29. Summary B: Measurement, Review, and Reference

This Summary B is what each chapter in the main text refers to as “Appendix C.” Whenever the main text directs you to “Appendix C,” come here.

If Summary A provides tools to fill in, this is the workbench you open when that writing is done. Measure the completed design in front of real people, catch where it leaks, and trace it back through principles and terminology. This summary is not an answer key but a checkpoint for finding where your first screen drives users away.

Terminology corresponds one-to-one with the main text. A gauge is an input I change by hand; the instrument panel contains the output metrics produced as results. The North Star is just one number among them. Payment and revenue are treated only as output metrics, never as thresholds of the first experience. The ladder of first moments begins at T0 and ends at T10, the first return. All example numbers are hypothetical.

1. Instrument-Panel Definition Sheet and Causal Map

Purpose. Define what to watch—the instrument panel—from Chapter 22, then connect how those numbers are moved in Chapter 23 through causality: gauge → experience → instrument panel. Instrument-panel values are output metrics that cannot be pushed directly by hand. Gauges are inputs I can change directly on the screen. Keep your hands on inputs and your eyes on outputs. The North Star is a single number on the instrument panel. Payment and revenue are not thresholds of the first experience; place them only as output metrics, at the edge of the panel or much farther downstream.

Instrument-panel definition sheet (five output metrics + one North Star). Set revenue and daily active users aside for now. Choose small, nearby results for which the first screen is directly responsible. Record the following for each metric:

Exactly what the number counts / measurement point / current value (hypothetical) / mark only one of the five as the North Star
Candidates: pass rate by stage / rate of reaching the first core fun (activation) / first-task completion rate / next-day return (D1) / seven-day return (D7) / time spent in one session / first-share rate

Gauge-to-panel causal map (five gauges → experience → one panel metric). Place one instrument-panel metric at the far right, write five gauges that can move it on the left, and put the experience changed by those inputs in the middle. Your hands always stay on the left—the inputs.

Split by segment (do not judge from one average). For every group, separately record the instrument-panel value, the point of greatest leakage, and the gauge to adjust for that group. Example: people arriving through advertisements / people who sought the game themselves. If the two rows differ greatly, work on the group with the lower row.

Three-step alignment with business goals. A business goal is a distant output you cannot push by hand. Bring it down into one nearby instrument-panel metric that foreshadows it and one gauge you can touch on today’s screen, then align all three in one line. Example: seasonal revenue (distant output) ← D1 return 8% (nearby result) ← add a “See you tomorrow” farewell immediately after completion (input to change today). In this line, keep your hands only on the item at the far right—the input you can actually change.

How to fill it in.

Select five output metrics and explain in one line exactly what each counts (“D1 is the number, among 100 people who opened the game for the first time yesterday, who open it again today”). Record the measurement point as well.
Assign the North Star to only one of the five. It is the one number that most directly reveals whether the first experience did its job. If it is revenue or daily active users, it is too far away for the first experience to reach, so choose again.
Place one selected panel metric, such as D1, at the far right, then write five gauges that can move it on the left. Include only things you can directly change on today’s screen: timing of the first reward, timing of the login request, where the excuse to return is planted, how restarting works after the first failure, and the form of the farewell.
Split panel values by group, then move your intervention to the gauge of the group that leaks.
If you need to include payment or revenue, write it only as an output metric at the edge of the definition sheet or on a separate line. Even if payment appears at the far right of the causal map, it is an output that follows after value is felt, not a handle to pull by hand.

Short example.

Metrics: first-completion rate 30% (completion-save log) / D1 return 8% (first launch on the next day, ● North Star) / thirty-minute retention 25% (session log).
Causal map: wave and say “See you tomorrow” → the farewell promises a next time → D1↑ / leave one small unfinished care task → curiosity → D1↑ / name it and make it mine → attachment → D1↑.
You cannot touch tomorrow’s return. You can touch today’s farewell. Payment is not a handle anywhere on this map.

2. Measurement Plan and User-Test Questionnaire

Purpose. Design the iteration in which you revise from what you measured and measure the revision again. Measurement needs two eyes: qualitative evidence from directly watching and listening to people, and quantitative evidence from numbers formed by accumulated behavior. Quantitative evidence answers where the leakage occurs; qualitative evidence answers why.

Qualitative plan (interviews and observation). Choose one screen identified by the numbers, then record what five first-time users ask, what they only stare at, and where they get stuck.

Quantitative plan (funnel and segments). Record how many people remain at each stage: first screen 100 → first input → first response → first completion → next day (D1). Identify the steepest interval and decide whether to split it into ad-driven and organic arrivals.

Iteration sequence (paper → prototype → measured behavior). You can measure before everything is built.

