Get Experiment Result

GET/v2/project/{project_key}/experiment/{test_id}/result

Return dashboard-parity conversion results for an experiment, with optional goal drilldowns, dashboard-compatible filters, and breakdowns.

Path parameters

Parameter	Description
`project_key`	Your project's API key
`test_id`	The numeric experiment ID

Query parameters

Parameter	Type	Description
`start_date`	string	Filter conversions from this date. Format: `YYYY-MM-DD`.
`end_date`	string	Filter conversions through this date. Format: `YYYY-MM-DD`.
`goal_key`	string	Drill into a specific primary or secondary goal by `goal_key`. If omitted, the experiment's attached primary goal is used when present; otherwise the project's global primary goal from Get Project Configuration is used.
`goal_profile_id`	string	Drill into a specific goal by internal goal ID. Prefer `goal_key` when available.
`filter_by`	string	Dashboard preset segment filter such as `mobile`, `desktop`, `new`, `returning`, `paid`, `search`, `direct`, or `country-US`.
`custom_filter`	string \| array	Dashboard custom segment rules in the same nested rule shape used by the dashboard segment builder.
`breakdown_by`	string	Segment results by a dashboard-supported dimension. See Supported breakdowns.
`event_attribute`	array	Dashboard event attribute filters for event, revenue, and script goals. Use the dashboard nested condition shape. See Event attribute filters.
`min_count`	number	Minimum matching event count per visitor for event-backed goals. Defaults to `1`.
`timeframe_days`	number	Event conversion window in days after test entry for event-backed goals.
`tz`	string	Dashboard timezone hour offset used for date filtering. Defaults to `0`.
`report_phase`	string	Which conversion period to include: `pre_deploy` (experiment split only), `post_deploy` (after Set winner & serve), or `all_time` (both). Defaults to `pre_deploy` when the experiment is actively serving a winner.

All query parameters are optional. Counts and conversion rates are produced by the same backend route as the Mida dashboard (/abtest/conversion). Date filters can be one-sided: if only start_date is supplied, results are filtered from that date through the current server date; if only end_date is supplied, results are filtered from the experiment creation date through that date. The previous Public V2 structured filters object is rejected for parity; use filter_by or custom_filter.

Supported breakdowns

device, browser, os, country, region, city, user_type, ref_page, landing_page, utm_campaign, utm_source, utm_medium, utm_content, utm_term

UTM values are public aliases. Public V2 validates them against this allowlist, maps them to dashboard-safe property.utm_* values before calling /abtest/conversion, and rejects unsupported values with 400.

user_type is derived from visitor age at test entry: new means the visitor record was created within 24 hours of the goal row timestamp; otherwise it is returning.

Supported filters

Use the same filter inputs as the dashboard:

filter_by for preset segments, for example mobile, desktop, new, returning, paid, search, direct, or country-US. Comma-separating different dimensions is AND (desktop,country-US = Desktop AND US); comma-separating multiple country-XX tokens is OR within country (desktop,country-US,country-CA = Desktop AND (US OR CA)).
custom_filter for the dashboard's segment-builder rules in the same nested array shape: inner arrays are OR groups, outer arrays are AND groups.

`custom_filter` criteria

These are the same criteria the dashboard's Custom Segment form exposes:

