# Analyses Pilot et recherche

Source : https://pilot.braike.com/docs/regies/analyses-pilot

Les outils que Pilot ajoute aux régies : la recherche dans votre espace et les analyses calculées par Pilot.

En plus des outils des régies, Pilot MCP sert 2 outils de recherche et 5 analyses calculées par Pilot. Ces outils ne dépendent pas d'une régie en particulier : ils lisent ce que Pilot a déjà collecté pour votre espace, sauf mention contraire.

Pilot complet a d'autres analyses, que l'offre Pilot MCP ne sert pas : celles qui reposent sur des fonctions de Pilot complet (objectifs, étiquettes de campagnes, propositions, automatisations), et les diagnostics créatifs, qui interrogent une régie en direct.

<Callout title="Langue des descriptions">
Pilot décrit ses propres outils en anglais : c'est le texte que votre assistant reçoit. Chaque fiche commence par un résumé en français.
</Callout>

## Recherche dans votre espace [#recherche]

### `search` [#search]

**Search Campaigns** · lecture · rejouable

Retrouve dans votre espace ce qui correspond à une recherche libre, par exemple le nom d'une campagne.

Search the workspace for advertising campaigns matching a free-text query. Returns a list of records (id, title, url) the client can then fetch.

| Paramètre | Type | Requis | Description |
| --- | --- | --- | --- |
| `query` | `string` | oui | Free-text search query. Matches campaign names. |

### `fetch` [#fetch]

**Fetch Campaign** · lecture · rejouable

Rend le détail d'un résultat de `search`, à partir de son identifiant.

Fetch a complete advertising campaign by ID (as returned by search), with its recent performance metrics.

| Paramètre | Type | Requis | Description |
| --- | --- | --- | --- |
| `id` | `string` | oui | The namespaced campaign ID returned by search (e.g. "campaign:meta:123"). |

## Analyses de Pilot [#analyses]

### `pilot_data_freshness` [#pilot_data_freshness]

**Data Freshness** · lecture · rejouable · lit les données collectées par Pilot

La fraîcheur des données que Pilot a collectées, et l'origine d'un trou : une collecte en panne, ou un compte qui ne diffuse plus.

Freshness of the data Pilot has collected, per ad platform, and above all who a gap is due to: a collection failing on Pilot's side, or an account that stopped delivering. Each platform is judged on its weakest account: data covered through, last collection and its status, delay in days, and a cause: up to date (data at most 3 days old), late on Pilot's side (older data, and the last collection failed or is more than 2 days old), no activity on the account's side (older data, but a collection succeeded within 2 days) or unknown (no coverage date yet). Collection dates are returned because they are what tells these causes apart. Reads Pilot's database only, without calling the ad platforms.

| Paramètre | Type | Requis | Description |
| --- | --- | --- | --- |
| `platform` | `string` | non | Restrict to one ad platform ('google', 'meta', 'google\_analytics'...). |

### `pilot_cross_platform_overview` [#pilot_cross_platform_overview]

**Cross-Platform Overview** · lecture · rejouable · lit les données collectées par Pilot

Dépense et performance consolidées de toutes vos régies, ratios recalculés sur les totaux. Refuse un total en argent quand les devises diffèrent.

Spend and performance consolidated across the connected ad platforms over the last days (default 30, at most 365): impressions, clicks, conversions, spend and conversion value per platform and in total, with CTR, CPC, CPM, CPA, ROAS and conversion rate recomputed on the totals, never averaged across platforms. Refuses to return a money total when several currencies coexist or a currency is unknown, rather than return a wrong figure; impressions, clicks, conversions, CTR and conversion rate stay. Analytics and Search Console traffic is left out unless include\_organic is true. Reads Pilot's database only, without calling the ad platforms.

| Paramètre | Type | Requis | Description |
| --- | --- | --- | --- |
| `days` | `integer` | non | Window in days. Default 30. |
| `platform` | `string` | non | Restrict to one ad platform. |
| `include_organic` | `boolean` | non | Include Analytics and Search Console (organic traffic). Default: false. |

### `pilot_diagnose_wasted_spend` [#pilot_diagnose_wasted_spend]

**Wasted Spend** · lecture · rejouable · lit les données collectées par Pilot

La dépense sans conversion et les coûts par acquisition aberrants, au-delà de deux fois le coût médian du compte.

Campaigns that spend without converting, and those whose CPA is above a threshold, over the last days (default 30). Only campaigns that spent at least min\_spend (default 50) are judged. The default CPA threshold is the account's median CPA times cpa\_multiple (default 2): the median, so that an outlier campaign does not move the threshold used to judge it; max\_cpa sets it instead. A conversion that was not collected is never counted as no conversion: such campaigns are listed apart as undetermined, with the reason. Up to limit campaigns per list (default 25, at most 100). Reads Pilot's database only, without calling the ad platforms.

| Paramètre | Type | Requis | Description |
| --- | --- | --- | --- |
| `days` | `integer` | non |   |
| `platform` | `string` | non |   |
| `min_spend` | `number` | non | Minimum spend for a campaign to be judged. Default 50. |
| `max_cpa` | `number` | non | CPA threshold to apply. Otherwise computed from the median CPA. |
| `cpa_multiple` | `number` | non | Multiple of the median CPA. Default 2. |
| `limit` | `integer` | non |   |

### `pilot_budget_pacing` [#pilot_budget_pacing]

**Budget Pacing** · lecture · rejouable · lit les données collectées par Pilot

Le rythme de dépense : ce qui plafonne, ce qui dort et ce qui reste indéterminé, avec la raison.

Where the budget is capped and where it sits idle: compares each campaign's average daily spend over the last days (default 30) with its daily budget. A campaign spending seuil\_plafond of its budget or more (default 0.9) is capped, one spending seuil\_sous\_utilisation or less (default 0.5) is underused, and the idle daily budget of underused campaigns is totaled. Campaigns whose budget or spend was not collected, whose daily budget is zero, or that run on a lifetime budget, are set apart with the reason, never ranked as underused: an unknown budget is not a zero budget. Reads Pilot's database only, without calling the ad platforms.

| Paramètre | Type | Requis | Description |
| --- | --- | --- | --- |
| `days` | `integer` | non |   |
| `platform` | `string` | non |   |
| `seuil_plafond` | `number` | non | Share of the daily budget spent from which a campaign counts as capped. Default 0.9. |
| `seuil_sous_utilisation` | `number` | non | Share of the daily budget spent up to which a campaign counts as underused. Default 0.5. |

### `pilot_account_concentration` [#pilot_account_concentration]

**Account Concentration** · lecture · rejouable · lit les données collectées par Pilot

Combien de campagnes portent 80 % de la dépense.

How many campaigns carry 80% of the spend over the last days (default 30): tells whether the account rests on a handful of campaigns or spreads its spend, with the top campaigns by spend and their share (limit, default 10, at most 50). Amounts are summed as collected, without currency conversion. Reads Pilot's database only, without calling the ad platforms.

| Paramètre | Type | Requis | Description |
| --- | --- | --- | --- |
| `days` | `integer` | non |   |
| `platform` | `string` | non |   |
| `limit` | `integer` | non |   |