Paper: ask people to point with a finger at a screen drawn on paper and see whether they read it as intended.
Prototype: use a mock screen that responds to presses and watch where they hesitate.
Measured behavior: use logs from real people to see where the actual leakage occurs.
Before changing anything, record the current number—the baseline. It is the starting point for every comparison.

Re-FTUE review (new, returning, and existing users). For each of three moments when someone becomes a beginner again, prepare one sentence that points out only “what changed,” rather than teaching everything again.

Major update — existing users facing a changed screen structure
Return after a long absence — users who have not played for some time
Season transition — users facing newly laid rules

Five-person interview questionnaire. Sit beside five people opening the game for the first time. Do not teach, do not intervene; only watch where their fingers stop.

(Before starting) “What do you think this game is about?” Check whether the promise made by the preview came across.
(First screen) “What do you want to do on this screen right now?” Check whether the identity is visible at a glance. Watch only the pointing finger; do not add an answer.
(At the point of difficulty) “What are you thinking right now?” Hear why they stopped: too many colors, uncertainty about whether progress will save, or boredom.
(Immediately after first completion) “What did you just do?” See whether the body learned the first fun or merely followed instructions.
(At the end) “Do you think you will open it again tomorrow? If so, why?” Hear whether a reason to return was planted in the first session.

How to fill it in.

First narrow the leakage quantitatively to one place. The numbers must say “this one place” before you know where to bring people for observation.
Hand that place to qualitative work. Logs answer where; observation and interviews answer why. Numbers first, people second.
Change only one field at a time. Choose the steepest interval, change only that, then measure again.
Measure re-FTUE separately too. Do not look only at newcomers’ D1; compare how many existing users return before and after a redesign.
Five people cheaply reveal large problems that catch almost everyone. Quantitative evidence fills in small differences and rare paths.

Short example.

Quantitative: first-completion rate 30%. When laid out by stage, the steepest drop is on the color-selection screen (first screen 100 → enter color selection 70 → complete choice 40 → character complete 30 → next day 8). Because these stage boundaries differ from the example funnel in Chapter 25, the fields and numbers differ too; use the comprehensive sample in Section 6 as the authoritative values.
Qualitative: three of five first-time users stopped on the color screen. “There are too many colors.” / “Will this be saved?” / one grew bored and closed it. The point where people abandon a choice midway is the candidate for revision.
Iteration: compare completion rates between a screen with fewer color options and the unchanged screen. Record the pre-change 30% as the baseline. Change only the color screen at one time.

3. Comprehensive Self-Review

Purpose. This sieve gathers in one place the design principles spread across Chapters 1 through 27. Once Design Document v1 is filled in, place a single first screen in front of you and check it line by line. Add no new principle; verify that none already covered was missed. Do not try to pass everything at once. Mark the lines that fail, then begin with the one closest to the first fun.

Borrowing familiarity (three familiar, one unfamiliar)

Has the new thing introduced by the first screen been narrowed to one?
Are the other three—among controls, screen composition, rewards, and worldview—left in already familiar forms?
Is there a place where two or more novelties overlap and merely feel difficult?
Does what the hand does come from the person’s former home, while the meaning made by that gesture arises only here?
For every familiarity borrowed, can we keep the expectation that its form brings?

Four principles (intuitive, familiar, consistent, low entry)

Intuitive: Before hearing an explanation, can the user guess what this screen is for?
Familiar: Does the screen’s action resemble something already done elsewhere?
Consistent: Does a rule that worked on one screen produce the same result on the next?
Low entry: Are the thresholds before the first action—login, terms, long explanations—low?
Scoring: Score intuitive, familiar, and consistent from 0 to 5, but let an extreme beginner rather than the team judge. Measure low entry through costs rather than a score, using the first-five-minute cost budget in Summary A, Tool 13. Record beside it the time to first input in seconds and notes on the difficulty of the first three rounds.
Example: intuitiveness 4; even a light user with no game experience reached “I should touch it,” and time to first input was 12 seconds (hypothetical).

Three layers of feedback

First layer (physical): Does the touched point respond immediately? Does an answer saying “your hand touched here” return within a brief moment?
Second layer (rules): Is it visible how that input was processed by the game’s rules?
Third layer (world): Does the world answer what that action meant inside it?
Do the three layers arrive in order: touched → reflected → meaningful? Is there a place where only the third is grand while the first is absent?
Is there a place where “the input registered” and “meaning arose” are mixed into one burst of fireworks, obscuring causality?
Sheet: choose five actions from the first five minutes. For each, record whether all three answers arrive—touched, reflected, meaningful—and mark any missing layer.
Example: when the character is stroked, the touched point depresses and sounds (physical), affection rises by one unit (rules), and the character giggles and follows (world). No layer is missing.