Criteria	Notes
`landing_page`, `ref_page`, `user_agent`	URL/string match against the visitor's landing page, referrer, or user-agent.
`queryparam`	Requires `key` (the query-string param name).
`source`	Fixed values: `direct`, `referral`, `social`, `search`, `nonpaid`, `paid`, `email`.
`device`	Fixed values: `mobile`, `desktop`, `tablet`.
`browser`	Fixed values: `chrome`, `firefox`, `safari`, `edge_chromium`, `edge_legacy`, `opera`, `ie`, `brave`, `vivaldi`, `yandex`, `chromium`, `samsung`, `uc`, `duckduckgo`, `silk`, `facebook`, `instagram`, `other`.
`os`	Fixed values: `windows`, `macos`, `ios`, `linux`.
`dayofweek`	Integer `0` (Monday) – `6` (Sunday). Also supports `>`, `>=`, `<`, `<=`.
`hourofday`	Integer `0` – `23`. Also supports `>`, `>=`, `<`, `<=`.
`visitor_type`	Fixed values: `new`, `returning`.
`country`	ISO 3166-1 alpha-2 uppercase (e.g. `US`, `GB`). Only `==`, `!=`, `*`, `!`.
`v.visitor_id`, `v.uuid`	Visitor identifiers. 1–80 chars of `[A-Za-z0-9_-]`. Only `==`, `!=`, `*`, `!`.
`attribute`	Visitor attribute. Requires `key`; supports extended operators including `gt`, `gte`, `lt`, `lte`, `bt`, `bf`, `days_gt`, `days_gte`, `days_lt`, `days_lte`.
`utm`	Requires `key` (the UTM param name).

Operators (unless otherwise noted): ==, !=, ** (contains), !* (does not contain), ex (exists), nx (does not exist).

region and city are not custom_filter criteria — use a filter_by regional preset (northamerica, europe, asiapacific, latinamerica, africa, middleeast) or breakdown_by=region/breakdown_by=city to split results by them.

event_attribute must be valid JSON in the nested dashboard shape: an outer OR array containing inner AND arrays of filter objects. Each object supports key, operand, and value; operator is accepted as an alias for operand.

Supported operands are =, !=, >, <, >=, <=, ** (contains), !* (does not contain), ex (exists), and nx (does not exist). Keys may contain only letters, numbers, _, ., and -, up to 80 characters. Values are capped at 500 characters.

Examples

curl "https://api-{region}.mida.so/v2/project/YOUR_PROJECT_KEY/experiment/1234/result" \
  -H "Authorization: Bearer YOUR_GENERATED_API_KEY"

With date filtering:

curl "https://api-{region}.mida.so/v2/project/YOUR_PROJECT_KEY/experiment/1234/result?start_date=2024-01-15&end_date=2024-02-15" \
  -H "Authorization: Bearer YOUR_GENERATED_API_KEY"

Filtered to mobile traffic:

curl -G "https://api-{region}.mida.so/v2/project/YOUR_PROJECT_KEY/experiment/1234/result" \
  -H "Authorization: Bearer YOUR_GENERATED_API_KEY" \
  --data-urlencode 'filter_by=mobile'

Custom segment — Desktop AND (US OR Canada) AND Chrome:

curl -G "https://api-{region}.mida.so/v2/project/YOUR_PROJECT_KEY/experiment/1234/result" \
  -H "Authorization: Bearer YOUR_GENERATED_API_KEY" \
  --data-urlencode 'custom_filter=[[{"criteria":"country","operator":"==","value":"US"},{"criteria":"country","operator":"==","value":"CA"}],[{"criteria":"device","operator":"==","value":"desktop"}],[{"criteria":"browser","operator":"==","value":"chrome"}]]'

Breakdown by device:

curl "https://api-{region}.mida.so/v2/project/YOUR_PROJECT_KEY/experiment/1234/result?breakdown_by=device" \
  -H "Authorization: Bearer YOUR_GENERATED_API_KEY"

Post-deploy results (after serving a winner):

curl "https://api-{region}.mida.so/v2/project/YOUR_PROJECT_KEY/experiment/1234/result?report_phase=post_deploy" \
  -H "Authorization: Bearer YOUR_GENERATED_API_KEY"

Specific goal with event attribute filters:

curl -G "https://api-{region}.mida.so/v2/project/YOUR_PROJECT_KEY/experiment/1234/result" \
  -H "Authorization: Bearer YOUR_GENERATED_API_KEY" \
  --data-urlencode 'goal_key=purchase' \
  --data-urlencode 'event_attribute=[[{"key":"revenue","operand":">","value":100},{"key":"currency","operand":"=","value":"USD"}]]'

