[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M71E7EBTC17V133RNX31VNMZ":3,"skill-same:01M71E7EBTC17V133RNX31VNMZ":86},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":28,"publishedAt":49,"linkOnly":50,"markdownContent":51,"markdownBody":52,"frontMatter":53,"refName":54,"commitSha":55,"skillDirectory":56,"forksCount":14,"repositoryHtmlUrl":57,"repositoryHomepageUrl":6,"ownerHtmlUrl":58,"ownerBlogUrl":59,"ownerDescription":60,"updatedAt":61,"install":62,"resources":75},"01M71E7EBTC17V133RNX31VNMZ","configuring-experiment-analytics",null,"Configures the analytics side of a PostHog experiment — exposure criteria (server-resolved default exposure event vs custom exposure events), primary and secondary metrics, the supported metric types (count, sum, ratio with `math` and `math_property`, retention with `retention_window_start` and `start_handling`), multivariate user handling (\"Exclude\" vs \"First seen variant\"), and how to read results once the experiment is live. Use when the user adds or edits a primary or secondary metric (e.g. \"add a secondary metric tracking 'downloaded_file' per user\"), sets up a ratio metric (e.g. \"revenue from purchase_completed \u002F pageviews\"), sets up a retention metric (e.g. \"$pageview → uploaded_file, 7-day window\"), configures custom exposure (e.g. \"only count users who hit \u002Fcheckout\"), changes multivariate handling, or asks \"who is in the analysis?\", \"how do I measure impact?\", \"is this winning?\", \"what's the confidence level?\", or \"should I ship?\".","patinaproject\u002Fskills","patinaproject","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F229996889?v=4","01M71E58JKTDNY1138WYZK44TR","Patina Project","01M71E58K5GH4K16MV6A8BQ6NZ",1,"MIT","https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fconfiguring-experiment-analytics\u002FSKILL.md",[18,23],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000013","分析与可视化","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":26,"taxonomyTypeDesc":27},"01JBSEEDCAT000000000000033","数据分析师","PROFESSION","适用职业",[29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48],"configuring experiment analytics","experiment","analytics","impact","exposure","when","window","sum","purchase","types","retention","checkout","how to","default","up","only","math","analysis","server","posthog","2026-11-10 10:12:30",false,"---\nname: configuring-experiment-analytics\ndescription: Configures the analytics side of a PostHog experiment — exposure criteria (server-resolved default exposure event vs custom exposure events), primary and secondary metrics, the supported metric types (count, sum, ratio with `math` and `math_property`, retention with `retention_window_start` and `start_handling`), multivariate user handling (\"Exclude\" vs \"First seen variant\"), and how to read results once the experiment is live. Use when the user adds or edits a primary or secondary metric (e.g. \"add a secondary metric tracking 'downloaded_file' per user\"), sets up a ratio metric (e.g. \"revenue from purchase_completed \u002F pageviews\"), sets up a retention metric (e.g. \"$pageview → uploaded_file, 7-day window\"), configures custom exposure (e.g. \"only count users who hit \u002Fcheckout\"), changes multivariate handling, or asks \"who is in the analysis?\", \"how do I measure impact?\", \"is this winning?\", \"what's the confidence level?\", or \"should I ship?\".\n---\n\n# Configuring experiment analytics\n\nThis skill answers: **Who is included in the analysis?** and **How to measure impact?**\n\n## Exposure criteria\n\nExposure criteria determine which users are counted in the experiment analysis.\n\n### Include people when\n\nTwo options:\n\n1. **Default exposure event** — users are included when the experiment's default exposure event fires for the experiment's flag: `$feature_flag_called`, or `$experiment_exposure` for newer experiments. Which one applies is resolved server-side — read `resolved_exposure_event` from `experiment-get` rather than assuming either name (both events carry the same properties). This is the standard approach — it means a user is included only when they actually encounter the feature flag in your code.\n2. **Custom exposure event** — users are included when a specific custom event fires. Use this when you want tighter control over who enters the analysis (e.g., only users who actually visit the page where the experiment runs).\n\n### Multiple variant handling\n\nWhen a user is exposed to multiple variants (e.g., due to flag changes or race conditions):\n\n- **Exclude multivariate users** — removes these users from the analysis entirely. Cleaner data, smaller sample.