Six principles of error and recovery

Have you avoided reducing the possibility of failure to zero, since it supplies tension and joy? Is the failure nonetheless nonfatal?
Recoverability: Is there a way back from the point of error—undo or retry?
Control: Can the user understand what is happening and stop or reverse it?
Information preservation: Is what the user invested—a decorated character, typed name, selected settings—protected from disappearing entirely after a small mistake?
Do not burden users with what they cannot fix: Are problems outside their control, such as server delay, kept from being blamed on them?
Flow preservation: Does a large window interrupt the action or send the user back to the beginning?
Where something smells of “game over” or “start again,” is there a sign that turns death from punishment into rhythm?
Ladder: choose a point of risk in the first five minutes and write three levels for handling the same mistake—bad, better, best—then mark the level of the current design.
Example: for a wrongly colored area, bad handling opens a warning and returns to the beginning; better handling offers an Undo button; best handling lets the character smile even at the wrong color and returns to the previous state in one action.

Distracted-attention design

Have you assumed that the user is distracted and in noise, rather than quietly focused?
What to press now: Can a newcomer see at a glance what to press?
State: Can they see at a glance what is happening—progress, waiting, or completion?
Error: Is an error signal visible at a glance, like a red warning lamp rather than a message that must be read?
Have you crossed out things at the screen edge that a newcomer cannot read or does not need now?

Convention teardown

Has every convention left on the first screen passed the four questions—premise, shared understanding, cost, replacement—in the convention teardown review, Summary A Tool 3? The home and empty table for those four questions are in Summary A.

One protagonist per screen (Eight Layers)

For every first-screen element, did you write one letter naming the layer among the Eight Layers to which it speaks?
Are four or more layers mixed on one screen, making it speak clearly to no one?
Did you choose one priority layer? Are you giving the operator what belongs to the viewer?
For that one layer, did you separate what to turn on now from what to defer to the next screen?

Put payment after trust

Draw a line where the first fun arrives. Have you moved the expensive costs attached before it—login, personal information, spending decisions—behind the line?
Does the first impression “look free”? Is a visible advance charge blocking entry?
Is payment treated as an output metric rather than the endpoint of the first experience? The ladder of first moments ends at T10.
Is the standard for showing, hiding, or previewing payment on day one not “How much can we extract today?” but “Where did this person feel value?”

Whole-book reconciliation (finish)

First-moment mapping (T0–T10): Did you identify the single threshold with the greatest leakage?
Gauges and instrument panel: Are your hands on inputs—the gauges—and your eyes on outputs—the panel? Are you wasting effort trying to pull a result number as if it were a handle?
Measurement and iteration: Did you choose the single steepest interval? Do you know why people leave, or only suspect?
Re-FTUE: After a major update or season transition, is there one sentence pointing out only “what changed” where existing users will become lost again?
Next action: Did you choose just one place on the first screen to address first, rather than revising several at once?
Conflict coordination: When two or more layers—world, system, sense, social, business—occupy one point and compete, did you select the protagonist there and defer, blend, or remove the rest?
Example: Immediately after first completion, a completion cheer and sign-up request appear together, making the sense layer compete with the business layer. The protagonist is sense—the joy of completion—so defer sign-up until the user attempts to save.

4. AI Application Module (Isolated)

Purpose. Open this module only when bringing AI into the first experience. Chapters 1 through 25 of the main text did not insert AI; Chapter 27 is the entrance, and the deeper tools are gathered here. There are three reasons for keeping them separate. First, the core machinery of this book—the Eight Layers, T0–T10, experience costs, experience simulation, gauges and instrument panels, and prior analogous experience—works unchanged with or without AI. AI is not a principle but a tool for implementing those principles. Second, concrete details of AI age quickly; scattering them through the main text shortens the book’s life. Third, keeping them together makes updates easy.

Update premise. This section will age faster than any other in the book. The patterns and risks here retain only principles; names of particular people and specific figures for particular products are deliberately omitted. When tools change, rewrite only this section. The main text can remain untouched. What does not change is the trust fence below and the application of principles.

Spectrum of AI involvement (manual → reactive → recommend → automatic → generative). The farther right you go, the more the machine decides and the farther human control recedes. Begin by identifying the current level of your first screen.

Manual — decides nothing; everyone receives the same screen. People choose everything. Risk: unable to read the newcomer’s blank page, leaving the first screen vague.
Reactive — answers a human action on the spot. The person leads and the machine responds. Almost no risk; this is the domain of feedback.
Recommend — reads the person’s disposition and places candidates in front. The person chooses from a narrowed view. Risk: poor narrowing exposes only irrelevant candidates.
Automatic — guesses at the person and lays out a tailored screen in advance. The person receives it, then turns it off or corrects it. Risk: a poor fit treats the person as someone else.
Generative — creates characters, dialogue, and scenes on the spot. The person receives them. Risks: falsehood, broken consistency, and difficult reversal.
Record where each level already appears on your first screen or where you intend to place it. The result becomes a decision table for choosing when and where to use selection, recommendation, and automation.
Judgment rule: Each higher level increases the sweetness of a good fit and the disappointment of a mismatch. Before moving up one level, ask: What happens if the machine misreads a person here? Can we detect the mismatch? Does the user have an exit? If any answer is missing, build the fence before raising the level.

