30 Claude Prompts That Build Product Dashboards

You are a retention analyst who builds cohort heatmaps a product team can read in a Monday review. <context> I need a cohort retention dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic decaying cohorts so I can click the working screen in the chat. </context> <inputs> - Product name: [PRODUCT] - Cohort grouping: [E.G. SIGNUP WEEK] - Retention window: [E.G. WEEKS 0-8] - Metric: [E.G. % STILL ACTIVE OR % STILL PAYING] - Number of cohorts: [E.G. LAST 8] </inputs> <task> Build the dashboard with: a header for [PRODUCT] and the metric; KPI cards (newest cohort size, week-1 rate, week-4 rate, estimated steady cell); a shaded heatmap (rows are cohorts, columns are periods, cells show percent); a line chart overlaying cohort curves; and a newest-versus-oldest callout. Generate decaying-but-plausible percents in labeled JS constants. Cell click lists that cohort's period value. Color plus the number so the map is not color-only. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Cohort matrix in labeled JS constants at the top. - Newest row first. Accessible contrast. - Do not invent revenue if I asked for active retention. Label the metric on the page. </constraints> <format> Return the full HTML as an artifact, then explain how to paste a real cohort matrix over the constants. </format>

You are a retention analyst who treats D1, D7, and D30 as three separate clocks. <context> I need a D1 D7 D30 retention dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic returning-user counts so I can click the working screen in the chat. </context> <inputs> - Windows to show: [RETENTION WINDOWS] - Product name: [PRODUCT] - Qualifying start event: [E.G. SIGNUP OR FIRST SESSION] - Date range of start days: [E.G. LAST 45 DAYS] - Targets per window: [E.G. D1 30%, D7 18%, D30 12%] </inputs> <task> Build the dashboard with: KPI cards for each window versus its target; a triple line of D1, D7, and D30 over start days; a table (start day, starts, returned D1, D7, D30, rates); and a falling-window callout that names which clock dropped most versus the prior four starts. Compute each rate as returned in that window over starts that have aged enough. Starts too new for D30 show as pending, not zero. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Start-day rows in labeled JS constants at the top. - Pending cells labeled pending. Accessible contrast. - Do not invent platform benchmarks I did not give. </constraints> <format> Return the full HTML as an artifact, then write the aged-enough rule and which array to replace. </format>

You are a retention analyst who builds churn-risk watchlists from usage and silence, not a vibe score. <context> I need a churn-risk dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic accounts so I can click the working screen in the chat. </context> <inputs> - Signals to score: [CHURN SIGNALS] - Product name: [PRODUCT] - Weights per signal: [E.G. USAGE 40, LOGIN GAP 30, TICKETS 20, NPS 10] - At-risk threshold: [E.G. SCORE OVER 60] - Accounts to list: [E.G. 20 SAMPLE ACCOUNTS] </inputs> <task> Build the dashboard with: KPI cards (accounts scored, at risk, median score, newly silent); a score histogram; a sortable watchlist (account, plan, last active, each signal, computed score, status); and a signal-weight note on the page. Compute the score in JS from my weights and normalized signal values. Filter chips: at risk / all. Selecting an account shows the signal breakdown. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Accounts, signals, and weights in labeled JS constants at the top. - Show the formula. Accessible contrast, not color-only. - Do not invent cancel dates or save-offer copy. Risk score only. </constraints> <format> Return the full HTML as an artifact, then write the score formula and which arrays to replace. </format>

You are a product analyst who builds stickiness dashboards from DAU, WAU, and MAU, not a vanity active total. <context> I need a stickiness dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic daily actives so I can click the working screen in the chat. </context> <inputs> - Stickiness target: [STICKINESS TARGET] - Product name: [PRODUCT] - Active definition: [E.G. COMPLETED CORE EVENT] - Date range: [E.G. LAST 90 DAYS] - Optional split: [E.G. PLATFORM OR PLAN] </inputs> <task> Build the dashboard with: KPI cards (latest DAU, WAU, MAU, DAU/MAU versus [STICKINESS TARGET]); a dual line of DAU and MAU; a stickiness line (DAU/MAU) with the target; a WAU/MAU secondary if space; and a table of weeks with those four numbers. Compute WAU and MAU from the daily active series using rolling 7-day and 28-day unique counts in the sample, or from pre-aggregated columns if I supply those. Label which method you used. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Daily series in labeled JS constants at the top. - Below-target stickiness flagged. Accessible contrast. - Do not invent a second activity definition. One rule on the page. </constraints> <format> Return the full HTML as an artifact, then explain DAU/WAU/MAU math and which series to replace. </format>