\n- **First seen variant** — assigns users to the first variant they were exposed to. Keeps all users in the analysis. Note that \"first seen\" can introduce other biases as\n  behavior cannot be clearly attributed to a single variant and is not recommended unless necessary.\n\n**Bias risk on uneven splits.** \"Exclude multivariate users\" combined with an uneven variant split can\nintroduce bias — multi-variant users are dropped asymmetrically and the smaller variant loses a larger\nfraction of its assignments. If those users behave differently from the rest, the smaller variant's\nmetrics will be skewed.\n\nThe right mitigation depends on experiment state:\n\n- **Not yet launched, or only exposed to a few users so far** — switch to an even variant split and\n  use the overall rollout percentage to limit test-variant exposure. This removes the bias and\n  preserves statistical power. See `configuring-experiment-rollout`.\n- **Live experiment with significant exposures** — changing the split mid-run reassigns users across\n  variants, which is bad for user experience and data quality. Switch this setting to \"First seen\n  variant\" instead — it keeps already-assigned users in their original variant (no reassignment) and\n  removes the asymmetric exclusion.\n\n### Filter test accounts\n\n`exposure_criteria.filterTestAccounts` (default: true) — excludes internal\u002Ftest users from the analysis.\n\n## Resolving experiments\n\nMetric changes require an experiment ID. If the user refers to an experiment by name\nor description (e.g. \"add metrics to the checkout test\"), load the `finding-experiments`\nskill to resolve it to a concrete ID before proceeding.\n\n## Metrics\n\nA metric reaches an experiment one of two ways, both via `experiment-update`:\n\n- **Inline metric** — defined directly on the experiment. Sent in the `metrics` array, which\n  **replaces** the entire inline list, so always get the current experiment first via `experiment-get`\n  to preserve existing metrics.\n- **Shared (saved) metric** — a reusable metric object that can be attached to many experiments.\n  Attached by ID via `saved_metrics_ids` (this list also **replaces** the experiment's existing\n  saved-metric links, so resend the full set — see Step 1).\n\n**Prefer reusing a shared metric over duplicating it inline.** Build a new inline metric only when\nno suitable shared metric already exists.\n\n### Step 1: Check for an existing shared metric (REQUIRED — match by definition, not name)\n\nBefore building any new inline metric, you MUST check whether the project already has a shared\n(saved) metric that measures the same thing, and reuse it. Duplicating a metric that already exists\nas a shared metric fragments measurement and is exactly what we want to avoid.\n\n**Reuse is decided by the metric _definition_ — the event or action plus the metric type — not the\nname.** Saved metrics are named by each team's own conventions, which you cannot guess, so you must\ncompare on what each metric measures (its `query`), never on its title.\n\n**Workflow:**\n\n1. **Know what you're about to build first.** Settle the target event(s)\u002Faction(s) and metric type\n   (mean \u002F funnel \u002F ratio \u002F retention) before searching — see Step 2 to confirm the event exists via\n   `read-data-schema`. You can only recognize a duplicate once you know the concrete event\u002Faction,\n   so this check runs _after_ you've pinned down the event, not before.\n2. **Search by the event, then compare each candidate's `query`.** Call `experiment-saved-metrics-list`\n   with `?event=\u003Cthe event you're measuring>` to find metrics that reference it — matched directly (an\n   `EventsNode`) **or** via the step events of any action a metric references, so action-based metrics are\n   found by the event their action fires on. Then for each returned row, inspect its **`query`** (not the\n   `name`\u002F`description`): a saved metric is a reuse match when its `query` measures the **same event or\n   action with the same `metric_type`** (and compatible `math`) as the metric you'd otherwise build, even\n   if its name is different.\n   - **Match on the event, not the action's name.** An action-based metric is discoverable by the event\n     the action fires on — pass that event, not the action's label.\n   - **Do not use `search` for this.** `search` matches only the metric's own `name` \u002F `description` \u002F tags —\n     never the underlying event or action — so it cannot find a definition match. Use `search` only when the\n     user names a specific saved metric to attach (name resolution, not a definition match).