Major patterns for using AI in the first experience (pattern — action — connected principle — caution). Names and counts will change as tools evolve. Look only at the principles.

Chapter 27’s three branches are the broad shapes in which AI enters. The five levels and seven patterns here divide those shapes into a finer practical decision table. Personalization leads to recommendation and automation, cold-start assistance, and taste inference; conversational guidance becomes the conversational-guidance pattern; on-the-spot generation leads to the generative level and on-the-spot-generation pattern. Adaptive difficulty and operations/review assistance are cross-cutting patterns that support all three branches.

Personalized first screen — prearranges different characters, actions, and tones for each person — Eight Layers and prior analogous experience — a good fit writes a good first line on the blank page; a poor fit is worse than leaving it blank.
Conversational guidance — guides through questions and answers instead of a fixed tutorial — three feedback layers and experience costs — adapts to the person’s pace, but a wrong answer ruins first recognition.
On-the-spot generation — instantly creates appearance and dialogue — experience simulation — easily becomes a spectacular sign without an event; beware inconsistency and falsehood.
Cold-start assistance — fills an information-free first entry by inference from similar people — T0–T10 first entry and first recognition — beware skew that misses uncommon groups.
Taste inference — lays out what comes next from selections and dwell time — Chapter 26, choice as a taste signal — read the signal as a hypothesis rather than a conclusion; do not confine the person.
Adaptive difficulty — quietly adjusts in response to how the person gets stuck — Chapter 17, failure as rhythm and flow — change things without the user noticing, but not in a way that feels frustrating or depriving.
Operations/review assistance — automatically produces and filters variations of the first screen — measurement and iteration — screens differ by person and are difficult to compare.

Three pillars of the trust fence. Without these, the kindness of tailoring hardens into interference the user cannot control. They remain unchanged even when tools change.

Why is this shown? (transparency): The user should have at least a glimpse of where the received first screen came from and what led the system to form this impression.
Can it be turned off? (control): When tailoring is intrusive or wrong, there must be a handle that stops it and returns to the standard first screen.
Can errors be corrected? (correctability): If the machine misreads the user, saying “That is not who I am” must change the result.
Review: Is the reason visible? / Can it be turned off? / Can an error be corrected? If even one line is blank, build the fence before adding AI there.

Four risks.

Bias and skew: The system fits frequently observed groups well and rare groups badly. One person’s first moment is built carefully; another’s is thrown together. Examine separately which dispositions are especially poorly served.
Falsehood: A generative screen makes nonexistent things look plausible. Without an event behind the sign, greater spectacle is exposed more quickly.
Overdependence: Someone who continually receives only tailored experiences loses opportunities to encounter qualities they did not know. A first moment made for one person can narrow rather than broaden that person.
Privacy: To guess more accurately, the system gathers dwell time and even where fingers stop. The personal-information cost can grow on an AI-built screen.

Measurement, review, and fairness for a personalized first moment.

Measurement: You know whether an AI-tailored screen is truly better only by comparing it alongside a standard, untailored screen.
Review: Preserve a record that reconstructs which first screen each user received. Without a way to look back, you cannot even know what went wrong.
Fairness: Measure how many people find the tailoring intrusive enough to turn it off and which dispositions are especially poorly served.
Applying principles: The Eight Layers, costs, and convention checkpoint do not disappear because AI built the screen. First decide what to accomplish, which layer to address, which costs to reduce, and which conventions to challenge; then find where AI can implement those decisions better. Do not let the tool determine the goal.

5. Multidisciplinary Principle Index

Purpose. This index gathers by field the established principles woven into the main text. Rather than relying on one source, this book borrowed evenly from several fields for FTUE design. The main text explains only the concepts in the author’s own words and includes neither book titles nor direct quotations. Here, every principle is accompanied by the chapter where it was used and a source at the level of person or institution. This is a guide for deeper exploration, not a list of footnotes.

The principles here are not the subjects of the main chapters. Each serves only as a light illuminating one question in the original memo. If one source becomes the star of an entire chapter, the chapter is overloaded.

