How Pray.com grows

The social network era required name, email and phone number to verify real people. For the consumer subscription funnel the phone number caused high drop-off, so it was removed from signup; they later found other ways to collect it.

Used TV and radio because the older, higher-LTV faith audience trusted 'as seen on TV' more than digital ads. Worked with large TV-buying agencies. Creatives included James Earl Jones reading the Bible plus in-house spots. Split spend: about 50% spread thinly across many channels and time slots for consistency, the rest on big bets like prime time, since a given slot can be the best one week and the worst the next. Watched by-minute real-time response feeds while ads ran.

Sold a private social network to faith organizations (mosques, synagogues, temples, parishes, churches) and onboarded their members in person with a booth, often teaching older members how to download an app. Reaching 80% penetration usually took more than one visit. Organizations gather one day a week (usually the same day), so a field sales force with a faith background was needed.

Noticed pastors posting short faith quotes with a devotional on the social network and that these engaged differently from normal posts. Tied them into text messages, added audio/video recording for pastors, then productized them as morning, noon and nightly prayers, and in 2019 created in-house devotionals. Longer-form content then became the basis of the subscription.

Followed the 'imitate, iterate, innovate' approach: copy formats already proven by Calm and Headspace (e.g. sleep content) to get benchmarks, but from a faith perspective: bedtime Bible stories, sleep psalms, meditative prayers. Then iterate through constant tests to learn what could be monetized and how far. Kept a free version for users who expect faith content to be free.

Redid the whole product platform so onboarding flows and paywalls are driven from the server, letting the team run rapid A/B iterations without shipping new app builds each time.

Ran tens of thousands of dollars a day on Meta with most of it behind one high-performing creative (showing five stars in the description) that Meta had approved. Meta later took it down without warning during 'Q5' (post-Christmas to end of January, the best season for subscription apps). Campaigns had to be paused, retooled and relaunched. Lesson: diversify creative like an investment portfolio, no single creative above roughly 30% of spend.

Built an in-house creative playbook first, then hired an agency to run it; Pray.com is among the agency's most aggressive clients on creative testing. Inspiration comes from other industries (gaming, dating) rather than their own category. Targeting is left broad so Meta's algorithm finds the audience, from the curious to weekly churchgoers; Meta stays the core channel after testing Google, TikTok and others because it is the most consistent post-IDFA.

The TV-buying platform supplied by-minute data on when ads aired and helped with media mix modeling for broadcast; the team blended it with performance marketing data. Used media mix modeling from the start rather than relying on deterministic attribution.

Every report or review, positive or negative, gets a response from a person: a dedicated support lead (Earl) runs it, and the CTO or CEO (Steve) gets on a call with users, including those complaining that the app just wants money. Labor-intensive, decided from day one. Paired with keeping a free version for people who expect faith content to be free.

No separate roadmaps and no engineers dedicated to growth: one living roadmap that accounts for all stakeholders, shows priority shifts and models the cost of delaying items. Performance marketing sat under the CTO at first, so engineering felt marketing's urgency; engineers move to wherever the need is.

In 2019-2020 image ads in the Facebook feed performed well. Meta reps pushed video; early video ads (far costlier to make) didn't resonate. After TikTok became mainstream and from listening to Meta's quarterly earnings calls, started investing in Reels early; Reels now carries most delivery across Meta's apps.

Onboarding asks questions (a light version of Noom's long quiz funnel) to understand where the user is in their faith journey and tailor the post-onboarding experience toward healthy spiritual habits.

Treat subscription as the core revenue line, ads as secondary, and one-time purchases/upsells (e.g. books) as a third, following Duolingo's subs + ads + consumables model. Revenue is the North Star; features are prioritized by which revenue line they grow, and value is shared with pastors on the platform.