Skip to main content
The most expensive game development mistake is building something players don’t want. This playbook validates the concept — player desires, core mechanics, art direction, feature priorities, and overall appeal — before you commit engineering and art resources, so your game design document is grounded in real player feedback. This is a pre-development sequence: each step informs the next, producing a prioritized, player- validated concept. It’s designed to be run by an AI agent (Claude, ChatGPT, Cursor, or any client with PickFu MCP / CLI / API access) or step-by-step by a human.

​
When to use this playbook

Use it whenSkip it when
You have a game concept but haven’t committed to developmentThe game is already built — use ASO to optimize the listing
You want to validate mechanics / art / features with real playersYou only need to test marketing assets
You can describe scenarios or show concept artYou have nothing concrete to show or describe yet
Audience targeting. These polls target mobile gamers (mobileyes) aged 25–44 and report by age and gender. Adjust to your game’s actual target players. Use list_available_targeting for options.

​
The sequence

1

Competitive research

Learn what players love and hate about existing games in your genre, and what would make them play daily.
2

Validate core mechanics

Rank gameplay scenarios to find the mechanic players find most fun.
3

Choose art direction

Rank visual style references to find the look that most attracts your players.
4

Prioritize features

Rank candidate features so you build the highest-impact ones first.
5

Validate the overall concept

A star-rating poll on the assembled concept confirms download intent before you build.
Sample size and cost. Defaults to 100 respondents per step for reliable ranking signal on pre-development decisions (these inform expensive build choices, so the extra confidence is worth it). On a tighter budget, 50 per step gives directional results. See sample size guidance.

​
Run this playbook with an AI agent

Copy prompt for AI

Paste this prompt into Claude, ChatGPT, Cursor, or any AI agent connected to the PickFu MCP server, CLI, or REST API. The agent will run the entire loop on your behalf — creating polls, reading responses, and iterating until a winning variation emerges.
You are running the PickFu "Game concept validation" playbook end-to-end.
Goal: validate a mobile game's mechanics, art direction, features, and overall
concept with real players before development begins.

Before starting, ask the user for:
- The game concept and genre (e.g. "4X strategy", "cozy farming sim")
- 3-4 gameplay-mechanic scenario descriptions (text)
- 3-6 art-style reference images
- A candidate feature list (5+ features)
- A game name + short description, and concept art for the final validation
- Target audience (default: mobile gamers 25-44; refine to the game's
  actual demographic if known)

Substitute [GENRE] with the user's genre in every question. Run in sequence:

1. COMPETITIVE RESEARCH (100 respondents, multi-question open_ended).
   Q1: "What do you like most about the mobile [GENRE] games you currently play?
   What keeps you coming back?"
   Q2: "What frustrates you most about current mobile [GENRE] games? What would
   make you stop playing?"
   Q3: "What kind of gameplay would you expect or want in a mobile [GENRE] game?
   What features would make you play every day?"
   Synthesize the responses into player desires, frustrations, and must-haves.

2. CORE MECHANICS (100 respondents, ranked, 3-4 options).
   Provide 3-4 mechanic scenario descriptions as text options. Question:
   "In a mobile [GENRE] game, which gameplay scenario would you prefer? Rank
   from most to least appealing, and explain why your top choice is the most
   fun." The winner is your core loop direction.

3. ART DIRECTION (100 respondents, ranked, 3-6 options).
   If the user doesn't have art references, GENERATE them: use generate_image
   to render the SAME subject/scene in 3-6 distinct art directions (e.g.
   realistic, stylized cartoon, pixel art, painterly, low-poly) so respondents
   react to style, not subject. Question: "Which visual style would you prefer
   for a mobile [GENRE] game? Which style makes you most interested in playing?"
   You can also feed the step-1 frustrations back into the prompts — generate
   styles that address what players said current [GENRE] games get wrong.

4. FEATURE PRIORITIES (100 respondents, ranked).
   Provide your candidate features as ranked options. Question: "Which of these
   features would you most want in a mobile [GENRE] game? Rank them from most to
   least important." This orders your development backlog by player demand.

5. CONCEPT VALIDATION (100 respondents, star_rating).
   Add the game name + description as context; upload concept art as the option.
   Question: "How likely would you be to download and play this mobile [GENRE]
   game? What do you like or dislike about the concept?"

