Five apps for one Friday night

Pedro, 21, spent an entire Friday trying to find something to do: messaging during class, calling two friends after dinner, scrolling WhatsApp groups. He ended the night at home, gaming. Letícia, 26, knew there was a party that weekend, a friend had mentioned it days earlier, but the details never arrived: no location, no time, no way to find out. Fernanda, 28, organized events for a living and rebuilt the same machine every single time: a Facebook event, a Sympla listing, payment via bank deposit or PicPay, and promotion across Instagram, WhatsApp, Facebook, Snapchat, Twitter and Tumblr.

The founding team at Ensight saw the gap. Nobody owned the full journey of an event (before, during, and after) in one product. Sympla sold tickets. WhatsApp moved logistics. Facebook had the social graph but had abandoned events. Instagram had discovery but no way to act on it. PicPay moved money with no connection to what the money was for.

I joined with a blank slate: no brief, no wireframes, no prior research. My job was to define what the product should be before anyone touched a screen.

Three personas from primary research, each with a documented day-in-the-life. Different users, the same fragmentation beating all of them.

13 products mapped across four functional layers: event discovery, social coordination, payment, and location. No single product owned more than two.

I interviewed users between 19 and 30 (students and young professionals, the product's target) along with event creators and venue owners on the supply side. Three personas came out of that work, each with a documented day-in-the-life journey.

A distinction shaped everything after it: what users said, what they actually needed, and what they didn't know they needed.

That third column became the product thesis. Nobody asked for unification because nobody imagined it as possible. Every user had built a personal workaround system across three to six apps and accepted the friction as the cost of having a social life.

The second finding shaped the supply side: event creators had a trust and logistics problem, and discovery had little to do with it. Fernanda's six-platform routine existed because no single platform was accountable for the whole event, and an attendee buying a ticket from an unknown producer was making a blind bet: no history, no ratings, no accountability anywhere.

Before the pandemic, the Brazilian events sector moved roughly R$250 billion/year in corporate events and R$17 billion in social events, employed about 6 million people, and represented around 4.3% of national GDP, growing about 6.5% per year (Abrafesta via G1; Abeoc/Sebrae).

Ensight set concrete targets before design started. Those numbers meant the product could not be a slow-burn social network that monetizes years later. It needed transaction revenue (ticket sales and in-event consumption) from early on. This constraint decided the central strategic question before it was even asked: which side of the marketplace to build first.

My first instinct was to build the consumer side first: get users in, make discovery good, then bring producers to the audience. That's wrong for a marketplace, and the business targets made it obviously wrong. A producer with no audience on a new platform has no reason to create events there. The product would die waiting for an audience that never came, all while burning acquisition budget against a R$2.00 CAC ceiling.

One producer brings their entire existing audience. One consumer brings only themselves.

The only way through was giving producers something worth migrating for before the audience existed: ticket sales, event analytics, page infrastructure, professional visibility. Revenue for producers from day one, even with a small crowd.

Private events between friends, individual professionals, and the companies that run events: three fundamentally different cases.

What shipped: three account types with hard boundaries. The partnership model connected the tiers: a football club hosting a concert partners with the production company running the show and lists performers as professional partners. Each entity keeps its own accountability and visibility.

The same structure served webinars: one company, multiple professional speakers listed as contributors.

In 2020, distance was the only ranking signal most event apps used. A rave and a classical concert at the same distance looked equally relevant to everyone. Momentum ranked the Explore feed on five inputs instead.

Producer credibility was the variable that mattered most and cost the most to design. It combined verification status, event history, ratings on past events, and recency of activity. The goal: make bad production expensive. A new account, even verified, ranked below a producer with ten well-rated events.

The priority matrix scored 14 problems on importance and viability. Most decisions followed the scores: the top of the table shipped in the core MVP, and anything that needed user volume to matter (chat, video feed, real-time location) was deferred regardless of quality.

In-event consumption, the comanda (attendees order and pay inside the app), scored 1/1: lowest importance and lowest viability of all 14 rows. By the framework's logic it should never have been built. It shipped as a late MVP feature anyway, because the business model outweighed the user score. Users didn't care, and the matrix was right about that. But producers did, and the R$2.70 revenue target needed more than ticket fees. The comanda was never a user feature. It was a supply-side acquisition tool wearing a user feature's clothes.

The matrix was a tool, not a rulebook. In-event consumption scored lowest on both axes and shipped anyway: the business model overrode the score.

Anything that needed users to work was deferred. Anything that made producers money shipped, even, once, against the matrix.

Four months from zero to MVP scope, documented before the first screen.

Someone in a loud venue, holding a drink, one hand, getting interrupted: possibly stressed, possibly not fully literate in the moment, possibly with a physical impairment. Discovery mapped five context states, and they all pointed at the same user.

The constraints that came out of that: content visual over textual (the target won't read paragraphs in a nightclub), audio banned as an information channel entirely, every primary action within one-hand reach, interfaces recoverable from any wrong tap in one action.

Navigation weight followed documented priority, not opinion, mapped onto the business value chain.

This was also the project where I worked as the only designer, reporting directly to the two co-founders and working daily with two engineers translating specs into product. That ratio, one designer to a small founding team, is what pushed me toward handoffs detailed enough to build from without a follow-up conversation.

Identity built alongside the product, not applied after: "uma plataforma social completa focada em eventos de todos os tipos e tamanhos."

Event page, the flagship. Every field from the spec in one scroll: producer identity and event history (the trust chain from §05), date and location with a spatial mini-map (the same layer Momentum ranks on), tags plus going/interested counts feeding producer credibility back into Momentum, a comment thread for social proof, and an in-app buy bar pinned to one-hand reach, the supply-side revenue that justified building producers first.

Explore / Momentum. The featured Momentum card surfaces producer credibility and engagement velocity as one glanceable unit; friend presence rides above it as avatars; the ranked feed sits below. On the map, pin size renders interaction volume in space, and a tap-to-preview card opens the full event page without leaving the map.

Onboarding fork. Two doors, two commitments. Basic signup gets a live product in seconds; verification unlocks the commercial stack: pages, unlimited sales, categories, analytics. No one is blocked at the door, and the people who bring events have a clear reason to verify: the supply-first bet, encoded in a screen.

Comanda: order and pay inside the event. Every item is a one-tap add, no forms, no card entry mid-event. Consumption accrues to an open tab the producer can capture instead of losing it to cash and card machines. This is the 1/1 feature from §07: scored lowest on the matrix by users who didn't care, shipped anyway because producers did.

meaple

meaple

meaple

Primary

#00C0FD #007CFF #5500FA #310097

Secondary

#FFFFFF #1C1C38

The mark reads as an "M" and a butterfly at once. The gradient runs cyan → blue → violet → deep purple: a blue-forward system, not the "violet/purple" the old brand doc called it.