← Back

Volleyball Community App

Making a recreational volleyball game reliably happen before building the rest of the marketplace.

Project overview

  • Type: Product · 10-week 0→1 · mobile (iOS + Android)
  • Project type: 0→1 · Community Marketplace · Trust & Reliability · Activation & Retention
  • Role: Lead Product Designer · Opportunity Framing · UX Research · Experiment Design
  • Methods: Diary study · JTBD interviews · fake-door test · A/B
  • Tools: dscout · Otter.ai · Figma + Dev Mode · Amplitude · Statsig
  • Case thesis: Designing a recreational volleyball app around one reliable, skill-balanced game that happens as planned, then expanding to coaches and courts from the attendance trust it builds.

The context

Recreational volleyball is organized today across scattered group chats, spreadsheets, and noticeboards, with no shared record of who plays, at what level, or who reliably shows up. A new app could connect players, coaches, and courts, but a multi-sided community lives or dies on whether its core interaction works before its breadth does.

The problem

Games fall apart before they start. In an analysis of 120 games proposed across 9 existing player groups, only 61% reached the minimum players needed to play, and soft RSVPs produced a no-show rate around 38% — that is, of the players who gave a soft "yes," roughly 38% didn't appear, leaving a 12-person game (6v6) about four to five players short on average, too few to field two full sides (behavioral, group-chat analysis). A first game that was short-handed or badly mismatched on skill was the most common reason players stopped coming back (attitudinal, survey).

The goal

Raise the share of proposed games that are played as planned with balanced teams, and turn a good first game into repeat play, measured by games-played rate, no-show rate, and player retention, rather than by the number of features shipped.

Empathize — Games fell apart before they started: only 61% of proposed games reached enough players, and no-shows ran near a third of every soft-confirmed roster

In this section: Research foundation · Key insights

Research foundation (method)

  • Phase 1 — Player interviews (n=18, ~35 min, recruited via dscout, transcribed in Otter.ai): how players currently find and commit to games, and what made them quit a group.
  • Phase 2 — Survey: ~918 panelists invited; 180 analyzable responses (19.6% response rate); ~123 completed every item (13.4% completion rate). Attitudinal percentages are computed on the 180 analyzable responses; question types (select-all vs single-select) are labeled per question in the appendix.
  • Phase 3 — Existing-group analysis (120 games across 9 chat groups, 8 weeks): coded each game for whether it reached minimum players and counted no-shows against soft RSVPs; the same cohort's return-play over the window is the source of the 44% baseline retention used later.
  • Phase 4 — Prototype pilot (Amplitude-instrumented, mobile, ~90 players, 14 organized games, 8 weeks): behavior on the rebuilt match loop, including the commitment A/B.

Key insights

1. The soft RSVP gives organizers no reliable count, so games collapse or overfill. Group-chat "who's in?" threads produced commitments that did not hold, so organizers over-invited to be safe and games still ran short or lopsided. Triangulation:

  • Behavioral: 61% of proposed games reached minimum players; no-show rate (share of soft-confirmed who didn't appear) around 38%.
  • Verbatim (P3, weeknight organizer, 5-yr group admin) — coded: Unreliable commitment: "Eight people say yes in the chat, four actually turn up, and now we can't play a real game."

2. A bad first game is what loses a player. Players who joined a game that was short-handed or far off their skill level rarely came back. The single experience set whether the community kept them.

  • Attitudinal: the top reason survey respondents gave for leaving a group was unreliable or mismatched games, ahead of cost or location.
  • Verbatim (P11, beginner, joined and quit two groups) — coded: First-game churn: "My first night, half the people didn't show and the rest were way better than me. I just never went back."

3. Players want to commit, and casual players balk at anything that feels like a contract. Interviewees described wanting a trustworthy headcount while resisting anything heavy. The pull toward reliability and the resistance to friction point the same way: commitment has to be light.

  • Verbatim (P7, casual player) — coded: Low-stakes commitment: "I'll confirm I'm coming, but I'm not putting money down to play a pickup game on a Tuesday."

Dashboard — How recreational games fail today

How recreational games fail today
Scope: 120 proposed games · 9 player groups · 8 weeks · coded
Guiding question: Why don't proposed games get played as planned?

  Reached minimum players ................... 61%
  Soft-RSVP no-show rate .................... 38%
  Avg players short of the 12 needed ........ ~4.5
  Top reason players quit a group ........... unreliable / mismatched games

Key Insight: The failure is upstream of any feature — the headcount can't
be trusted, so games either collapse or run unbalanced.

Define — The first job was making a single game reliably happen with players near the same level

In this section: POV · How Might We · Principles · Insight→decision map

POV statement. Recreational players need a game that reliably happens with players near their level, because a single short-handed or mismatched game is what stops them from coming back.