Software UI/UX (principle — meaning / chapters used / source)

Jakob’s Law — users spend more time in other apps and expect yours to work in ways they already know / Chapters 1 and 5 / Jakob Nielsen, Nielsen Norman Group.
Invisible design — good design goes unnoticed; affordances and signifiers / Chapters 1, 13, and 15 / Donald Norman. The terms signifier and signified themselves are generally traced to Saussurean semiotics.
Usability heuristics — visibility of system status, match with the real world, user control, consistency, error prevention, recognition over recall, and error recovery / Chapters 2, 15, 16, and 17 / Jakob Nielsen, Nielsen Norman Group.
Hick’s Law — more choices slow decisions / Chapters 8 and 11 / generally traced to research by Hick and Hyman.
Fitts’s Law — large, nearby targets are faster to press / Chapter 16 / psychologist Paul Fitts.
Progressive disclosure and empty states — do not show everything at once; open it when needed / Chapters 3 and 21 / common across the UX field, including Nielsen Norman Group.

Marketing, advertising, and promotion

Funnel from attention to action — attention → interest → desire → action / Chapters 18 and 23 / generally traced to a model by advertising pioneer E. St. Elmo Lewis.
Positioning — occupy one place amid competition / Chapters 1 and 18 / Al Ries and Jack Trout.
The three-second rule and hooks in advertising — capture people in the first few seconds / Chapter 18 / common practice in video advertising.
Free trials, demos, and starters — let people taste value first / Chapters 11 and 13 / common industry practice.
Primacy effect — information received first is disproportionately weighted / Chapters 11 and 18 / generally traced to Solomon Asch’s research on impression formation.
Expectation disconfirmation — a gap between preview and reality causes disappointment / Chapters 11 and 18 / generally traced to Richard Oliver’s expectation-disconfirmation model.

Branding

Brand consistency — the same experience wherever people encounter it / Chapter 12 / common branding practice.
Trust from first impressions — coherence and completeness become trust / Chapters 4 and 14 / generally traced to Stanford Web Credibility Research led by B. J. Fogg.

Business and product

Value proposition — the first screen answers “Why should I use this?” / Chapters 1 and 6 / common in business strategy and product practice.
Retention cohorts — how many remain after the first day, seven days, and one month / Chapters 3, 22, and 25 / common in growth analytics.
Activation and time to first value — the first completion of a core task (aha) and the time required to reach it / Chapters 3, 20, and 22 / generally traced to Dave McClure’s pirate-metrics framework.
Organic and paid acquisition — motivation for arrival determines expectations of the first experience / Chapter 23 / common in growth analytics.
Foot-in-the-door — a small agreement invites a larger one, though the first payment remains an output metric / Chapter 24 / generally traced to experiments by Freedman and Fraser.

Cognitive and behavioral psychology (general)

Curse of knowledge — a person who knows cannot see the bewilderment of one who does not / Chapter 1 / generally traced to Elizabeth Newton’s Stanford tapper-and-listener experiment.
Hedonic adaptation — repeated exposure wears down the same stimulus / Chapter 18 / generally traced to discussion by Brickman and Campbell.
Loss aversion — fear of losing what was invested inhibits action / Chapters 11 and 17 / Kahneman and Tversky.
Variable ratio — irregular numbers of actions before a reward produce stronger attraction / Chapter 13 / Skinner.
Flow — immersion arises when difficulty and skill match / Chapter 14 / Mihaly Csikszentmihalyi.
Motivation, ability, and trigger — behavior occurs when all three meet at one point / Chapter 15 / B. J. Fogg’s behavior model.
Baby schema and the uncanny valley — safety signals in a character seen for the first time / Chapters 10 and 14 / baby schema is traced to Konrad Lorenz, and the uncanny valley to Masahiro Mori.

Supporting references (ideas only). Some passages in this book borrow only ideas from other work by the same author: the sense that ordinary people compete over small fragments of spare time, the distinction between gamers and ordinary people, the metaphor of turning tourists into residents, and the concept of re-FTUE.

Only the ideas were borrowed. The main text does not directly expose book titles, proprietary terms from those works, or operations and monetization language; it internalizes them in the author’s own words.
This book has one primary source: the original memo at the beginning of Chapter 1. Supporting references remain references, nothing more.

6. Comprehensive Sample: Hypothetical Casual Game (Authoritative)

Purpose. This is one fully completed version of the Summary A master workbook, filled from beginning to end for a single hypothetical game. It is a casual game in which players decorate a cute animal character, name it, and care for it each day; its primary users are ordinary people who rarely play games. These samples S0–S15 are authoritative for examples scattered across the three summaries. The short examples in Summary A are excerpts from this sample. If numbers—such as the hypothetical funnel of first completion 30, thirty-minute retention 25, and D1 8—or sense vocabulary appear inconsistent, follow this sample. It demonstrates how to fill the blanks; it is not an answer key.

S0. FTUE scope statement. Start: the moment a user sees one social-media image saying, “You can make a cute character like this yourself.” End: the moment they first complete, name, and save their own character. In one sentence: It begins with the first image of a cute character and ends when I first complete and save that character with my own hands.

