Big Brother 28 premieres tonight, and I built a thing.
BBFantasy is a fantasy draft league for Big Brother. You draft the house, make a quick pick before each episode, and compete with your friends all season while the app does the scoring. It is live now at bigbrotherfantasy.com. It is a sibling of SurvivorTribes, my Survivor fantasy app, forked from the same codebase, so I thought I knew exactly what I was building.
Then I looked at the data and got humbled. That is the part worth writing about.
I built the wrong thing first, confidently
I built the league features first. Drafts, scoring, leaderboards, the whole commissioner setup, the stuff a fantasy app is supposed to have. I was proud of it.
Then I opened the analytics and found the uncomfortable truth: about 98.9% of the people who show up never log in. They arrive on episode nights, poke around houseguest profiles, and leave. The single stickiest feature in the whole app, a tool that compares two houseguests, was one almost nobody could even find. I had built depth for a tiny logged-in core while the actual crowd, the superfans who show up to argue and predict, had nothing structured to do.
So the honest question stopped being "what league feature do I build next" and became "how do I make the free, no-login side worth coming back to every single week." That reframe is the whole strategy. The opportunity was never new audience. It was funnel conversion and a weekly habit.
A North Star that refuses to flatter me
When the audience is 99% anonymous, signups are a vanity metric. So the North Star for BBFantasy is Weekly Engaged Players: the count of unique people who take at least one real action in a given episode week. A pick, a poll vote, a compare. Active only, on purpose, because counting passive page views would let a good Google week inflate the number without anyone actually forming the habit I care about.
It is a deliberately unflattering metric. Right now it is near the floor: weekly retention sits in the single digits. But a North Star that is currently terrible is a good North Star, because all of its room is still ahead of it. It also spans free and paid, which keeps me building the funnel instead of polishing a paywall that 99% of visitors never reach.
The feature I protected against its own score
I ran RICE on the backlog to rank it honestly. The cheap, high-reach activation work rose to the top, as it should: an episode countdown, surfacing that compare tool everywhere, a homepage that offers something to do in the first three seconds.
The one thing I am shipping out of rank order is a named weekly ritual called Thursday Picks: a no-account poll plus a solo pick that anyone can play in a couple of taps, live every episode week. It scored fifth. I am treating it as a first-wave must-ship anyway, and I wrote down exactly why: it is the only item that moves the hardest input, weekly retention, and my whole North Star depends on it. Good prioritization is not blind obedience to a spreadsheet. It is knowing which single override you are making and being able to defend it out loud.
The pre-mortem named the real threats too, and they are not technical. The likeliest killer is me: a solo builder with too many "Now" items shipping all of them half-done. The most Big-Brother-specific one is that a player's picks get evicted from the house and they leave regardless of what I build, because reality TV evicts your reason to care. Naming those early is how you build the mitigations in instead of discovering them in September.
The bigger thing
I ran the same 16-step product process on this that I run for clients. JTBD, personas, a North Star, an opportunity tree, RICE, a roadmap, a pre-mortem. Pointed at my own reality-TV habit, on nights and weekends, mostly solo with coding agents doing the heavy lifting alongside me.
That is the part I keep coming back to. I can take an idea from "wouldn't it be fun if" to a live, instrumented, strategy-backed product in days now, and hold it to the same rigor I would bring to any team I work with. It is the most fun I have had making things in years, and it has changed what I think one product person can ship.
If you watch Big Brother, come draft a house with me. And if you are building something and want a product leader who can actually ship, I would love to talk.
Skills for this topic
Read next
I built a 16-step product strategy process from skills I use with clients. Then I ran it on k8mak.com. The biggest finding: the site I thought I was building was not the site I needed.
I built k8mak.com the same way I build products for clients: JTBD research, OKR alignment, opportunity solution trees, and a 15-step product strategy process. Here is what I built, what I cut, and why.
AI product strategy is not about what the model can do. It is about what the user needs done. The distinction sounds obvious. It is not. Here is the framework I use to tell the difference.
Want to work together?
I help teams ship better products. Let's talk about your situation.
Get in touch