You are a product-finance analyst who builds revenue-retention boards a founder can defend. <context> I need a GRR and NRR dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic MRR movements so I can click the working screen in the chat. </context> <inputs> - Plan or segment to lead with: [PLAN OR SEGMENT] - Other segments: [LIST OR "ALL"] - Months: [E.G. LAST 12] - Currency: [E.G. USD] - Starting MRR by segment: [OR "INVENT PLAUSIBLE"] </inputs> <task> Build the dashboard with: a segment filter defaulting to [PLAN OR SEGMENT]; KPI cards (opening MRR, GRR, NRR, expansion, contraction, churned MRR); a waterfall or stacked bar of movements by month; a GRR versus NRR line; and a table (month, open, new, expand, contract, churn, close, GRR, NRR). Compute GRR as (open + expand - contract - churn) / open, excluding new logos from the numerator. Compute NRR as (open + expand - contract - churn) / open, still excluding new. Show both formulas on the page. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Monthly movements in labeled JS constants at the top. - Currency labeled. Accessible contrast. - Do not invent tax or FX. MRR math only. </constraints> <format> Return the full HTML as an artifact, then write GRR and NRR in one paragraph and which array to replace. </format>

You are a product conversion analyst who builds signup-to-activated funnels from product events. <context> I need a signup-to-activated funnel dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic step counts so I can click the working screen in the chat. </context> <inputs> - Funnel steps in order: [FUNNEL STEPS] - Product name: [PRODUCT] - Period: [E.G. LAST 8 WEEKS] - Activation definition: [ONE LINE] - Optional weekly series: [YES / INVENT] </inputs> <task> Build the dashboard with: a header for [PRODUCT]; a horizontal funnel using [FUNNEL STEPS] with count and step-to-step rate; KPI cards (started, activated, overall rate, biggest leak); a weekly overall-rate line; a step table (count, from previous, from top); and a leak callout. Compute every rate from the counts. Flag the largest drop. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Step counts in labeled JS constants at the top. - Largest drop visually flagged. Accessible contrast. - Do not invent marketing spend. Product events only. </constraints> <format> Return the full HTML as an artifact, then point out which drop to investigate first and how to paste real step counts. </format>

You are a product analyst who builds feature-adoption funnels that end on repeat use, not a first click. <context> I need a feature adoption funnel as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic feature events so I can click the working screen in the chat. </context> <inputs> - Feature name: [FEATURE NAME] - Steps in order: [E.G. EXPOSED, CLICKED, COMPLETED, USED AGAIN IN 7 DAYS] - Eligible users: [E.G. ACCOUNTS ON A PLAN THAT HAS THE FEATURE] - Period: [E.G. LAST 30 DAYS] - Repeat rule: [E.G. 2+ USES IN 7 DAYS] </inputs> <task> Build the dashboard with: a header for [FEATURE NAME]; KPI cards (exposed, completed, repeat users, repeat rate); a funnel of the steps; a bar of drop at each step; a weekly exposure-to-repeat line; and a table of eligible users still stuck before repeat. Compute rates from counts. Show the repeat rule on the page. Clicking a step lists sample stuck users. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Event counts and a small stuck list in labeled JS constants. - Repeat is not the same as first click. Accessible contrast. - Do not invent a second feature. One feature only. </constraints> <format> Return the full HTML as an artifact, then explain the repeat rule and which constants to replace. </format>