S1. Familiarity inventory and conventions to dismantle. Familiarity to borrow: swipe up (short-form video), select and apply colors (coloring), stroke with a hand (doll play), and a card-like illustrated catalog (collection apps). What to break: do not put combat, competition, or failure on the first screen. Convention to dismantle, in one line: the cultural assumption that “games are challenge and overcoming.” Remove difficulty, failure, and game over from the first screen; replace them with warmth that responds to touch.

S2. First-moment mapping T0–T10. T0 word of mouth: “Apparently you can make characters like this yourself” / T1 one character image in a social-media timeline / T2 “Can I make one too?” / T3 install and launch, then start making without login (★ placing login here creates the greatest leak) / T4 “I am making a cute character” / T5 choose a color and stroke / T6 it giggles when touched / T7 an accidental press nearly erases the decoration / T8 my own character now exists (aha) / T9 name it and show it to someone / T10 return the next day to see how it is doing. Identified leak: T3 when login is placed there. Let people create first; request sign-up when they save.

S3. Borrowing prior analogous experiences. Short-form video (immediacy, swiping, free access → swipe up; risk: a long load or intro invites a swipe away) / coloring (what I color becomes the result → choose and apply colors; risk: too many colors become overwhelming) / doll play (touch produces an attached response → stroke; risk: no response makes a dead doll) / character-fan activity (favorite character, belonging, daily companionship → catalog, decoration, checking in; risk: calling a character by a classification label feels wrong).

S4. Eight user layers and priority. Recipient of each first-screen element: large character image (Viewer), color and stroke buttons (Button Presser), “Create My Character” (Customer), naming (Character). Priority layer: L3 Viewer. A newcomer came to see “This character is so cute.” Defer: turn off operations, settings, and full-catalog overview (L6) on the first screen.

S5. Experiencer and mistaken assumptions. Primary target: an ordinary person who rarely plays games and may have little game experience. Unknowns: game controls and genre abbreviations. First hurdle: does not know what to press first. Taboo: calling the character a “collectible” or sending it into combat. Mistaken assumption: gamers think “tutorials are for skipping,” but this user does not expect a tutorial at all. Let them meet immediately instead of guiding them.

S6. Migration candidates and worldview map. Three candidates: (a) a person who collects all kinds of cute-character content, (b) a light user who likes cute animal characters, and (c) a person who watched and now wants to try making one. Primary worldview in one sentence: a warm world where cute characters wait in their own places to be cared for and raised every day. Two adjacent worlds: character/decoration communities and pet-raising/decoration games. Worldview information prohibited on the first screen: proper nouns that must be memorized, long lore cutscenes, and world history that is not yet needed.

S7. Character-sense roster. (first-screen character — sense — content — origin that will like it — sentence translated into sense)

Puppy — protection and care — stroking and decoration — people who like warm characters — “A soft child you want to hold and keep looking after.”
Frog — play and response — tapping and rhythm — people who like lively characters — “A playful child that hops in answer to a tap.”
Cat — exploration and discovery — looking around and finding things — people who like cool charm — “A child who does not approach first but stays a long time once it lets you close.”
Select senses only from the fixed list of eight clusters. Put the puppy at the very front of the first screen because protection and care are the senses to which a newcomer most easily reaches out.

S8. Experience-cost budget. Time immediately after entry (loading) → show a cute character image while loading / attention when creation begins (colors and accessories) → one at a time, with defaults already set / personal information at save attempt (sign-up) → ★ defer here, after creation is finished / no cost after first completion → line where the first fun arrives. Draw the line: place the first fun—character completion—before sign-up.

S9. Three content-experience types. Among the ten experiences, the primary user is accustomed to ownership, relationship, and decoration (sensation). Three to give in the first thirty minutes: ownership (my character now exists), sensation (the feel of touching and coloring), relationship (the character recognizes me). Defer: daily care, growth, and placing several characters together; display and discovery come after the user stays.

S10. Experience-simulation analysis. Heart button (points to attachment / a heart alone does not produce attachment / create it through naming and stroking) / completion celebration (joy of ownership / the real event of my character coming into existence is present / match the scale of the presentation to the event) / catalog-fill indicator (reward of collection / the event is too small on the first screen / reduce the sign and defer it until after the user stays).

S11. One hundred first impressions. Experience promised by the preview: “You make a cute character like this yourself.” Experience returned by the first screen: complete a cute character with my own hands. Are the two sentences the same? Yes. The preview does not promise combat or competition, so the first screen keeps its promise. Three meeting points: social-media image, first store image, and screen immediately after launch. Each must reveal at a glance what this is for.