Success response

{
  "success": true,
  "test_id": 1234,
  "test_name": "Homepage CTA Button Color Test",
  "status": "live",
  "status_code": 1,
  "start_date": "2024-01-15T10:30:00.000Z",
  "end_date": null,
  "days_running": 15,
  "primary_goal": {
    "goal_name": "CTA Button Click",
    "goal_type": "clickOnElement",
    "goal_value": ".cta-button"
  },
  "stats_engine": "bayesian",
  "confidence_threshold_percent": 95,
  "serving_variant_id": "1",
  "is_serving": true,
  "report_phase": "pre_deploy",
  "traffic_allocation": 100,
  "distribution_mode": "equal",
  "is_mab": false,
  "total_visitors": 2980,
  "total_conversions": 268,
  "has_revenue": true,
  "total_revenue": 12450.00,
  "total_orders": 312,
  "rpv": 4.18,
  "aov": 39.90,
  "currency": "USD",
  "variants": [
    {
      "variant_id": "0",
      "name": "Control",
      "visitors": 1500,
      "conversions": 120,
      "conversion_rate": 8.0,
      "revenue": 5800.00,
      "orders": 145,
      "rpv": 3.87,
      "aov": 40.00,
      "currency": "USD",
      "traffic_weight": 50,
      "variant_status": "active",
      "is_control": true
    },
    {
      "variant_id": "1",
      "name": "Red CTA Button",
      "visitors": 1480,
      "conversions": 148,
      "conversion_rate": 10.0,
      "improvement": 25.0,
      "revenue": 6650.00,
      "orders": 167,
      "rpv": 4.49,
      "aov": 39.82,
      "currency": "USD",
      "revenue_per_visitor_improvement": 16.02,
      "traffic_weight": 50,
      "variant_status": "active",
      "is_control": false
    }
  ]
}

Response fields

Field	Type	Description
`test_id`	integer	Unique experiment ID
`test_name`	string	Experiment name
`status`	string	Human-readable label: `"draft"`, `"live"`, or `"inactive"`
`status_code`	integer	Numeric status: `9`=draft, `1`=live, `0`=inactive
`start_date`	string	When the experiment was created (ISO 8601)
`end_date`	string \| null	The date the experiment was formally concluded, or `null`. See note below.
`days_running`	integer	Days since creation. If the experiment is concluded, counts to the `end_date`; otherwise counts to now.
`stats_engine`	string	Statistical framework configured for this test: `"bayesian"` or `"frequentist"`. Mirrors the dashboard's per-test stats mode. Consumers (AI agents, dashboards, custom integrations) should speak in the matching framework — chance-to-beat-Control / expected loss for Bayesian tests, p-values / confidence intervals for Frequentist tests.
`confidence_threshold_percent`	number	Per-test deploy threshold (e.g. `95`). For Bayesian tests this is the chance-to-beat-Control needed to call a winner; for Frequentist tests this is the significance level (winners require `p < 1 - threshold/100`).
`serving_variant_id`	string \| null	When set, the variant id served at 100% after Set winner & serve. Omitted when not serving.
`is_serving`	boolean	`true` when `serving_variant_id` is set on the experiment.
`report_phase`	string	Echoes the `report_phase` query used (`pre_deploy`, `post_deploy`, or `all_time`).
`primary_goal`	object \| null	Primary conversion goal definition, or `null` if none set
`traffic_allocation`	number	Percent of eligible site visitors included in the experiment (0–100)
`distribution_mode`	string	Configured split mode: `"equal"`, `"custom"`, or `"mab"`
`is_mab`	boolean	`true` when smart optimization (MAB) is enabled
`total_visitors`	integer	Total visitors across all variants
`total_conversions`	integer	Total conversions across all variants
`has_revenue`	boolean	`true` when revenue data is present for the selected goal (Purchase/order events with a revenue property). When `false`, revenue fields are omitted.
`total_revenue`	number	Currency-normalized total revenue across all variants. Present when `has_revenue` is `true`.
`total_orders`	integer	Total order count across all variants. Present when `has_revenue` is `true`.
`rpv`	number	Revenue per visitor (`total_revenue / total_visitors`). Present when `has_revenue` is `true`.
`aov`	number	Average order value (`total_revenue / total_orders`). Present when `has_revenue` is `true`.
`currency`	string	Project reporting currency code (e.g. `"USD"`). Present when `has_revenue` is `true`.
`variants`	array	Per-variant result objects (control listed first)
`breakdown_by`	string	Present when `breakdown_by` is requested
`breakdowns`	array	Per-segment totals and variant results, present when `breakdown_by` is requested