You are a product-led growth analyst who builds trial-to-paid funnels a founder can run before the trial clock ends. <context> I need a trial-to-paid dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic trial rows so I can click the working screen in the chat. </context> <inputs> - Trial length: [TRIAL LENGTH] - Product name: [PRODUCT] - Steps: [E.G. TRIAL START, ACTIVATED, HIT PAYWALL, PAID] - Activation event: [E.G. FIRST VALUE] - Period: [E.G. TRIALS STARTED LAST 45 DAYS] </inputs> <task> Build the dashboard with: KPI cards (trials, activated, paid, convert rate, activated-but-unpaid); a funnel of the steps; a convert-by-start-week line; a days-left risk list of open trials (id, day of trial, activated yes/no, hours since last use); and a table of ended trials with convert or expire. Compute remaining days from start date, [TRIAL LENGTH], and a now constant. Filter: open / converted / expired. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Trial rows in labeled JS constants at the top. - Unactivated open trials flagged. Accessible contrast. - Do not invent discount copy or card-on-file facts I did not give. </constraints> <format> Return the full HTML as an artifact, then explain remaining-day math and which array to replace. </format>

You are a product monetization analyst who builds upgrade funnels from in-product plan changes. <context> I need a plan-upgrade funnel dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic plan-change events so I can click the working screen in the chat. </context> <inputs> - Plan tiers in order: [PLAN TIERS] - Intent start: [E.G. HIT LIMIT / OPENED BILLING / CLICKED UPGRADE] - Later steps: [E.G. VIEWED PLANS, STARTED CHECKOUT, PAID] - Period: [E.G. LAST 60 DAYS] - Product name: [PRODUCT] </inputs> <task> Build the dashboard with: a header for [PRODUCT]; a funnel from intent through paid; KPI cards (intents, upgrades, upgrade rate, stalled in checkout); a from-tier to-tier matrix using [PLAN TIERS]; a stalled-workspace list (workspace, current plan, last step, days since intent); and a weekly upgrade-rate line. Compute rates from event counts. A from-tier filter updates the funnel. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Events and workspaces in labeled JS constants at the top. - Stalled means intent without paid inside an editable day window. Accessible contrast. - Do not invent list prices I did not give. Mark missing prices as [VERIFY]. </constraints> <format> Return the full HTML as an artifact, then explain stall math and which arrays to replace. </format>

You are a product conversion analyst who compares the same in-product funnel across two segments. <context> I need a two-segment product funnel comparison as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with two realistic series so I can click the working screen in the chat. </context> <inputs> - Funnel steps in order: [4-7 STEPS] - Segment A: [SEGMENT A] - Segment B: [SEGMENT B] - Period: [E.G. THIS QUARTER] - Product name: [PRODUCT] </inputs> <task> Build the dashboard with: a header and a control that focuses [SEGMENT A], [SEGMENT B], or both; side-by-side funnels or a grouped bar per step; a comparison table with counts, step-to-step rates, and a delta column; a line of overall conversion for each segment over time; and a callout of the step where the two segments diverge most. Generate sample data for both segments in labeled JS constants and compute rates and deltas. The control must update the focused view without a reload. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Both segment series in labeled JS constants at the top. - Largest gap flagged. Accessible contrast. - Do not invent a third segment. </constraints> <format> Return the full HTML as an artifact, then explain the delta math and where to paste both segment counts. </format>

You are a product analyst who ranks features by reach and frequency, not raw event volume. <context> I need a feature-adoption scorecard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic feature totals so I can click the working screen in the chat. </context> <inputs> - Features to score: [FEATURE LIST] - Product name: [PRODUCT] - Period: [E.G. LAST 30 DAYS] - Eligible users: [COUNT OR "INVENT"] - Adoption target: [E.G. 25% REACH] </inputs> <task> Build the dashboard with: a header for [PRODUCT]; KPI cards (features tracked, features at target, median reach, median weekly frequency); a horizontal bar of reach by feature; a scatter or grouped view of reach versus frequency; a scorecard table (feature, unique users, reach %, events, events per user, versus target); and a laggard callout. Compute reach as unique users over eligible. Clicking a feature shows a tiny weekly sparkline from sample series. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Feature rows in labeled JS constants at the top. - Below-target features flagged. Sortable table. Accessible contrast. - Do not invent revenue per feature. </constraints> <format> Return the full HTML as an artifact, then write the reach formula and which array to replace. </format>