   Success target: overall concept averages 4.0+ stars. Below that, revisit the
   weakest dimension (mechanics, art, or features) before committing to build.

Final report must include:
- Player desires / frustrations / must-haves from step 1
- The validated core mechanic (step 2) and art direction (step 3)
- A player-prioritized feature list (step 4)
- The overall concept star rating (step 5) and the top likes/dislikes
- A go / iterate / rethink recommendation for the game design document

Tools to use:
- save_survey + publish_survey         — create and launch each poll
- get_survey_responses                 — read responses
- generate_image                       — render art-style variations (step 3)
- list_available_targeting             — confirm targeting codes
- upload_media                         — art references, concept art
Where AI fits in this playbook. Two places: (1) synthesis — the agent turns the open-ended competitive research (step 1) into a structured map of player desires, frustrations, and must-haves that shapes every later step; and (2) creative generation — the agent renders the art-direction candidates (step 3) with generate_image, including styles that address the frustrations surfaced in step 1. The final concept validation (step 5) stays a single one-shot gate — it’s a go/no-go decision, not a loop, so re-testing it repeatedly just burns budget on a decision you’ve already made.
Want to run this manually? The same sequence is available as a one-click template in the PickFu app (Start the concept-validation playbook) with pre-filled poll URLs.

​
Step-by-step (human operator view)

​
1. Competitive research

SettingValue
Poll typeMulti-question (3 open-ended)
Q1”What do you like most about the mobile [GENRE] games you currently play? What keeps you coming back?”
Q2”What frustrates you most about current mobile [GENRE] games? What would make you stop playing?”
Q3”What kind of gameplay would you expect or want in a mobile [GENRE] game? What features would make you play every day?”
AudienceMobile gamers, 25–44 (report by age + gender)
Sample size100
What you’ll get: the player desires, frustrations, and daily-play hooks that should shape every later decision. See an example research survey →

​
2. Validate core mechanics

SettingValue
Poll typeRanked choice
Question”In a mobile [GENRE] game, which gameplay scenario would you prefer? Rank from most to least appealing, and explain why your top choice is the most fun.”
Options3–4 mechanic scenario descriptions (text)
AudienceMobile gamers, 25–44
Sample size100
See an example mechanics poll →

​
3. Choose art direction

SettingValue
Poll typeRanked choice
Question”Which visual style would you prefer for a mobile [GENRE] game? Which style makes you most interested in playing?”
Options3–6 art-style reference images
AudienceMobile gamers, 25–44
Sample size100
No art references yet? Use generate_image to render the same scene in 3–6 distinct styles (realistic, stylized cartoon, pixel, painterly, low-poly) so respondents react to the style, not the subject. Feed your step-1 frustration findings into the prompts to generate directions that address what players say current [GENRE] games get wrong.
See an example art-style poll →
PickFu ranked poll comparing three art styles for a mobile strategy game with vote share and AI insights

​
4. Prioritize features

SettingValue
Poll typeRanked choice
Question”Which of these features would you most want in a mobile [GENRE] game? Rank them from most to least important.”
OptionsYour candidate features (replace the defaults with your own)
AudienceMobile gamers, 25–44
Sample size100
See an example feature-ranking poll →

​
5. Validate the overall concept

SettingValue
Poll typeStar rating
Question”How likely would you be to download and play this mobile [GENRE] game? What do you like or dislike about the concept?”
OptionsConcept art as the option; game name + description as context
AudienceMobile gamers, 25–44
Sample size100
Success criteria: a confirmed core mechanic (step 2), validated art direction (step 3), a player-prioritized feature list (steps 1 & 4), and an overall concept averaging 4.0+ stars (step 5) — ready to inform your game design document. See an example concept-validation poll →

​
Troubleshooting

Not necessarily — diagnose first. Cross-reference the step-5 dislikes against your earlier results. If the mechanic (step 2) and art (step 3) scored well but the concept didn’t, the problem is usually how they’re combined or communicated, not the parts themselves. Re-pitch with a sharper description before abandoning.
That’s the playbook doing its job — it surfaces demand before you’ve sunk cost. Either build toward what players want, or understand the gap (sometimes the appealing mechanic is a hook and your intended one is the retention driver). Run a follow-up ranked poll if you need to disambiguate.
This validates the concept before development. The ASO playbook optimizes the store listing once the game exists. Run this one first; run ASO when you’re preparing to launch.