get_data_health

Read-only.

It answers “can I trust this number?” Call it before reporting a $0, a missing revenue figure or a partner count, so the user hears what heyBTW has not measured instead of a zero. Follows the heyBTW MCP rules.

When to use it

“Why does this event show no pipeline?”
“How many of our sign-ups can’t be matched to a company?”
“Which events have attribution and which don’t?”
“Is our partner-sourced count complete?”
“Before I send this report, what’s missing?”

Parameters

Parameter	Type	Required	Notes
`contract_version`	string	No	Omit for `v1`.

No other parameter. The response is one row, never dropped.

Response

data[0] itemizes coverage facts. It computes nothing of its own: each fact is another tool’s, and fact_sources names the source of each. It carries no composite, grade or tier.

Field	Meaning
`facts`	Opens the row. `unresolved_people` says in one `sentence` how many sign-up records have no resolvable company, with counts in `fields`, `via: heybtw` and a `basis`.
`next`	When the count is above zero, `next.doors` names two ways to close the gap, in order: add company and title from the user’s own tools and bring the rows back through `record_signups` or `import_people`, then enrich in heyBTW from the dashboard Enrichment page (`path`).
`attribution_coverage`	`events_with_canonical_attribution` and `events_without_canonical_attribution`, both named with `channel_id`. The same lists `get_event_attribution` names.
`unmatched_partner_source`	`signup_count`: sign-ups whose partner source matches no active partner.
`enrichment_coverage`	Sign-up records with and without a resolvable company. Counted per record (one per person per event), not per person.
`unmeasured_facts`	Facts heyBTW does not measure yet. Today: `provider_sync_freshness`, with its own gap.
`fact_sources`	The resolver behind each fact.
`upstream_coverage_notes`	Notes forwarded from the source tools, each with its `source_tool`.

The sample below is trimmed from a real response on the heyBTW workspace. Every sign-up there has a company, so facts says so and no next is served. The fact’s basis sentence is left out of the sample; it says the count is over sign-up records, one per person per event, not distinct people.

{
  "tool": "get_data_health",
  "result_count": 1,
  "data": [
    {
      "facts": [
        {
          "id": "unresolved_people",
          "sentence": "Every sign-up on this workspace has a company.",
          "fields": {
            "signup_records_without_resolvable_company": 0,
            "signup_records_total": 154
          },
          "via": "heybtw"
        }
      ],
      "attribution_coverage": {
        "source_tool": "get_event_attribution",
        "event_count": 4,
        "events_with_canonical_attribution": [
          { "event_name": "ScaleUp.ai Field Dinner: Austin", "channel_id": "cmvifxsyp000j07l2hmp06n1u" },
          { "event_name": "ScaleUp.ai x Partners: NYC Tech Week Open House", "channel_id": "cmvifxuom000o07l2hioks7wf" }
        ],
        "events_without_canonical_attribution": [],
        "events_with_invalid_canonical_projection": []
      },
      "unmatched_partner_source": { "source_tool": "get_event_attribution", "signup_count": 0 },
      "enrichment_coverage": {
        "signup_records_with_resolvable_company": 154,
        "signup_records_without_resolvable_company": 0,
        "signup_records_total": 154
      },
      "unmeasured_facts": ["provider_sync_freshness"]
    }
  ],
  "metadata": { "contract_version": "v1" },
  "status": "partial",
  "gaps": [
    {
      "code": "partial_coverage",
      "message": "Data-health fact not measured: provider_sync_freshness",
      "what_to_do": "Treat provider_sync_freshness as unknown, never as fresh or stale. Retrying will not change this result.",
      "retryable": false
    },
    {
      "code": "partial_coverage",
      "message": "Canonical attribution coverage is incomplete for this response",
      "what_failed": "Canonical attribution data is missing for 2 of 4 events that have not happened yet, so nothing is attributed to them until their event day - their sourced, influenced and won figures are 0 until then, and registered_account_pipeline is the open pipeline of the accounts registered for them, which is not attribution: ScaleUp.ai x Marlowe Cloud: LA Revenue Leaders Dinner, ScaleUp.ai x Partners: SF Tech Week Open House.",
      "retryable": false
    }
  ]
}

Rules the agent must follow

When unresolved_people is above zero, relay its sentence and offer both next.doors in their order before any CRM join. The join misses those people.
Treat provider_sync_freshness as unknown. Never call a sync fresh or stale.
An event with no canonical attribution has unknown revenue, not $0. Name it from attribution_coverage.
An upcoming event shows 0 attributed until its event day. That is the rule, not missing data. Its registered_account_pipeline (from get_event_attribution) is never attribution.
Counts here are sign-up records. Do not report them as distinct people.

Errors and gaps

Code	What it means	What to do
`partial_coverage` (unmeasured fact)	A fact with no live source, such as `provider_sync_freshness`. `retryable: false`.	Say it is not measured.
`partial_coverage` (attribution)	Forwarded verbatim from `get_event_attribution`, for example events that have not happened yet.	Relay it; scope revenue questions to covered events.
`partial_coverage` (Sponsor Mode)	Partner events restricted by the host’s Sponsor Mode are not counted.	Point to the dashboard’s partner view.
`scope_truncated`	More than 500 events, or a list too long for the response.	Say the lists are capped.
`internal_error`	Failure on heyBTW’s side.	Retry once.

get_event_attribution: the source of the attribution facts.
get_program_report: serves this row verbatim as its coverage block.
record_signups and import_people: the first door for unresolved people.
Playbook: Did anyone follow up, and what did the event do for pipeline.

Get started

How heyBTW MCP works

Playbooks

Tools: orientation

Tools: events and rosters

Tools: people and invites

Tools: after the event

Tools: accounts and partners

Tools: pipeline

get_data_health

When to use it

Parameters

Response

Rules the agent must follow

Errors and gaps

​When to use it

​Parameters

​Response

​Rules the agent must follow

​Errors and gaps

​Related

When to use it

Parameters

Response

Rules the agent must follow

Errors and gaps

Related