Alex Schultz on Meta Growth, Product Experiments and Leading with Candor

In this 2023 interview, Meta chief marketing officer and analytics leader Alex Schultz discusses growth, experimentation, product decisions and his experience as a gay man in technology. He speaks from his role at Meta, including when assessing its products and competitors.

Schultz began building paper-airplane websites as a teenager and learned search optimization when one attracted traffic. While studying physics at Cambridge, he experimented with advertising and eBay affiliate marketing to support himself, buying search ads on his credit card and earning commissions.

After considering a PhD, he applied to eBay and showed the team his affiliate work. He says the company offered him a job and held it open for a year. His early site was initially about the pleasure of attracting visitors; monetization became important later.

At eBay he learned data-driven search marketing, affiliate marketing, merchandising, customer relationship management and presentation skills. He credits several managers and colleagues for shaping his work before he moved to Facebook.

Harry asks about Mark Zuckerberg’s praise for Schultz’s contribution to Facebook reaching a billion users. Schultz says the counterfactual cannot be tested and credits numerous colleagues. He identifies a data-informed growth model as a more important collective achievement.

The model accounted for registrations plus returning users, minus churn. Schultz says the team discovered both churn and returning users were much larger than expected. That shifted attention from acquiring users at any cost toward retaining them and bringing them back.

At eBay, changing the target from confirmed registrations to activated users changed marketing decisions. When the team prioritized people who bid, bought or listed, it sent searchers to relevant results rather than a registration page. Registrations fell, he says, but the number completing meaningful actions increased.

To identify actions associated with retention, Schultz suggests examining correlations with the desired outcome and applying common sense. For Facebook that might involve friending, for WhatsApp messaging and for Instagram following. A correlation alone does not prove that forcing an action causes retention.

Harry asks whether content discovery has displaced the social graph. Schultz sees a mix: recommendations matter more, but friends, follows and likes still supply useful signals. His assessment is a Meta executive’s view of a competitive market at the time.

Schultz recalls that Facebook’s web messaging was strong, but mobile messaging growth concealed a loss of overall share as web use declined and other mobile services expanded. The team had been measuring the transition too narrowly.

Research suggested users expected faster responses when the recipient used a dedicated Messenger app, whereas messages inside another app behaved more like direct messages. The team tested separating Messenger in selected countries, observed an initial dip followed by increased usage, and refined the rollout.

Schultz says large mature companies can run longer experiments and segment their results, while startups may need to act with less data. He now thinks he asked for too many Messenger tests and moved too slowly even when the evidence was clear.

Harry asks whether it is better to move second and learn from a competitor. Schultz says launching, catching up and then innovating can work, but waiting too long risks missing a wave. Speed also creates more opportunities to run experiments and learn.

He says the first version of Reels, placed within Stories, failed, and subsequent iterations taught the team more. He claims Meta moved faster than some rivals and is beginning to innovate on the format. These are his account and competitive claims as a Meta executive, not independently established comparisons.

For analytics leaders, Schultz says the task is to explain evidence clearly enough to influence decisions without concealing important caveats. When a leader chooses another path, an analyst can later compare the outcome with the prediction and gradually build trust in the learning process.

He uses Messenger as an example of over-caution: he wanted another round of tests despite having enough information to act. He thinks the delay may have limited its success, while acknowledging he cannot determine the long-term counterfactual.

Schultz says Facebook advertising evolved into a combination of self-service and sales rather than a simple replacement of one with the other. Small businesses could boost posts or create ads themselves, while sales teams helped promising advertisers adopt more products.

He recalls a French translation that effectively changed “advertise” to “create an ad.” He says advertiser growth in France was 30–40% higher in that test, and the company adopted the wording more broadly. This is his account of a historical test, not a general guarantee about copy changes.

The self-service funnel and later sales support, he says, helped advertisers move from simple promotions to more sophisticated formats. Schultz argues Meta is especially useful for small businesses; the claim reflects his commercial role at the company.

Asked whether product teams need psychologists, Schultz values behavioral expertise but also stresses trying many ideas. Even experienced teams cannot reliably predict how people will respond to a button, wording or flow. In some experiments a longer flow with more targeting options converted better than a shorter one.

