developers and content creators (UA, monetization, player engagement tools, etcetera), plummeted after alerting investors to “issues” with Audience Pinpointer, the machine learning tool at the heart of its strategy for dealing with Apple’s privacy changes.
Pinpointer relies on algorithms that look at real-time in-app behavior data at the time of an ad request to determine the bid most likely to generate performance. Which is fine, but Unity lost the value of a portion of its training data “due in part to us ingesting bad data from a large customer,” John Riccitiello, Unity’s CEO and president, told investors on Tuesday.
Performance suffered, so customers spent less. Like a flywheel, but in reverse. The impact to Unity’s business will be roughly $110 million this year, with no carryover impact to 2023, and Riccitiello said fixing the issue is a top priority.
AppLovin’s investors were understandably curious if its Axon tool is exposed to the same potential problems.
“Machine learning obviously has opportunity and sensitivities and really does require a good data feed, and a lot of this just comes down to execution,” Foroughi said. “We’re focused on our own execution, not what others around us are doing.”
MoPublishers on the platform
In other news, AppLovin’s acquisition of MoPub closed in early January, and the MoPub exchange is now fully integrated into MAX.
More than 90% of MoPub publishers have made the transition, which nets out at just over 700 million daily active users in one exchange, Foroughi said.
But AppLovin didn’t give MoPub customers much time to migrate to MAX, just 90 days. To protect publishers from revenue loss as they rapidly moved their inventory onto MAX – a nontrivial affair – AppLovin paid out $210 million in non-recurring publisher bonuses, which it views as an investment to create cross-selling opportunities.
“We’re hopeful that those publishers will become customers on our AppDiscovery platform,” said AppLovin CFO Herald Chen, who noted that if you add the $210 million in bonus payouts to the MoPub sticker price, the total for that acquisition was $1.26 billion.
“That’s a very attractive price for such a strategic and financially lucrative asset,” Chen said.
And a quick nothingburger alert: AppLovin’s shares were briefly halted for trading right after the market closed on Wednesday, but the issue was quickly resolved. Apparently, there was a glitch in getting the earnings press release posted, and the SEC temporarily halted the stock to be safe.