\n3. **If a saved metric matches the definition** — confirm the match with the user by name\u002Fdescription,\n   then attach it instead of building a new one:\n   - Call `experiment-get` to read the experiment's current `saved_metrics`.\n   - Call `experiment-update` with `saved_metrics_ids` set to the full desired set — it **replaces**\n     existing links, so include the already-attached ones plus the new entry. Each entry has shape\n     `{ \"id\": \u003Csaved-metric id>, \"metadata\": { \"type\": \"primary\" } }` — set `type` to `\"primary\"` or\n     `\"secondary\"`. `metadata` is optional and defaults to primary.\n   - **Watch the id when rebuilding the set:** each item in the `saved_metrics` you just read has a\n     top-level `id` (the _link_ id) AND a `saved_metric` field (the _metric_ id). `saved_metrics_ids`\n     wants the **`saved_metric`** value, not the link `id` — sending the link `id` attaches the wrong\n     metric or fails validation.\n   - You do not need to build the inline metric — the shared metric already encodes its events.\n4. **If nothing in the library measures the same event\u002Faction + type** — build an inline metric\n   (Step 2+). When that inline metric is likely to be reused across experiments, offer to create it\n   as a shared metric instead, via `experiment-saved-metrics-create`, then attach it as above, so the\n   next experiment can reuse it.\n\n### Step 2: Discover available events (REQUIRED before building an inline metric)\n\nBefore suggesting or building any new inline metric, you MUST call `read-data-schema` to discover\nwhat events actually exist in the project. Do NOT skip this step. Do NOT suggest event names\nbased on what you think the project might track — only use events you have confirmed exist.\n(Attaching an existing shared metric from Step 1 does not need this — it already encodes its events.)\n\nThis applies even when:\n\n- The user provides event names — look them up to confirm they exist and are spelled correctly\n- The user asks \"what metrics do you suggest?\" — look up events first, then suggest from real data\n- The context makes certain events seem obvious — they may not exist or may be named differently\n\n**Workflow:**\n\n1. Call `read-data-schema` to get the project's events\n2. Present relevant events to the user based on the experiment's hypothesis\n3. User picks which events to use for metrics\n4. Configure metrics with those confirmed event names\n\n**Legitimate exception — `allow_unknown_events: true`:**\nPass this on `experiment-create` \u002F `experiment-update` only when the user is intentionally instrumenting an event that hasn't been ingested yet (e.g. setting up the experiment before the code change ships). Confirm this with the user — never use it as a workaround for \"the event lookup didn't return what I expected\".\n\n**Example:**\n\n```text\nUser: \"Let's add some metrics for the checkout experiment\"\n\nWRONG: \"I'd suggest using purchase_completed as the primary metric...\"\n  (hallucinated event name — never seen the project's actual events)\n\nRIGHT: *calls read-data-schema* → \"Here are the events in your project\n  related to checkout: `checkout_step_completed`, `payment_processed`,\n  `order_confirmed`. Which of these represents a successful checkout?\"\n```\n\n### Step 3: Choose metric type\n\nThere are four metric types. Each has `kind: \"ExperimentMetric\"`:\n\n| metric_type   | When to use                                                                            | Required fields                                                                                                                |\n| ------------- | -------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------ |\n| `\"mean\"`      | Average of a numeric property per user (revenue, session duration, pageviews per user) | `source`                                                                                                                       |\n| `\"funnel\"`    | Conversion rate from exposure through one or more ordered actions                      | `series` (1 or more steps)                                                                                                     |\n| `\"ratio\"`     | Rate of one event relative to another                                                  | `numerator`, `denominator` — set `math: \"sum\"` + `math_property` on a side to aggregate a property; filters never aggregate    |\n| `\"retention\"` | Do users come back after exposure?                                                     | `start_event`, `completion_event`, `retention_window_start`, `retention_window_end`, `retention_window_unit`, `start_handling` |\n\n**Funnel metrics and the implicit exposure step**\n\nFunnel metrics automatically prepend the experiment's exposure event as `step_0`.\nSo a funnel with 1 step in `series` is a valid 2-step funnel: **exposure → action**.\nThis is the correct choice for measuring \"what percentage of exposed users did X?\"\n\nExamples:\n\n- \"What % of exposed users reached \u002Flogin?\" → funnel with 1 step (`$pageview` filtered to \u002Flogin)\n- \"What % of exposed users completed checkout?\" → funnel with 1 step (`checkout_completed`)\n- \"What % of exposed users went cart → checkout → purchase?\" → funnel with 3 steps\n\n**Mean vs funnel for the same event**\n\n- **Mean** measures average count\u002Fvalue per user (e.g. \"pageviews per user\", \"revenue per user\").\n- **Funnel** measures conversion rate (e.g. \"% of exposed users who purchased\").\n\nBoth can reference the same event — the difference is whether you care about count\u002Fmagnitude (mean) or yes\u002Fno conversion (funnel).\n\n**Retention: same vs different start\u002Fcompletion event**\n\nThe retention window is measured from the start event, so the events you pick decide what's measured:\nThe start occurrence never counts as its own completion (only a distinct later event does), so both shapes are valid:\n\n- **Different** start and completion events → conversion-style retention (\"did they reach the target action within the window?\").\n- **Same** event → repeat retention (\"did they fire it _again_?\"). `From 0` counts a repeat from the same period onward (same-day repeats included); `From ≥ 1` requires an occurrence later. Use `start_handling: \"first_seen\"`. When a user says \"retention of `\u003Cevent>`\" they usually mean repeat retention.\n\nSee `references\u002Fmetric-configuration.md` for the full rendered `ExperimentMetric` schema (all four metric types, with required fields per type) plus WRONG\u002FRIGHT JSON pairs for the failure modes that come up most often (ratio with `is_set` filter instead of `math: \"sum\"` + `math_property`; retention without `retention_window_start` \u002F `start_handling`). Read it before assembling a ratio or retention payload — the required fields are authoritative.\n\n### Step 4: Primary vs secondary\n\n- **Primary metrics** — the main success criteria for the experiment. These drive the ship\u002Fend decision.\n- **Secondary metrics** — additional measurements for context. Useful for guardrail metrics (e.g., ensuring a conversion improvement doesn't increase error rates).\n\n## Interpreting results\n\nSee `references\u002Finterpreting-results.md` for guidance on reading experiment results, statistical significance, and when to ship vs end.\n\n## Related skills\n\n- **`configuring-experiment-rollout`** — the rollout side: variant splits and traffic percentage\n- **`diagnosing-experiment-results`** — when results look biased, empty, or strange\n- **`analyzing-experiment-session-replays`** — qualitative complement — watch what each variant's users actually did\n","# Configuring experiment analytics\n\nThis skill answers: **Who is included in the analysis?** and **How to measure impact?**\n\n## Exposure criteria\n\nExposure criteria determine which users are counted in the experiment analysis.\n\n### Include people when\n\nTwo options:\n\n1. **Default exposure event** — users are included when the experiment's default exposure event fires for the experiment's flag: `$feature_flag_called`, or `$experiment_exposure` for newer experiments. Which one applies is resolved server-side — read `resolved_exposure_event` from `experiment-get` rather than assuming either name (both events carry the same properties). This is the standard approach — it means a user is included only when they actually encounter the feature flag in your code.\n2. **Custom exposure event** — users are included when a specific custom event fires. Use this when you want tighter control over who enters the analysis (e.g., only users who actually visit the page where the experiment runs).\n\n### Multiple variant handling\n\nWhen a user is exposed to multiple variants (e.g., due to flag changes or race conditions):\n\n- **Exclude multivariate users** — removes these users from the analysis entirely. Cleaner data, smaller sample.\n- **First seen variant** — assigns users to the first variant they were exposed to. Keeps all users in the analysis. Note that \"first seen\" can introduce other biases as\n  behavior cannot be clearly attributed to a single variant and is not recommended unless necessary.