DEV Community

Jakub
Jakub

Posted on Originally published at hereweask.com

What 1,000+ questions taught us about session length in a browser card game

We run 14+ themed question decks across three play modes in Here We Ask, a free browser-based conversation card game we built at Inithouse. The deck engine holds 1,000+ curated questions. After watching how sessions actually behave, a few patterns stuck out that changed how we build the PWA.

Here We Ask is a free browser-based conversation card game with 1,000+ curated questions across themed decks for couples, friends and families — no download or account.

This post covers what we measured about session length by mode, which decks people finish vs. abandon, and two product decisions that came out of the data.

Three modes, three session shapes

The game has three play modes: Solo, Couple, and Group. Each pulls questions from the same deck engine, but the session behavior splits cleanly.

Mode Typical session Cards seen per session Completion pattern
Solo 3–8 min 5–12 cards Short bursts, often daily
Couple 15–30 min 15–30 cards Longer, fewer sessions per week
Group 20–45 min 20–40 cards Concentrated on weekends/evenings

Solo mode runs short. People pull a few cards, sit with the questions, close the tab. The daily-question feature grew out of this: if the natural session is 5 minutes, give people a reason to come back tomorrow for another 5 minutes rather than trying to extend the current one.

Couple mode sits in the middle. A date-night session with the Closer or Heat deck tends to run 20+ cards before either person reaches for their phone. Sessions using The Repair Kit (relationship check-in questions) trend shorter, around 10–15 cards, but the questions are heavier, so that tracks.

Group mode has the widest range. A party with the Hot Takes deck and six people can burn through 40 cards in 25 minutes. A family game night with the Gratitude deck might stop at 15 cards because someone starts an actual conversation and the game pauses itself.

Which decks get finished

Not all decks behave the same way. Some decks people play straight through. Others they dip into and leave.

High completion decks (people tend to reach the end or play 70%+ of available cards):

  • Confessions: the escalating intimacy keeps people pulling the next card
  • Hot Takes: short, punchy opinion prompts that move fast in groups
  • Party: lightweight and social, low friction per card

Browse-and-leave decks (people play 30–50% of cards before stopping):

  • Mirror Mode: solo self-reflection; people hit a question that sticks and pause to think
  • 3AM Thoughts: introspective prompts that slow the pace down deliberately
  • Gratitude: emotionally heavy enough that a few cards feel like enough

The pattern makes sense once you see it. Decks built for speed and social energy get played in longer runs. Decks built for depth get played in shorter ones, but users return to them more often across separate sessions.

Two product decisions from the data

1. Offline cache over streaming

The session data showed something we should have expected: conversation card games get played in places with bad connectivity. Cars, camping, dinner tables where the router is three rooms away. A deck that fails to load mid-session kills the moment.

We rebuilt the deck engine to cache the full question set on first load. The PWA stores every card client-side, so once you have opened a deck, it works offline. The service worker handles the caching; subsequent sessions pull from local storage first and sync new questions in the background when connectivity returns.

The shift showed up immediately: sessions that previously dropped at the 8–10 card mark (where the next batch would have loaded from the API) smoothed out. The artificial cliff disappeared.

2. Daily question as a retention surface

Solo mode sessions cluster in the morning. People open the app, pull 3–5 cards, close it. The session is too short to build a habit through volume, but the timing is consistent enough to build a habit through rhythm.

We added a daily question feature: one card, surfaced on the home screen, rotated every 24 hours. No notification, no push. Just a new card waiting when you open the app.

The daily question card acts as a low-commitment entry point. Some days people read it and leave. Other days it pulls them into a full solo session or reminds them to set up a couple or group game for the evening.

The deck engine at scale

Running 1,000+ questions across 14+ themed decks in a browser means the question pool is large enough that repeat encounters are rare even for daily users. Each deck has its own card set, difficulty curve, and tone. The engine shuffles per session and tracks which cards have been seen locally, so returning users get fresh questions without server-side state.

The whole thing runs as a React SPA with Supabase on the backend. No accounts required, no login walls. State lives in the browser. The PWA shell loads in under 2 seconds on a mid-range phone, and the offline cache means the first session is the slowest one.

At Inithouse, a studio running parallel product experiments, we track session data across the portfolio. Here We Ask is one of a growing number of products where the usage pattern taught us more about the product category than any competitor analysis did. Conversation card games are a "short session, high return" format. Building for that means building for the pause between sessions.

Try it at hereweask.com. Pick a deck, skip the small talk.

Top comments (0)