Personalization is the product. Every member sees a different homepage, a different roster, a different register.
The fashion-curious member who'd otherwise drown in choice. Discovery solves "what's good?" — Tailored solves "what's good for me?" Same primitives, different output. Your homepage isn't a homepage; it's a portrait.
A taste graph per member, built from explicit signals (onboarding swipes on classic + Fobou pieces, followed curators, designers, votes, purchases) and implicit ones (dwell, return, drop participation). ML-tuned, re-fits on every interaction. The graph drives every surface: the homepage feed, the residency roster you see first, the register density, the language register, the campaigns ranked by you-fit.
Compounding signal. Day-1 onboarding asks five questions; week-2 swipes calibrate; month-2 onwards the system reads. Personalization scales smoothly — you can always see the canonical Fobou ("show me everything"), but the default is always yours.
Different pitch, different moat. Avenue 3 (Discovery) bets on editorial taste; Avenue 6 (Tailored) bets on data graph. The two compound: Discovery makes the canon legible; Tailored makes the canon yours. Defensibility scales with engagement — once your graph knows you, switching costs are personal, not technical.
Same site, same primitives, same drops. Two members, two completely different homepages. The taste graph reads, the surface adapts. The longer you stay, the less default you see.
Your taste graph is built from two kinds of signal — what you tell us, and what we infer from your behaviour. The graph is the product. Everything else is what we render with it.
Three people, three taste graphs, one cohort of drops. The system reads each. The output is the same primitives — feed, residency roster, ranked campaigns — assembled per person.
Each surface adapts on the same axis as the graph. The site you see is constructed in real time. The canon view is always available — one click — for those days you want everyone's homepage instead of yours.
Three pieces, ranked by graph match × recency × campaign urgency.
Cohort cards reordered by register match. Same six houses, different priority.
Ranked first by you-fit, then by days-remaining. Yours-firsts, never theirs.
Curators surfaced match your saved-lineage shape. Closest first; adjacent second.
The lineage chain you see opens at the era you've spent the most time in.
The diary's seven entries are reordered. Still same week, still same picks; you start where you'd want.
Quieter graphs see less per scroll. Louder graphs see denser cards. Spacing adapts; type ladder adapts.
Voice shifts subtly. Some readers see fewer adjectives; others see more. Captions length scales.
Sketches that match your taste rise. You vote on what's near you, not on a global average.
Sometimes today's piece is from your direction; sometimes deliberately not — calibrated for stretch.
Push only when the new drop matches your graph above a threshold. Else: it'll wait.
Your year, read back. Spotify Wrapped's vibe, Letterboxd's discipline. December.
Personalization done well makes a member feel read. Done badly it makes them feel watched. The difference is who controls the data, who can see it, and how easy it is to walk away.
Export anytime, full graph + signal history, in a format competitors can read. No lock-in.
Every implicit signal is visible in your settings. "We think you lean here because of these eight saves."
Down-vote a direction. Pin a register. Forget a season. You correct the graph; the graph corrects the site.
Whenever you want everyone's Fobou — for an evening, for a season, forever — switch off. Comes back on when you ask.