strategy · hard
Should Meta build a dating product?
Should Meta build a dating product?
The question is already resolved: Meta built it. Facebook Dating launched in 2019 and is now live in 52 countries with 21.5 million daily active users, the first time Meta disclosed the number publicly. A candidate who opens with “I’d evaluate whether Meta should enter dating” has already signaled they did not prepare. The real question is whether Meta should scale aggressively, hold steady, or exit. The correct answer is to scale, but with a specific structural condition attached.
Structure a strong answer
strong
"Facebook Dating already exists and is growing. At 21.5M DAU across 52 countries, it has already passed Hinge at 15M. The 24% year-over-year growth in conversations among 18-to-29-year-olds is the most important number in the dataset, because that cohort is the one Meta is losing to TikTok (12 minutes per day on Facebook versus 58 on TikTok). So the question I'd frame for the interviewer is: should Meta double down, hold, or exit? My recommendation is to double down, but only on one specific positioning: free AI matchmaking for a subscription-fatigued market. Here is the logic. Tinder, Bumble, and Hinge all monetize via paywalls. Bumble and Match Group both posted double-digit revenue declines in 2025. The TAM is real at around $6B globally, but the incumbents are weakening. Meta's structural advantage is not the social graph per se. It is that behavioral signals from Marketplace, Events, and Groups allow Meta's AI matchmaker to surface compatibility that a cold-start dating app cannot replicate. 'Meet Cute,' the weekly AI-curated match Meta launched in 2025, is the right product direction. The North Star metric is not matches created per week. That is a vanity metric. A match that leads nowhere creates noise and privacy exposure. The correct North Star is real-world meetings within 30 days, because anything short of that fails the lovable bar and becomes a brand liability. Now, the core risk, which I would name explicitly: the users most likely to need a dating app are often the ones who distrust Meta most with intimate personal data. That tension is not a footnote. The mitigation is full product isolation: a separate app, end-to-end encryption, and a clear policy that Facebook profile data is not used for ad targeting without explicit opt-in. That is not a nice-to-have; it is the condition under which the product reaches the users who matter. Monetization stays off subscriptions entirely. The value to Meta is ad signal enrichment from high-intent preference data (hobbies, lifestyle, relationship goals) and retention of a demographic it is visibly losing. Counter metrics: data misuse reports, opt-out rate from profile linkage, and conversation-to-meeting conversion by country. I would expand to the roughly 140 countries where it is not yet live, starting with markets where incumbents have weak penetration."
weak
"Yes, Meta should build a dating product because they have a huge user base and lots of data. They could use their social graph to make better matches than Tinder. The North Star metric would be successful matches per week." This fails on four counts: it treats a resolved question as hypothetical; 'huge user base' is a truism that every Meta pitch starts with and no interviewer learns from; 'matches per week' is a vanity metric that does not reach lovable and generates privacy exposure without relationship outcomes; and it ignores the brand-trust liability entirely. Interviewers call this framework soup with no judgment.
The three questions that structure this answer
Treating it as a binary “build or not” question is the tell. The sharp candidate breaks the decision into three distinct layers.
Is it viable? The global dating market is around $6B and the subscription incumbents are weakening. Meta’s revenue model is not subscription at all: dating enriches ad signal from behavioral intent data. That makes viability a different calculation than it would be for a pure dating startup.
Is it lovable? In 2026, feasibility is solved; the bar is lovable. Lovable in dating means compatibility that leads to real meetings, not just matches. Meta’s AI matchmaker has access to behavioral signals (saved Marketplace items, event attendance, group membership) that reveal preference more accurately than a dating profile. The ‘Meet Cute’ algorithm launched in 2025 is the first credible implementation of this. The lovable test is the 30-day meeting rate.
What is the brand-trust condition? This is the question most candidates skip. Dating is the highest-trust context Meta has ever entered, and Meta’s privacy record is the strongest objection a user can raise. Naming this risk, and naming the mitigation (isolated app, E2E encryption, explicit opt-in for data use), is what separates a senior judgment call from a feature pitch.
The 2026 reframe
This is no longer a 2019 “should we enter dating” question. Feasibility is solved. The AI matchmaking layer is live. The incumbents are in decline. The real strategic risk is not failure to compete with Tinder; it is a product that creates matches but no relationships, which poisons the Meta brand in the most intimate category it has ever touched. The PM who understands that “usable” now means anticipating needs (suggesting a coffee shop near a shared neighborhood, flagging a mutual event) without being obnoxious (no proximity notifications) will give a differentiated answer in 2026.