You are a product analyst who reads feature health as frequency, not a single adoption percent. <context> I need a feature-frequency dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic per-user event counts so I can click the working screen in the chat. </context> <inputs> - Feature name: [FEATURE NAME] - Period: [E.G. LAST 4 WEEKS] - Frequency buckets: [E.G. 1, 2-4, 5-10, 11+] - Power-user rule: [E.G. 11+ IN THE PERIOD] - Product name: [PRODUCT] </inputs> <task> Build the dashboard with: a header for [FEATURE NAME]; KPI cards (users, median uses, power users, one-and-done share); a histogram of uses per user using the buckets; a weekly repeat line (users with 2+ that week); a power-user table; and a one-and-done list (id, first use date, no second use). Compute buckets from a per-user count array. Bucket chips filter the lists. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Per-user counts in labeled JS constants at the top. - One-and-done is exactly one use in the period. Accessible contrast. - Do not invent session length. </constraints> <format> Return the full HTML as an artifact, then explain bucket rules and which array to replace with an event log rollup. </format>

You are a product analyst who splits the base into usage tiers a roadmap conversation can share. <context> I need a power-versus-casual dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic user tiers so I can click the working screen in the chat. </context> <inputs> - Usage tiers and rules: [USAGE TIERS] - Product name: [PRODUCT] - Period: [E.G. LAST 28 DAYS] - Features to compare across tiers: [3-6 FEATURES] - Migration window: [E.G. PRIOR 28 DAYS] </inputs> <task> Build the dashboard with: KPI cards (users per tier, share of events from power, casual share); a stacked share bar over weeks; a feature-reach table by tier; a migration matrix (prior tier to current tier); and a migration list of users who dropped a tier. Compute tiers in JS from a usage score or event count using my rules. Clicking a tier filters the feature table and the list. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Users and rules in labeled JS constants at the top. - Show the tier rules on the page. Accessible contrast. - Do not rank people as good or bad. Counts only. </constraints> <format> Return the full HTML as an artifact, then write the tier rules in comments and which array to replace. </format>

You are a product analyst who measures whether a feature still gets used after the first week, not whether it was clicked once. <context> I need a feature-retention dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic first-use cohorts so I can click the working screen in the chat. </context> <inputs> - Feature name: [FEATURE NAME] - First-use cohort grain: [E.G. WEEK OF FIRST USE] - Retention windows: [E.G. WEEK 1, WEEK 2, WEEK 4] - Used again means: [E.G. 1+ EVENT IN THAT WEEK] - Product name: [PRODUCT] </inputs> <task> Build the dashboard with: a header for [FEATURE NAME]; KPI cards (first-time users, week-2 retained, week-4 retained, fade from week 1 to week 2); a cohort heatmap of feature retention; a curve overlay; and a faded-user list from the newest cohort (id, first use, last use, weeks since last). Compute retained as users with a use in that window after first use. Show the used-again rule on the page. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - First-use rows in labeled JS constants at the top. - Color plus percent labels. Accessible contrast. - Do not mix this with account-level retention. Feature only. </constraints> <format> Return the full HTML as an artifact, then explain first-use cohort math and which array to replace. </format>

You are a product analyst who builds unused-feature boards that force a teach, hide, or kill decision. <context> I need an unused-feature graveyard dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic feature reach so I can click the working screen in the chat. </context> <inputs> - Features in the set: [FEATURE SET] - Unused rule: [E.G. REACH UNDER 5% AND LAST USE OVER 30 DAYS] - Eligible users: [COUNT] - Ship dates if known: [OR "UNKNOWN"] - Product name: [PRODUCT] </inputs> <task> Build the dashboard with: KPI cards (features in set, unused now, median reach, features never used in 30 days); a reach bar sorted lowest first; a table (feature, unique users, reach, last use, days since last use, shipped, suggested action teach/hide/kill from simple rules); and a kill-or-teach list. Compute unused from my rule. Features shipped under 30 days ago get a too-new tag, not unused. Action rules live in editable constants. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Feature rows in labeled JS constants at the top. - Suggested actions are rules I can edit, not a strategy memo. - Accessible contrast. Sort default: reach ascending. </constraints> <format> Return the full HTML as an artifact, then list the unused and too-new rules and which array to replace. </format>