S12. Time-based design. First 30 seconds (give identity and safety / defer long explanation and sign-up) / up to 3 minutes (first input and response / defer the whole catalog) / up to 10 minutes (first achievement = character completion / defer operations and settings) / up to 30 minutes (big picture = several characters, things to decorate, and choice / defer payment) / one day later (reason to return = check on and care for the character each day).

S13. Instrument panel and causal map. Five result numbers: first-completion rate, sign-up conversion, next-day return, first share, and session length. North Star: next-day return; revenue and daily active users are too far away, so exclude them. Input → arrow: put creation before sign-up → first completion↑ / plant something to check on immediately after completion → return↑ / immediate response to touch → first completion↑ / share button at completion → first share↑ / set color defaults → attention cost↓.

S14. Measurement, iteration, and re-FTUE. Retention by stage: entry → start creation → first completion → sign-up → return. Mark the steepest interval. If why people leave is still only a guess, sit beside five first-time users and watch. One re-FTUE sentence: when a season adding a new character opens, do not teach existing users everything; point out only “The new character is here.”

S15. Choice entrusted to the user and AI fence. One entrusted choice: colors, accessories, and expression on the first screen. This choice makes the user a creator and becomes the first taste signal. Read the signal as a hypothesis rather than a conclusion and always leave a way to choose again. If AI is involved, it can read selected colors and gestures and instantly create a character suited to that taste. A good reading creates a one-of-a-kind character; a poor reading makes the user turn away. Write three lines: Can the user understand why they received this character? / Can they turn tailoring off? / Can they correct it when wrong? If even one line is blank, build the fence first.

7. FTUE Risk List by Genre

Purpose. After choosing a genre, review in advance the risks common to that genre’s first five minutes and the costs ordinary people pay there. Gamers pass these risks through context; ordinary people pay the cost at the same point and leave. A “common risk” is something the genre traditionally places on the first screen; “cost paid” is what an ordinary person loses there. The instruction is not to eliminate the risk, but to reduce it on the first screen and defer it until after people stay.

RPG

Common risks: long opening cutscenes and spoon-fed lore, a flood of proper nouns to memorize, yellow-exclamation-point quest markers, and skill trees or stat screens on the first screen.
Costs paid: time (waiting for cutscenes), attention (memorizing proper nouns), lack of context.
First-five-minute remedy: introduce the world’s atmosphere with one line, and give the first input before the cutscene. Defer lore until after the user stays.

Puzzle

Common risks: lengthy written rules before play, high difficulty in the first round, and immediate exposure to three-star and score pressure.
Costs paid: attention (must read before starting), pride (blocked in the first round), social cost (pressure from score comparisons).
First-five-minute remedy: let the user make one move instead of reading an explanation, and make the first round necessarily solvable. Show scores and stars only after the first success.

Simulation

Common risks: a first screen crowded with menus and numbers, freedom so broad that it overwhelms, and long resource or construction tutorials.
Costs paid: attention (information overload), time (tutorial), lack of context (no goal).
First-five-minute remedy: give one clear first goal and turn off the other menus. Open freedom after one thing has been accomplished.

Collection

Common risks: gacha and probability displays on the first screen, an empty catalog, and simultaneous explanations of soft and hard currencies.
Costs paid: spending anxiety (pressure to draw on the first screen), aversion (complex probability and currencies), social cost (deprivation from an empty catalog).
First-five-minute remedy: simply give one first character before any draw. Defer currency explanations and show the catalog with at least one field already filled.

Idle

Common risks: no visible indication of what is progressing, a first screen filled only with numbers, and no reason to return planted on day one.
Costs paid: attention (interpreting numbers), lack of meaning, no motivation to return.
First-five-minute remedy: make what is growing visible at a glance and plant a day-one excuse to return: “Come back shortly and this will be ready.”

Social

Common risks: sign-up, friend linking, and permission requests on the first screen; immediate ranking and comparison; an empty feed.
Costs paid: personal information (first-screen sign-up and permissions), social cost (early ranking and sharing pressure), lack of meaning (empty feed).
First-five-minute remedy: allow looking around and private use first; defer sign-up and linking until after value is felt. Open rankings late enough that a beginner is not intimidated.

Interactive story / visual novel

Common risks: dumping long prose and character introductions in the first episode, failing to show whether choices change the story, showing choices locked behind currency in the first episode, and explaining only gamer-oriented Skip and Auto buttons.
Costs paid: time (long opening), attention (memorizing people and setting), spending anxiety (locked choices), lack of meaning (choices go nowhere).
First-five-minute remedy: keep the first episode short, offer the first choice early, and immediately show its effect on the screen. Remove locked choices and currency signs from the first episode; introduce them only after attachment forms.