\n\n**Bias risk on uneven splits.** \"Exclude multivariate users\" combined with an uneven variant split can\nintroduce bias — multi-variant users are dropped asymmetrically and the smaller variant loses a larger\nfraction of its assignments. If those users behave differently from the rest, the smaller variant's\nmetrics will be skewed.\n\nThe right mitigation depends on experiment state:\n\n- **Not yet launched, or only exposed to a few users so far** — switch to an even variant split and\n  use the overall rollout percentage to limit test-variant exposure. This removes the bias and\n  preserves statistical power. See `configuring-experiment-rollout`.\n- **Live experiment with significant exposures** — changing the split mid-run reassigns users across\n  variants, which is bad for user experience and data quality. Switch this setting to \"First seen\n  variant\" instead — it keeps already-assigned users in their original variant (no reassignment) and\n  removes the asymmetric exclusion.\n\n### Filter test accounts\n\n`exposure_criteria.filterTestAccounts` (default: true) — excludes internal\u002Ftest users from the analysis.\n\n## Resolving experiments\n\nMetric changes require an experiment ID. If the user refers to an experiment by name\nor description (e.g. \"add metrics to the checkout test\"), load the `finding-experiments`\nskill to resolve it to a concrete ID before proceeding.\n\n## Metrics\n\nA metric reaches an experiment one of two ways, both via `experiment-update`:\n\n- **Inline metric** — defined directly on the experiment. Sent in the `metrics` array, which\n  **replaces** the entire inline list, so always get the current experiment first via `experiment-get`\n  to preserve existing metrics.\n- **Shared (saved) metric** — a reusable metric object that can be attached to many experiments.\n  Attached by ID via `saved_metrics_ids` (this list also **replaces** the experiment's existing\n  saved-metric links, so resend the full set — see Step 1).\n\n**Prefer reusing a shared metric over duplicating it inline.** Build a new inline metric only when\nno suitable shared metric already exists.\n\n### Step 1: Check for an existing shared metric (REQUIRED — match by definition, not name)\n\nBefore building any new inline metric, you MUST check whether the project already has a shared\n(saved) metric that measures the same thing, and reuse it. Duplicating a metric that already exists\nas a shared metric fragments measurement and is exactly what we want to avoid.\n\n**Reuse is decided by the metric _definition_ — the event or action plus the metric type — not the\nname.** Saved metrics are named by each team's own conventions, which you cannot guess, so you must\ncompare on what each metric measures (its `query`), never on its title.\n\n**Workflow:**\n\n1. **Know what you're about to build first.** Settle the target event(s)\u002Faction(s) and metric type\n   (mean \u002F funnel \u002F ratio \u002F retention) before searching — see Step 2 to confirm the event exists via\n   `read-data-schema`. You can only recognize a duplicate once you know the concrete event\u002Faction,\n   so this check runs _after_ you've pinned down the event, not before.\n2. **Search by the event, then compare each candidate's `query`.** Call `experiment-saved-metrics-list`\n   with `?event=\u003Cthe event you're measuring>` to find metrics that reference it — matched directly (an\n   `EventsNode`) **or** via the step events of any action a metric references, so action-based metrics are\n   found by the event their action fires on. Then for each returned row, inspect its **`query`** (not the\n   `name`\u002F`description`): a saved metric is a reuse match when its `query` measures the **same event or\n   action with the same `metric_type`** (and compatible `math`) as the metric you'd otherwise build, even\n   if its name is different.\n   - **Match on the event, not the action's name.** An action-based metric is discoverable by the event\n     the action fires on — pass that event, not the action's label.\n   - **Do not use `search` for this.** `search` matches only the metric's own `name` \u002F `description` \u002F tags —\n     never the underlying event or action — so it cannot find a definition match. Use `search` only when the\n     user names a specific saved metric to attach (name resolution, not a definition match).