Consent source figures

The figures do not represent current consent status, unique-member counts or marketing acquisition attribution.

History events and daily summaries#

processAndSaveConsentChannelActivitySummary counts ConsentHistory rows whose createdAt falls within the reporting day. Rows are grouped by channel, consent ID and type within the tenant.

The writer uses explicit dataDate or yesterday, with configured timezone boundaries. It saves daily eventCount plus cumulative accepted and withdrawn fields. This report reads only eventCount.

Multiple changes or consent types for one member count separately. For example, one member accepting 2 consent types through My page contributes 2 accepted mypage events.

The collector returns early if any summary exists for the same tenant and dateOfRecord. No daily events means no daily source rows are saved; cumulative history alone does not create a row. Source: source-summary writer.

Query and selection rules#

historical-consent-channel-summary sums eventCount by channel and dateOfRecord period. It filters tenant, nondeleted summaries and the requested type exactly.

Accepted and withdrawn events are separate queries. Withdrawals remain positive counts here; Consent changes displays them negatively and counts updated records instead of history events. Source: historical query.

The UI offers currently active consent types. Supplied consentIds filter the query; missing or empty IDs default to active IDs.

If the active-ID list is empty, the repository adds no consent predicate. Historical inactive types can remain in the tenant's results. Sources: selection default, active picker.

What the channel identifies#

The graph displays channel strings without translation. Defined values include mypage, onboarding, newsletter, admin, mailchimp, make, m2m, mailmojo and import.

mypage identifies member self-service; admin identifies administrator actions. These are the channels supplied by those consent forms, not a member's original acquisition source. Sources: channel values, My page caller, admin caller.

Groups, totals and dates#

The service reserves one display slot for Others: view 5 selects 4 named channels; All selects at most 99. Selection uses the largest counts in the latest returned period.

Others appears only when that period exceeds the named-channel limit. Without it, older channels outside the selection disappear. Total sums returned datasets and cannot recover omitted channels. Sources: limit, ranking and Others.

Defaults are accepted status, weekly periods and view 5. Daily ranges end yesterday; coarser periods include the current unfinished bucket. Timeline windows are capped at 550 periods.

The earliest-date lookup uses active consents but does not filter accepted versus withdrawn. All time can therefore begin before the first event for the selected status. Source: timeline extent.

Diagnose a discrepancy#

• Source totals exceed members: compare events per member and consent; the report does not deduplicate people.
• Different values from Consent changes: compare history createdAt events here with current member-consent updatedAt rows there.
• An old source disappears: check latest-period ranking and whether Others is emitted before treating historical events as missing.
• Counts remain with no active types: inspect resolved consentIds; an empty array leaves the consent predicate absent.
• Blank graph: check the API response and values. Invalid responses render empty; zeros become null.
• Blank Total after hiding a source: restore the legend. Hidden returned groups still contribute to aggregation.

Tooltip shares use visible source counts, not all members or an acceptance-rate denominator. Sources: chart mapping, tooltip arithmetic.

The endpoint requires Dashboard read permission; menu access also depends on Analytics access and Insights. Source: endpoint gate.

Sources and verification#

Source checked 28 September 2026 at 6cd1850569be3b1702ca736a19320c7626efb7bc. Prior UI evidence confirms navigation, heading and accepted/default controls. No fresh UI, numeric, empty-selection, provider-attribution or ordinary-role test was performed.

For graph interpretation and controls, see Consent source figures.