You can read every market report on Indian wedding photography and still build the wrong thing. We know, because we nearly did. What corrected us wasn't research — it was sitting with the studios in our PhotoSelect pilot and watching how they and their clients actually behave. A few of those lessons rearranged the product. Here are the ones that mattered most.
The gallery was never the bottleneck
We came in assuming the hard, valuable problem was the gallery — making photos look beautiful, load fast, feel premium. It is a solved problem, and nobody in the pilot was losing sleep over it. What they were losing was the three weeks after delivery, spent asking for the balance. The product's centre of gravity moved almost immediately: from how photos are shown to how money and files change hands. We've written separately about why we landed on pay-to-unlock. The pilot is where that conviction came from.
A family is a quorum, not a client
Software loves a single user with a single account. An Indian wedding has neither. The bride decides with her mother, her sister, an aunt, sometimes the groom's side too — and they do it on four different phones, at four different times, often in four different cities.
Our early selection flow quietly assumed one person choosing. It was wrong in a way that's obvious in hindsight and invisible from a spec. Watching a shortlist get assembled by a whole family, asynchronously, taught us that "multi-device selection" isn't a nice-to-have feature in India. It's the actual shape of the decision.
"Bad network" is the default condition
Every studio told us some version of the same story: the venue had no usable WiFi, the hall basement had no signal, the guests were on mid-range Android phones nursing their data. We had treated poor connectivity as an edge case to handle gracefully. It is not an edge case. It is the normal operating environment for the most important moments.
That reframing changed how we think about the whole guest experience. The question stopped being "does it work when the network is good?" and became "does it work when the network is bad?" — because bad is the baseline.
The thing they valued wasn't a feature
Here's the lesson that surprised us most. We expected studios to ask for more automation — more auto-reminders, more auto-everything. Some of that helps. But the thing they reacted to most strongly was simpler: knowing. Knowing which album was opened, who had selected, who had paid, who had gone quiet. Not a dashboard for vanity — a clear answer to "where does this stand?" across every event in flight.
We'd been thinking of visibility as a reporting feature. The studios treated it as the product. A photographer who can see the state of every album doesn't feel behind, and doesn't have to call a client to find out. That calm is worth more than any individual automation.
What we're still getting wrong
We won't pretend it's all figured out. Pricing in this market is genuinely unsettled — subscriptions, per-event packs, and per-photo models all coexist, and studios have strong, differing instincts. We're testing, not asserting. And there's a real tension between automating the awkward parts of getting paid and leaving studios in control of their own client relationships. We lean toward control, and we keep checking that instinct against what actually happens.
That's the honest state of it. The pilot didn't hand us a finished product. It handed us a sharper set of questions — which, for a first product from a small lab, is exactly what it was supposed to do.