\n3. **If a saved metric matches the definition** — confirm the match with the user by name\u002Fdescription,\n   then attach it instead of building a new one:\n   - Call `experiment-get` to read the experiment's current `saved_metrics`.\n   - Call `experiment-update` with `saved_metrics_ids` set to the full desired set — it **replaces**\n     existing links, so include the already-attached ones plus the new entry. Each entry has shape\n     `{ \"id\": \u003Csaved-metric id>, \"metadata\": { \"type\": \"primary\" } }` — set `type` to `\"primary\"` or\n     `\"secondary\"`. `metadata` is optional and defaults to primary.\n   - **Watch the id when rebuilding the set:** each item in the `saved_metrics` you just read has a\n     top-level `id` (the _link_ id) AND a `saved_metric` field (the _metric_ id). `saved_metrics_ids`\n     wants the **`saved_metric`** value, not the link `id` — sending the link `id` attaches the wrong\n     metric or fails validation.\n   - You do not need to build the inline metric — the shared metric already encodes its events.\n4. **If nothing in the library measures the same event\u002Faction + type** — build an inline metric\n   (Step 2+). When that inline metric is likely to be reused across experiments, offer to create it\n   as a shared metric instead, via `experiment-saved-metrics-create`, then attach it as above, so the\n   next experiment can reuse it.\n\n### Step 2: Discover available events (REQUIRED before building an inline metric)\n\nBefore suggesting or building any new inline metric, you MUST call `read-data-schema` to discover\nwhat events actually exist in the project. Do NOT skip this step. Do NOT suggest event names\nbased on what you think the project might track — only use events you have confirmed exist.\n(Attaching an existing shared metric from Step 1 does not need this — it already encodes its events.)\n\nThis applies even when:\n\n- The user provides event names — look them up to confirm they exist and are spelled correctly\n- The user asks \"what metrics do you suggest?\" — look up events first, then suggest from real data\n- The context makes certain events seem obvious — they may not exist or may be named differently\n\n**Workflow:**\n\n1. Call `read-data-schema` to get the project's events\n2. Present relevant events to the user based on the experiment's hypothesis\n3. User picks which events to use for metrics\n4. Configure metrics with those confirmed event names\n\n**Legitimate exception — `allow_unknown_events: true`:**\nPass this on `experiment-create` \u002F `experiment-update` only when the user is intentionally instrumenting an event that hasn't been ingested yet (e.g. setting up the experiment before the code change ships). Confirm this with the user — never use it as a workaround for \"the event lookup didn't return what I expected\".\n\n**Example:**\n\n```text\nUser: \"Let's add some metrics for the checkout experiment\"\n\nWRONG: \"I'd suggest using purchase_completed as the primary metric...\"\n  (hallucinated event name — never seen the project's actual events)\n\nRIGHT: *calls read-data-schema* → \"Here are the events in your project\n  related to checkout: `checkout_step_completed`, `payment_processed`,\n  `order_confirmed`. Which of these represents a successful checkout?\"\n```\n\n### Step 3: Choose metric type\n\nThere are four metric types. Each has `kind: \"ExperimentMetric\"`:\n\n| metric_type   | When to use                                                                            | Required fields                                                                                                                |\n| ------------- | -------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------ |\n| `\"mean\"`      | Average of a numeric property per user (revenue, session duration, pageviews per user) | `source`                                                                                                                       |\n| `\"funnel\"`    | Conversion rate from exposure through one or more ordered actions                      | `series` (1 or more steps)                                                                                                     |\n| `\"ratio\"`     | Rate of one event relative to another                                                  | `numerator`, `denominator` — set `math: \"sum\"` + `math_property` on a side to aggregate a property; filters never aggregate    |\n| `\"retention\"` | Do users come back after exposure?                                                     | `start_event`, `completion_event`, `retention_window_start`, `retention_window_end`, `retention_window_unit`, `start_handling` |\n\n**Funnel metrics and the implicit exposure step**\n\nFunnel metrics automatically prepend the experiment's exposure event as `step_0`.