He credits engineering leaders with encouraging reviews that learn from honest mistakes rather than assigning blame. Repeating the same error can still prompt a performance conversation, but people should feel able to acknowledge a mistake and experiment again.

Harry asks what to do with an employee who is adequate but not ready for the next stage. Schultz says the answer depends on whether the issue is skill, motivation or role fit. He favors a direct conversation about strengths, available roles and compensation, and says separation may sometimes be necessary.

He finds difficult feedback uncomfortable himself and recalls receiving coaching to have a hard conversation. Avoiding the issue until a poor review surprises someone is a management failure, he says. Caring means giving clear guidance along the way and considering the other person’s perspective.

Schultz says the Facebook-to-Meta name change addressed confusion between the company and the Facebook app. The company also wanted people to associate the corporate brand with products beyond one application, including WhatsApp, Instagram and virtual reality.

Harry asks whether distinct product brands create internal silos. Schultz says that depends on management, not the company name. He describes analytics leads meeting across product groups, weekly insight reviews and shared leaders who transfer lessons between teams.

He presents Meta’s brand as connecting people through different experiences: social discovery on Facebook, creativity on Instagram and private conversation on WhatsApp, with virtual and augmented reality as longer-term possibilities. This is the company’s framing, not an independent assessment of its privacy or products.

Schultz believes the rebrand achieved its goal of broadening perceptions and cites an external ranking as one indicator, while cautioning that a sharp short-term change may not last. A goal and the metric used to evaluate it are different things, he says.

Every metric has flaws: registered users may never activate, and active-user counts can include fake accounts. Teams should understand the weaknesses, use counter-metrics and watch for Goodhart’s law, where a measure becomes less useful once it becomes a target.

For early founders, Schultz suggests learning from methods developed by earlier teams without assuming that every mature-company process is necessary. He believes founders should be precise about what they want to achieve and how an incentive might drive unintended behavior.

Harry reflects that achieving a financial milestone did not resolve his unhappiness. Schultz says new information can justify changing goals, although he prefers changing goals and metrics infrequently. He thinks work should be economically viable, but beyond a base level people should also consider impact and colleagues.

When teams miss goals, Schultz says the goal itself matters: a team that consistently hits targets while its business declines may be measuring the wrong thing. A demanding goal can disappoint people when missed, but a comforting target on a losing path is not a solution.

Harry describes a team target of meetings per week and worries it can be gamed. Schultz suggests pairing a volume measure with a quality counter-metric, while acknowledging that no combination is perfect. Leaders still have to inspect behavior and use judgment.

Schultz came out at university after falling in love, then says he went back into the closet at work after being advised that being openly gay might hurt his career in the United States. He describes that period as isolating and psychologically difficult.

At Facebook, supportive posts from colleagues including Blake Ross and Mark Slee helped him feel safe enough to come out at work. He later sponsored product and community work supporting LGBTQ+ people. He describes those colleagues’ visible support as personally transformative.

He says being a white cisgender gay man gives him advantages relative to many in the LGBTQ+ community, and believes trans people often face greater difficulties. At the same time, he recognizes that each person’s experience differs and describes still feeling isolated as an openly gay senior leader in some settings.

Schultz says his parents initially needed time to adjust their expectations of his future, but became deeply supportive. He credits them with teaching him to do what he believes is right even when it is difficult, and describes his father as an example of patient listening and being quietly right.

In the quick-fire round, he says startup speed matters and analytics leaders need patience when persuading others. He recalls Facebook creating a growth team after growth stopped in 2007, and says his team celebrates successful outcomes rather than merely shipping a feature.

He attributes some long working relationships to honest feedback and caring about colleagues’ long-term development, while acknowledging the approach does not work for everyone. He credits mentors with teaching him to adapt his tone and pay attention to whether others actually act on the message.

Looking ahead five years from the 2023 discussion, Schultz hopes Reels becomes a major monetization engine and that Meta can demonstrate value from virtual and mixed reality. He also wants public decisions and appropriate laws on content moderation and well-being. These are ambitions and policy preferences, not established outcomes.

Harry thanks Schultz for the conversation and advice; Schultz returns the thanks.