Meta plans to stay scrappy.
The company has laid off more than 10,500 people since the start of the year (and roughly 20,000 total since November). But now that the deepest cuts are over, the focus for the rest of 2023 is on “removing barriers that slow us down and introducing new AI-powered tools to speed us up,” CEO Mark Zuckerberg told investors during the company’s most recent earnings call in late July.
As a leaner machine with a flatter organizational structure, it’s possible for smaller teams to more quickly develop products and bring them to market.
“We’re finding we can be more agile with relatively small teams and ship products fast,” Justin Osofsky, Meta’s head of online sales, operations and partnerships, told AdExchanger.
The recent launch of Threads is a good example of this new “flatter is faster” mantra (which is arguably catchier than “move fast with stable infrastructure” and less reckless than Facebook’s original motto of “move fast and break things”).
Meta released the text-based (purported) Twitter killer Threads – built in-house by the Instagram team – in just seven months from start to finish.
“We’re running a leaner, scrappier company overall,” said Osofsky, who moved into his role at the Meta mothership in February after four years as COO of Instagram.
AdExchanger spoke with Osofsky to get a pulse on the state of Meta’s business.
On what’s next for Threads: “The reception it’s had so far means that it’s meeting a need. That said, we’ve got a lot of work to do to make the product great. We’ve been shipping new features fast over the past few weeks, but we also need to make sure that retention is there. There are multiple things that have to happen before we think about how to monetize.”
On the role of automation: “At first there’s an education component. Advertisers want to know what our AI-driven ad products do, how it makes their jobs easier and how it’s different from business as usual. That’s as true for the Advantage+ suite as it is for other products, like our Conversions API, or CAPI. In that case, advertisers want to know how they can get better measurement and better understand their results.”
On using AI to rank and recommend content: “We’re building large language models and using training data to produce results that are hopefully better for people. For example, we can recommend content from accounts you don’t follow but that we know you would be interested in. We call this ‘unconnected content.’ Recommending unconnected content allows people to discover new interests, which in turn creates a better experience and leads to more engagement.”
On Reels monetization: “We’re seeing a lot of engagement here, and I think that’s evidence our investments in AI are paying off. There are more than 200 billion plays of Reels every day across Instagram and Facebook. But while it’s good to see organic growth, we’re also continuing to improve the ad products and the formats, including overlays and experimenting with AR.
“Whenever we layer in monetization, it has to be in a way that feels consistent with the experience. We’re doing that with Reels, and it’s how we would do it with Threads if we monetize at some point in the future.”
Answers have been lightly edited and condensed.