\nSo a funnel with 1 step in `series` is a valid 2-step funnel: **exposure → action**.\nThis is the correct choice for measuring \"what percentage of exposed users did X?\"\n\nExamples:\n\n- \"What % of exposed users reached \u002Flogin?\" → funnel with 1 step (`$pageview` filtered to \u002Flogin)\n- \"What % of exposed users completed checkout?\" → funnel with 1 step (`checkout_completed`)\n- \"What % of exposed users went cart → checkout → purchase?\" → funnel with 3 steps\n\n**Mean vs funnel for the same event**\n\n- **Mean** measures average count\u002Fvalue per user (e.g. \"pageviews per user\", \"revenue per user\").\n- **Funnel** measures conversion rate (e.g. \"% of exposed users who purchased\").\n\nBoth can reference the same event — the difference is whether you care about count\u002Fmagnitude (mean) or yes\u002Fno conversion (funnel).\n\n**Retention: same vs different start\u002Fcompletion event**\n\nThe retention window is measured from the start event, so the events you pick decide what's measured:\nThe start occurrence never counts as its own completion (only a distinct later event does), so both shapes are valid:\n\n- **Different** start and completion events → conversion-style retention (\"did they reach the target action within the window?\").\n- **Same** event → repeat retention (\"did they fire it _again_?\"). `From 0` counts a repeat from the same period onward (same-day repeats included); `From ≥ 1` requires an occurrence later. Use `start_handling: \"first_seen\"`. When a user says \"retention of `\u003Cevent>`\" they usually mean repeat retention.\n\nSee `references\u002Fmetric-configuration.md` for the full rendered `ExperimentMetric` schema (all four metric types, with required fields per type) plus WRONG\u002FRIGHT JSON pairs for the failure modes that come up most often (ratio with `is_set` filter instead of `math: \"sum\"` + `math_property`; retention without `retention_window_start` \u002F `start_handling`). Read it before assembling a ratio or retention payload — the required fields are authoritative.\n\n### Step 4: Primary vs secondary\n\n- **Primary metrics** — the main success criteria for the experiment. These drive the ship\u002Fend decision.\n- **Secondary metrics** — additional measurements for context. Useful for guardrail metrics (e.g., ensuring a conversion improvement doesn't increase error rates).\n\n## Interpreting results\n\nSee `references\u002Finterpreting-results.md` for guidance on reading experiment results, statistical significance, and when to ship vs end.\n\n## Related skills\n\n- **`configuring-experiment-rollout`** — the rollout side: variant splits and traffic percentage\n- **`diagnosing-experiment-results`** — when results look biased, empty, or strange\n- **`analyzing-experiment-session-replays`** — qualitative complement — watch what each variant's users actually did\n",{"name":5,"description":7},"main","e6ee720f466a8344a52abd87679191fadde3b63f",".agents\u002Fskills\u002Fconfiguring-experiment-analytics","https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills","https:\u002F\u002Fgithub.com\u002Fpatinaproject","https:\u002F\u002Fwww.patinaproject.com","The best way to pick your next pair of boots or shoes.","2026-11-12 00:44:02",{"source":8,"skillName":5,"commands":63,"prompt":73,"zipAvailable":74},[64,67,70],{"tool":65,"command":66},"npx","npx skills add patinaproject\u002Fskills --skill configuring-experiment-analytics",{"tool":68,"command":69},"bunx","bunx skills add patinaproject\u002Fskills --skill configuring-experiment-analytics",{"tool":71,"command":72},"pnpm","pnpm dlx skills add patinaproject\u002Fskills --skill configuring-experiment-analytics","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fconfiguring-experiment-analytics\u002FSKILL.md",true,[76,82],{"path":77,"resourceType":78,"resourceTypeDesc":79,"sizeBytes":80,"sourceHtmlUrl":81},"references\u002Finterpreting-results.md","REFERENCE","参考文档",2223,"https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fconfiguring-experiment-analytics\u002Freferences\u002Finterpreting-results.md",{"path":83,"resourceType":78,"resourceTypeDesc":79,"sizeBytes":84,"sourceHtmlUrl":85},"references\u002Fmetric-configuration.md",115558,"https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fconfiguring-experiment-analytics\u002Freferences\u002Fmetric-configuration.md",[87,95,121,133,154,178],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":94},"01M72KDQJR99954NE224VB3KZ2","code-review","Review