You are a product operations analyst who builds roadmap-health boards a weekly product review can trust. <context> I need a roadmap health dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic initiatives so I can click the working screen in the chat. </context> <inputs> - Quarter or cycle: [QUARTER OR CYCLE] - Initiatives: [6-12 NAMES OR "INVENT REALISTIC ONES"] - Statuses: [ON TRACK, AT RISK, SLIPPED, BLOCKED, SHIPPED] - Owners: [NAMES OR SQUADS] - Promised dates: [OR ASK CLAUDE TO INVENT A CONSISTENT SET] </inputs> <task> Build the dashboard with: a header for [QUARTER OR CYCLE]; KPI cards (initiatives, on track, slipped, blocked); a status donut; a swim-lane or card board by status with owner and promised date; a table (name, status, owner, promised, days to date, last update); and a slipped-plus-blocked watchlist. Compute days to promised from a today constant. Status chips filter the board and table. Clicking a card opens the row. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Initiatives in labeled JS constants at the top. - Slipped and blocked visually loud. Accessible contrast. - Do not invent customer quotes. Dates and status only. </constraints> <format> Return the full HTML as an artifact, then explain days-to-date math and which array to replace with a tracker export. </format>

You are a product operations analyst who builds delivery-confidence boards from remaining work and the calendar, not a gut emoji. <context> I need a delivery-confidence dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic remaining work so I can click the working screen in the chat. </context> <inputs> - Initiatives: [INITIATIVES] - Remaining work unit: [E.G. POINTS / DAYS] - Weekly throughput: [NUMBER OR "INVENT PLAUSIBLE"] - Target ship dates: [PER INITIATIVE] - Confidence bands: [E.G. HIGH OVER 80, MED 50-80, LOW UNDER 50] </inputs> <task> Build the dashboard with: KPI cards (initiatives, low confidence, weeks of work left, first miss); a confidence bar per initiative; a remaining-versus-capacity view; a table (initiative, remaining, throughput, weeks needed, weeks left, confidence 0-100, band); and a ship-risk list where weeks needed exceed weeks left. Compute weeks needed as remaining / throughput. Compute confidence from that gap using a commented formula. Sort default: lowest confidence first. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Remaining work and dates in labeled JS constants at the top. - Show the formula. Accessible contrast. - Do not invent hiring. Show the gap only. </constraints> <format> Return the full HTML as an artifact, then write the confidence formula and which constants to replace. </format>

You are a product analyst who puts shipped work next to the outcome it was supposed to move. <context> I need an outcome-versus-output dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with a realistic metric series and ship log so I can click the working screen in the chat. </context> <inputs> - Outcome metric: [OUTCOME METRIC] - Target or direction: [E.G. RAISE ACTIVATION TO 40%] - Weeks: [E.G. LAST 12] - Shipped items: [8-15 NAMES WITH WEEK SHIPPED OR "INVENT"] - Product name: [PRODUCT] </inputs> <task> Build the dashboard with: a header naming [OUTCOME METRIC]; KPI cards (latest value, change since first ship in range, ships this cycle, weeks the metric moved the wrong way); a line of the outcome with ship markers; a ship table (name, week, owner, intended effect in my words if given); and a callout of ships that landed in a flat or falling window. Keep the metric series and the ship log in separate labeled constants. Markers must be clickable and highlight the row. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Do not claim a ship caused a move. Show timing only. - Accessible contrast. No causal language in labels. - Do not invent a second outcome metric. </constraints> <format> Return the full HTML as an artifact, then explain how markers bind to weeks and which arrays to replace. </format>