Rhythm / variety casual

Common risks: demanding exact rhythm in the first round so failure comes first, exposing terms such as combo and judgment grades without explanation, immediate scores and rankings, and ads between every short round.
Costs paid: failure (deprivation in the first round), attention (interpreting judgment terms), social cost (score comparison), time (ads between rounds).
First-five-minute remedy: make first-round judgment generous enough for everyone to reach the end, then introduce judgment terms one at a time after the first success. Defer ads and rankings until after the first fun arrives.

8. Glossary

Purpose. A concordance gathering the book’s distinctive terms and abbreviations in one place. It corresponds one-to-one with terminology in the main text. Do not invent new terms.

Abbreviations and time axis

First-Time User Experience (FTUE) — from first exposure to first core completion; in the working definition, T1 through T8.
New User Experience (NUX) — a newcomer’s settlement, return, and habit formation; T3 through T10.
Out-of-Box Experience (OOBE) — installation, opening, and setup; beginning without obstruction. A subset of T3.
User Experience (UX) — the entire lifecycle: meaning, satisfaction, relationships, and continuation. From T0 without an endpoint.
Onboarding — moving someone into the world and its habits; a means that continues FTUE.
Tutorial — one form for teaching controls and rules. It can belong within FTUE but is not FTUE itself.
The first eleven thresholds (T0–T10) — eleven chronological stages from First Rumor (T0) to First Return (T10). The single source of truth in this book.
Three phases — three groups used to speak quickly about T0–T10: before the game / first session / settlement.
Aha — the moment when the first core task is completed and the user thinks, “So that’s what this is.” T8.

People axis

Eight user layers — eight identities one user holds simultaneously; deeper involvement from bottom to top.
Customer (L1): asks whether it is worth money and time / User (L2): low friction and utility / Viewer (L3): something to see and immediate comprehension / Button Presser (L4): immediate response and confidence / Player (L5): challenge, mastery, and control / Console/Operator (L6): overview and sense of operation / Character (L7): a personal story inside the world / Resident (L8): a world and relationships in which to stay.
Eight experiencer types — eight branches within the group called ordinary people, separated by what they do not know: genre, franchise, platform, worldview, economy, social, returning, and spectator. A people axis orthogonal to the time axis.
Ordinary people’s fragments of spare time — the majority who did not come to play a game but to fill a short opening in the day. The primary target of this book.
From tourist to resident — turning someone who came to look into someone who stays; the metaphor for first-screen design.
Migration candidate — someone brought from another world into ours; a game-specific extension of the persona.

Physics of experience

Five layers of experience — experience is not flat; its layers are world, system, sense, social, and business.
Experience costs — users do not consume experience for free: attention, time, opportunity cost, social cost, failure, personal information, and spending anxiety.
Experience simulation — a sign that imitates an experience. Without an event behind it, it is not an experience.
Signifier — a sign pointing to an experience, such as a heart button / signified — the original experience it points to, such as attachment or recognition.
Prior analogous experience — an experience already familiar from another medium; both an asset that lowers entry cost and a trap.
New-content relativity — games are not an isolated medium. Users arrive after passing through video, social media, webtoons, fandom, and commerce.
Defamiliarization — stopping familiar terms and asking about them again as if seen for the first time; the methodology of this book.
Three feedback layers — three layers in which the world answers an action: physical (touched), rules (reflected), world (meaning).
Four principles — four qualities that make the first screen readable: intuitive, familiar, consistent, and low entry.
Three familiar, one unfamiliar — make only one thing introduced by the first screen new; keep the rest already known.
Flow — immersion that arises when difficulty and skill match.

Instrument-panel axis

Gauge — an input directly operated by the designer, my decision / instrument panel — output metrics appearing as results, the accumulated result of other people’s decisions.
Preview of experience — expectations planted before the first screen is opened. A preview is a promise.
One hundred first impressions — first exposures created by combining channel, layer, and threshold; two axes of selection and endurance.
Selection — the strategy of being chosen through a first impression / endurance — a first impression that does not wear down under repeated exposure.
North Star — the single metric that most directly reveals whether the first experience did its job.
Re-FTUE — moments after a major update, return, or season transition when an existing user becomes a beginner again.
Activation — the state of having reached the first completion of the core task; connected to T8.

Identity terms of this book

New content — game-like content that borrows game grammar while targeting people who are not gamers.
Convention teardown — examining conventions inherited by games with four questions: premise, shared understanding, cost, and replacement. The engine of this book.
Trust fence — three pillars that keep an AI-built first experience from becoming interference: transparency, control, and correctability.
First-time novice designer — a state, not a job title: someone who has seen so much that they pass basic principles as obvious.
My FTUE Design Document v1 — one sheet gathering fields completed from Chapter 1 through Summary A. It is a beginning, not a completion, and grows into v2 and v3.

For a broader chapter-by-chapter review, continue to “Summary C: Reference Cases by Chapter,” a collection that expands each chapter’s argument through real examples from inside and outside games.