How Might We

  1. How might we give an organizer a trustworthy count of who will show up?
  2. How might we match players into balanced games by skill without gatekeeping beginners out of the community?
  3. How might we cut no-shows with a commitment light enough that casual players still join?

Design principles (each traceable to an insight)

  • Reliability before breadth. The first loop makes one game dependable; coaches, courts, and tournaments expand from there. (the goal)
  • Liquidity before breadth, too. A reliable game still needs enough nearby players to fill it, so launch concentrates density in one group, court, and night before spreading. (cold-start reality — see below)
  • Balanced rosters, beginners included. Skill bands place players into level-appropriate games and keep beginners in the pool. (Insight 2)
  • Light commitment. No-shows are cut with reputation and confirmation, keeping money out of a casual game. (Insight 3)
  • Trust from attendance. Verified attendance history, and ratings segmented by quality such as leadership and teaching, are earned from real games. (Insight 1)

Insight → decision map

Insight (from Empathize) Concrete design decision
Soft RSVPs don't hold; only 61% of games filled A game has confirmed spots, a minimum-to-play threshold, and a waitlist, so the organizer sees a count that means something
A mismatched first game loses the player Games carry a skill band; matching keeps a roster within one band so beginners and advanced players land in fitting play
Players want commitment without heavy friction A non-monetary reliability score plus a one-tap confirmation reminder, instead of a deposit

Ideate & Craft — The product was built around one reliable game before coaches and courts

In this section: Design execution · Bootstrapping liquidity · Before → after · Other deliverables

Design execution

  • The game object — an organizer sets time, court, skill band, and minimum-to-play; players claim confirmed spots and join a waitlist when full, so the headcount is real.
  • Reliability score — a lightweight, non-monetary reputation built from attended-versus-missed history, surfaced on each player so organizers and teammates can trust a confirmation.
  • One-tap confirmation — a reminder before each game asks players to confirm or release their spot, so the waitlist fills empty seats in time to still play.
  • Skill bands and segmented ratings — players self-place and are adjusted by post-game ratings; coach and teammate ratings are segmented by leadership and teaching so the signal stays specific.
  • Sequenced expansion — court booking, coach discovery, and tournaments are designed to extend the same trusted game and attendance record, shipping after the core loop holds.

Bootstrapping liquidity (the cold-start problem, addressed head-on)

A 12-person game has the hardest liquidity requirement in consumer marketplaces: twelve people, one court, one time slot. Reliability is worthless if there aren't enough nearby players to fill the first game, so the launch model is density-first, not geography-wide: onboard one existing group at a time (importing its players and their informal reputations), anchor it to a single recurring court and night, and only open a second night or court once the first reliably fills. The same logic that orders features — make one loop dependable before adding breadth — orders the rollout: make one game full before adding the next. Coaches, courts, and tournaments come later precisely because they need this density to exist first.

Before → after

Before (group chat) After (the app)
Headcount "Who's in?" with no real commitment Confirmed spots, minimum-to-play, waitlist
Cutting no-shows Guilt and reminders Reliability score + one-tap confirm/release
Skill matching Hope it works out Games carry a skill band; rosters stay within one
Trust signal Reputation by word of mouth Verified attendance history and segmented ratings

Other deliverables

Built in Figma with Dev Mode handoff: the game-object component set, the reliability-score and confirmation states, the skill-band matching flow, and a clean, sporty visual system with a high-energy palette tuned to stay readable outdoors on a phone at a court.

Dashboard — Confirmation and reputation make games hold

Confirmation and reputation make games hold
Scope: Last 8 weeks · mobile · pilot (~90 players, 14 games)
Guiding question: Did light commitment make games happen as planned?

  Games reaching minimum players .... 61% → 86%
  No-show rate ...................... 38% → 14%
  Games within one skill band ....... — → 91%

Key Insight: Confirmed spots plus a reputation signal raised filled games
to 86% and cut no-shows to 14%, without charging players to commit.

Prototype / Test — A deposit cut no-shows to 8% but halved sign-ups, so fewer games happened; a reputation score fixed no-shows while keeping liquidity

In this section: The experiment · What it taught

No-shows were the clearest problem, so the first fix was a refundable deposit to force commitment: players put a small amount down to claim a spot, refunded on attendance. It was A/B tested against the non-monetary reliability score in Statsig across the pilot.

The failed variant. The deposit worked on the metric it targeted, dropping no-shows to 8%, the lowest of any variant. It also cut sign-ups per game by 52%, because casual players would not put money down for a pickup game, so most games no longer reached minimum players and fewer games were played at all. The deposit optimized one number and starved the liquidity the community runs on.