You are a product operations analyst who builds slip-and-scope watchlists that show late and fatter as two different problems. <context> I need a slip-and-scope dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic initiatives and added tickets so I can click the working screen in the chat. </context> <inputs> - Original ship date focus: [SHIP DATE] - Initiatives: [LIST] - Scope unit: [E.G. TICKETS / POINTS] - Added-after-promise rule: [E.G. TICKETS CREATED AFTER COMMIT DATE] - Cut options I allow: [E.G. DEFER, CUT, KEEP] </inputs> <task> Build the dashboard with: KPI cards (days late on the focus date, added scope after promise, items already cut, items still in); a grouped bar of promised versus current scope; a slip table (initiative, promised date, current date, days late, added scope, owner); and a cut list of added items with a suggested defer/cut/keep from simple rules I can edit. Compute days late from promised versus current. Compute added scope from items flagged after the commit date. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Initiatives and items in labeled JS constants at the top. - Suggested cuts are rules, not a mandate. Accessible contrast. - Do not invent stakeholder quotes. </constraints> <format> Return the full HTML as an artifact, then write late and added-scope math and which arrays to replace. </format>

You are a product operations analyst who builds discovery-versus-delivery mix boards a team can correct next cycle. <context> I need a discovery-versus-delivery dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic cycle hours so I can click the working screen in the chat. </context> <inputs> - Team: [TEAM] - Cycle: [E.G. LAST 6 WEEKS] - Discovery share target: [E.G. 20-30%] - Work items with type DISCOVERY or DELIVERY: [OR "INVENT"] - Aging bet rule: [E.G. DISCOVERY OPEN OVER 21 DAYS] </inputs> <task> Build the dashboard with: a header for [TEAM]; KPI cards (discovery share, delivery share, versus target band, aging discovery bets); a weekly stacked bar of the two buckets; an aging-bet table (item, age, owner, last note); a starved-research callout when share sits under the band for 3+ weeks; and a simple item table with type chips. Compute share from hours or from item counts, and label the unit on the page. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Items and hours in labeled JS constants at the top. - Under-band weeks flagged. Accessible contrast. - Do not invent research findings. Mix and age only. </constraints> <format> Return the full HTML as an artifact, then write share and aging rules and which arrays to replace. </format>

You are a product research analyst who builds NPS trend boards that show score and sample size together. <context> I need an NPS trend dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic monthly scores so I can click the working screen in the chat. </context> <inputs> - Product name: [PRODUCT] - Months: [E.G. LAST 12] - Sent and responded per month: [OR "INVENT PLAUSIBLE"] - Score range: [0-10] - Target NPS: [E.G. 40] </inputs> <task> Build the dashboard with: a header for [PRODUCT]; KPI cards (latest NPS, change versus prior, response rate, promoters minus detractors); a line of monthly NPS with the target; a stacked bar of promoter / passive / detractor counts; a month table (sent, responses, rate, promoters, passives, detractors, NPS); and a thin-sample flag when responses sit under an editable threshold. Compute NPS as (% promoters minus % detractors) from 0-10 scores. Promoters 9-10, passives 7-8, detractors 0-6. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Monthly rows or raw scores in labeled JS constants at the top. - Show the NPS formula. Accessible contrast. - Do not invent industry benchmarks I did not give. </constraints> <format> Return the full HTML as an artifact, then write the NPS formula and which constants to replace with a survey export. </format>

You are a product research analyst who reads NPS by the segment a team can actually change. <context> I need an NPS-by-segment dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic tagged responses so I can click the working screen in the chat. </context> <inputs> - Segment to lead with: [SEGMENT] - Other segment cuts: [E.G. PLAN, TENURE, PLATFORM] - Period: [E.G. LAST 90 DAYS] - Product name: [PRODUCT] - Minimum responses to show a score: [E.G. 20] </inputs> <task> Build the dashboard with: a cut selector defaulting to [SEGMENT]; KPI cards (overall NPS, best slice, worst slice, slices under the minimum); a bar of NPS by slice; a volume bar; a table (slice, responses, promoters, detractors, NPS, versus overall); and a worst-cohort callout. Compute NPS per slice from tagged 0-10 scores. Slices under the minimum show n/a, not a wild score. Clicking a slice lists sample scores. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Responses with tags in labeled JS constants at the top. - n/a labeled when thin. Accessible contrast. - Do not invent a segment I did not list. </constraints> <format> Return the full HTML as an artifact, then explain the minimum-n rule and which array to replace. </format>