About end_date and is_completed

end_date is only populated when the experiment has been formally concluded from the Mida dashboard (marked as complete). Deactivating an experiment via the API (status: 0) stops data collection but does not set end_date — it remains null. The days_running calculation uses end_date if set, otherwise counts to the current time.

Variant fields

Field	Type	Description
`variant_id`	string	Variant identifier (`"0"` = control)
`name`	string	Variant display name
`visitors`	integer	Unique visitors assigned to this variant
`conversions`	integer	Number of visitors who converted
`conversion_rate`	number	Conversion rate as a percentage (e.g. `10.0` = 10%)
`improvement`	number	Relative lift vs the control in percent (e.g. `25.0` = +25%). Only present on non-control variants when the control has data.
`revenue`	number	Currency-normalized total revenue for this variant. Present when `has_revenue` is `true`.
`orders`	integer	Order count for this variant. Present when `has_revenue` is `true`.
`rpv`	number	Revenue per visitor (`revenue / visitors`). Present when `has_revenue` is `true`.
`aov`	number	Average order value (`revenue / orders`). Present when `has_revenue` is `true`.
`currency`	string	Project reporting currency code. Present when `has_revenue` is `true`.
`revenue_per_visitor_improvement`	number	Relative RPV lift vs control in percent. Only present on non-control variants when the control has RPV data.
`traffic_weight`	number \| null	Configured share of experiment traffic for this arm (0–100). `null` when `is_mab` is `true`.
`variant_status`	string	`"active"` or `"paused"` (configured weight is 0)
`is_control`	boolean	Whether this is the control variant

Breakdown fields

Each breakdown item includes criteria, totals, conversion rate, and variants with lift calculated against Control within that segment. When revenue data is present for that segment, breakdown items also include has_revenue, total_revenue, total_orders, rpv, aov, and currency.

Revenue metrics

Revenue fields use the same dashboard /abtest/conversion engine as the Mida results page RPV view. They appear automatically for revenue-bearing goals (e.g. a Purchase event goal) — no separate revenue-type goal is required. Prefer RPV over conversion rate when has_revenue is true and AOV differs across variants.

Error responses

Status	Meaning
`400`	Invalid `test_id`, malformed query parameters, unsupported filters, use of structured `filters`, or date parameters are malformed / out of order
`401`	Invalid or missing API key
`404`	Experiment not found or belongs to a different project

Serving a winner

To deploy a winning variant at 100% on the same experiment (without creating a separate deploy test), use Serve Experiment Winner. Then query with report_phase=post_deploy to read post-deploy stats, or report_phase=all_time for the full timeline.

Next step

Need only configured traffic splits without result stats? Use Get Experiment Distribution. Need secondary goal results? Use Get Experiment Metrics. Done reviewing results? You can deactivate the experiment (status: 0), serve the winner, or get a public share link to send to stakeholders.

Path parameters​

Query parameters​

Supported breakdowns​

Supported filters​

custom_filter criteria​

Event attribute filters​

Examples​

Success response​

Response fields​

Variant fields​

Breakdown fields​

Error responses​