the changes since a fixed point (commit, branch, tag, or merge-base) along five axes: Standards (does the code follow this repo's documented coding standards?), Spec (does the code match what the originating issue\u002Fspec asked for?), Evidence (does each claim in the pull request have evidence that shows it?), Bugs (does the changed code have logic errors, unhandled states, races, or regressions?), and Security (does the change open injection, auth, secret, input, or CI workflow holes?). Runs the reviews in parallel sub-agents, tags every finding Important, Nit, or Pre-existing, and reports them side by side with what blocks merge. Given an earlier reviewed head and its findings, it reviews only the newer commits, rechecks the earlier findings, and never repeats one. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to \"review since X\".","https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002Fplugins\u002Fcode\u002Fskills\u002Fcode-review\u002FSKILL.md",[],[],"2026-11-10 22:39:42",{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":107,"publishedAt":49},"01M71ESEDVHD9EXN6787XHP7HD","writing-for-pstack","Write or tighten an operator prompt for a patina-mode run. Use when the operator asks to write, revise, or review a prompt that will be passed to patina-mode.","https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002Fplugins\u002Fwork\u002Fskills\u002Fwriting-for-pstack\u002FSKILL.md",[101,104],{"categoryNo":102,"categoryName":103,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":105,"categoryName":106,"taxonomyType":26,"taxonomyTypeDesc":27},"01JBSEEDCAT000000000000036","内容创作者",[108,109,110,111,112,113,114,115,116,117,118,119,34,120],"for","pstack","writing","writing for pstack","review","run","mode","write","operator","use","tighten","prompt","revise",{"skillNo":122,"name":123,"nameZh":6,"description":124,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":125,"categories":126,"keywords":129,"publishedAt":49},"01M71ESCJ0SSJFFENF8HYDKBTJ","writing-for-agents","Writing documents for agents. Use when creating or editing skills, or modifying AGENTS.md or CLAUDE.md.","https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002Fplugins\u002Fwork\u002Fskills\u002Fwriting-for-agents\u002FSKILL.md",[127,128],{"categoryNo":102,"categoryName":103,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":105,"categoryName":106,"taxonomyType":26,"taxonomyTypeDesc":27},[130,108,110,131,117,132,34],"writing for agents","editing","documents",{"skillNo":134,"name":135,"nameZh":6,"description":136,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":137,"categories":138,"keywords":139,"publishedAt":49},"01M71ESA1NN9WVMD2HA5Z1RRDY","wizard","Generate an interactive bash wizard that walks a human through steps only they can perform. Use when provisioning infrastructure, setting up credentials or CI secrets, walking an unfamiliar third-party dashboard, or running a one-off migration or cutover. Don't invoke this for steps the agent can perform itself.","https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002Fplugins\u002Fwork\u002Fskills\u002Fwizard\u002FSKILL.md",[],[135,140,141,142,143,144,34,145,146,147,148,149,117,150,151,43,108,152,153,44],"interactive","can","credentials","one","Bash","invoke","dashboard","generate","human","secrets","provisioning infrastructure","ci","setting","running",{"skillNo":155,"name":156,"nameZh":6,"description":157,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":158,"categories":159,"keywords":160,"publishedAt":49},"01M71ES8FAN567N3JEPG8W6AS1","wayfinder","Plan a huge chunk of work (more than one agent session can hold) as a shared map of decision tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.","https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002Fplugins\u002Fwork\u002Fskills\u002Fwayfinder\u002FSKILL.md",[],[156,161,162,163,164,165,166,167,168,169,170,171,172,173,174,175,176,141,177,143],"more","true","decision","session","time","chunk","as","shared","map","tickets","destination","at","tracker","plan","work","agent session","one at a time",{"skillNo":179,"name":180,"nameZh":6,"description":181,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":182,"categories":183,"keywords":184,"publishedAt":49},"01M71ES6PWJPKBW8R24YFH4PY6","wait-what","Stop. That last message did not land: re-pitch it.","https:\u002F\u002Fgithub.com\u002Fpatinaproject\u002Fskills\u002Fblob\u002Fmain\u002Fplugins\u002Fwork\u002Fskills\u002Fwait-what\u002FSKILL.md",[],[185,186,187,188,189,190,191,192,162,193],"wait","wait what","stop","it","land","re","pitch","message","last"]