You are a product research analyst who builds promoter-passive-detractor boards with movement when a prior score exists. <context> I need a promoter-passive-detractor dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic current and prior scores so I can click the working screen in the chat. </context> <inputs> - Survey window: [SURVEY WINDOW] - Product name: [PRODUCT] - Prior window if any: [E.G. PRIOR 90 DAYS OR "NONE"] - Accounts or users: [SAMPLE SET] - Follow-up owner field: [E.G. CSM OR PM] </inputs> <task> Build the dashboard with: KPI cards (promoters, passives, detractors, newly detracted if prior exists); a three-bucket donut; a movement matrix when prior scores exist (from bucket to bucket); a detractor table (id, score, prior score, comment snippet, owner); and chips for bucket. Compute buckets from 0-10. Movement only when both scores exist. Missing prior shows as first score, not a drop. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Score rows in labeled JS constants at the top. - Detractor rows visually first. Accessible contrast. - Do not invent comments I did not seed. Empty comment is allowed. </constraints> <format> Return the full HTML as an artifact, then explain bucket and movement rules and which array to replace. </format>

You are a product research analyst who turns NPS verbatims into a theme board, not a quote collage. <context> I need a verbatim-theme dashboard as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic tagged comments so I can click the working screen in the chat. </context> <inputs> - Themes: [FEEDBACK THEMES] - Product name: [PRODUCT] - Period: [E.G. LAST 90 DAYS] - Sample comments: [OR ASK CLAUDE TO INVENT PLAUSIBLE SHORT QUOTES] - Score on each row: [0-10] </inputs> <task> Build the dashboard with: KPI cards (comments tagged, top theme, theme with worst NPS, untagged count); a bar of mention volume by theme; a bar of NPS among rows tagged with that theme; a quote drawer for the selected theme (quote, score, date); and a table (theme, mentions, share, NPS, promoters, detractors). A row may have one theme for simplicity unless I ask for multi-tag. Clicking a theme filters the quotes. Compute theme NPS from those rows only. </task> <constraints - One self-contained HTML file; one CDN charting library only. - Comments and themes in labeled JS constants at the top. - Quotes stay short. Accessible contrast. - Do not invent a theme I did not list. Untagged is its own bucket. </constraints> <format> Return the full HTML as an artifact, then explain theme NPS math and which arrays to replace with real replies. </format>

You are a product research analyst who builds close-the-loop boards that keep promised replies from going stale. <context> I need a close-the-loop tracker as one self-contained HTML file with inline CSS, vanilla JS, and a CDN chart library, seeded with realistic follow-up rows so I can click the working screen in the chat. </context> <inputs> - Owner team: [OWNER TEAM] - Product name: [PRODUCT] - Statuses: [E.G. NEW, ASSIGNED, REPLIED, WONT REPLY, CLOSED] - SLA to first reply: [E.G. 5 BUSINESS DAYS] - Survey window: [E.G. LAST 60 DAYS] </inputs> <task> Build the dashboard with: a header for [OWNER TEAM]; KPI cards (open follow-ups, past SLA, median age, closed this week); a stacked bar of open rows by owner; an aging table; a promised-reply list (id, score, theme if any, promised date, owner, last customer-facing send); and a case table with search. Age from survey date, never from last internal note. Status and owner filters must work. Selecting an owner filters every widget. </task> <constraints> - One self-contained HTML file; one CDN charting library only. - Follow-up rows in labeled JS constants at the top. - Past-SLA rows red. Search works. Accessible contrast. - Do not invent refund or legal language. Follow-up status only. </constraints> <format> Return the full HTML as an artifact, then explain the age and SLA rules and which array to replace with a follow-up log. </format>