The deposit fixes no-shows and starves the games
Scope: Statsig A/B · mobile · pilot · 2 variants
Base: 14 games / ~90 players (~7 games, ~45 players per arm)
Guiding question: Which commitment mechanism maximizes games played as planned?

  Variant A — Refundable deposit
    No-show rate .................... 8%
    Sign-ups per game ............... −52% vs B
    Games reaching minimum players .. fell below baseline

  Variant B — Reputation score + confirmation
    No-show rate .................... 14%
    Sign-ups per game ............... baseline
    Games reaching minimum players .. rose to 86%

Read with care: ~7 games per arm. This is a DIRECTIONAL result; the decision
rested on the clear, consistent trade-off — the deposit fixed the wrong
number — not on statistical power.

Key Insight: A 14% no-show rate with a full roster beats an 8% rate with
half the players; reputation carried the commitment money couldn't.

What it taught. In a casual community marketplace, a commitment mechanism has to match how much the user has at stake; monetary friction that suits high-intent transactions drains the liquidity a low-stakes community depends on. The reputation-based model shipped.

Outcomes & reflections

In this section: Causal chain · Limitations · Competitive context · Reflections · References

Causal chain (pilot, ~90 players, 14 games, 8 weeks)

The reliability score and one-tap confirmation cut the no-show rate from 38% → 14%, which raised the share of proposed games that reached minimum players and were played as planned from 61% → 86%, so more players had a good, balanced first game. That lifted 30-day player retention from 44% → 61% (the 44% baseline measured from the same group-chat cohort over the study window) and grew repeat play, with median games per active player per month rising from 1.3 → 3.1.

Metric Before After Δ
No-show rate 38% 14% −24 pts
Games played as planned 61% 86% +25 pts
Games within one skill band — 91% balanced rosters
30-day player retention 44% 61% +17 pts
Median games per active player / month 1.3 3.1 ~2.4×

Scale note: in a community marketplace, repeat play compounds — a player who returns brings teammates, so lifting games-per-player from 1.3 to 3.1 multiplies the matches the network can fill.

Limitations (stated, because a portfolio claim is only as strong as what it concedes)

  • Small pilot. ~90 players, 14 games, ~7 games per A/B arm, 8 weeks. The no-show and games-played effects are clear and consistent; the 30-day retention lift is directional, not statistically powered.
  • Single-cohort baseline. The before figures come from existing groups, not a held-out control, so they describe the status quo the design improved on rather than a randomized comparison.
  • Liquidity is the gating risk. The whole model assumes enough nearby players to fill a 12-person game; the density-first rollout mitigates this but does not remove it, and a thin market is where the product would fail first.
  • Reliability ≠ safety. A reputation score handles flakiness, not the real-world fact that strangers meet in person to play; verification and reporting are required adjacent work, scoped but not the focus here.
  • Reputation can be gamed. Self-placed skill and attendance scores need guards (organizer confirmation of attendance, cooldowns) so the trust signal stays honest.

Competitive context

Organizing pickup volleyball is a crowded space, and skill-level matching is already table stakes: Volo Sports runs leagues and pickup with skill divisions, GoodRec hosts open-play volleyball with notifications, ENDALGO organizes volleyball groups with RSVPs and payments, Grab a Game uses explicit skill bands, and several pickup apps match by location and skill filter. Because skill-banding is common, it is not the wedge. This product's differentiation is the reliability layer: a non-monetary reputation built from verified attendance, a confirm/release loop that protects the headcount, and a deliberate sequence that makes one game dependable and full before adding breadth. The bet is that reliability, not feature breadth or yet another skill filter, is what turns a first game into repeat play.

Reflections (transferable principles)

  • In a community marketplace, the retention lever is the reliability of the core transaction, so the first build should make one loop dependable before adding adjacent features — and concentrate liquidity before spreading it.
  • A commitment mechanism has to match the user's stake; reputation can carry commitment where money would drain the liquidity a casual community needs. (The A/B is the evidence: the deposit fixed no-shows and starved the games.)
  • Skill matching earns trust by balancing a game so beginners land in level-appropriate play, which keeps the community larger than a strict skill filter would.

References (for the competitive claims)

  • Volo Sports — adult recreational leagues and pickup with skill divisions (volosports.com).
  • GoodRec — open-play volleyball sessions with hosts and notifications (goodrec.com).
  • ENDALGO — volleyball group organizing with RSVPs and payments (endalgo.com).
  • Grab a Game — pickup volleyball with explicit skill bands (grabagame.com).
  • Pickup Sports — location- and skill-filtered pickup games (App Store).