The 180,000-contact database that only mails 13,000 people
Contoso Facilities migrated 180,000 contacts from a legacy CRM into Customer Insights - Journeys. The legacy system tracked a single 'DoNotEmail' checkbox. Marketing has since defined three purposes in real-time marketing: Commercial (newsletters and offers), Transactional (service updates) and Tracking (link and image tracking). Six weeks after go-live, 62% of contacts show consent as 'Not configured' and every commercial send skips them, so Finance sees roughly 13,000 sends from a 180,000-contact database. Separately, 1,400 people who unsubscribed through the preference center still received the March newsletter; on their contact records the legacy checkbox is set to Yes and a consent row exists. The migration also stamped an 'Allow' consent row for every contact where the checkbox was No. Marketing wants to set the Commercial purpose default to Allow to 'unlock' the 62%.
1. Explain why the 1,400 preference-center unsubscribers still received the March newsletter, and state the first three remediation actions in priority order.
Read the worked answer
Diagnosis (needs data, not a guess)
Consent is not a field on the contact. Real-time marketing evaluates contact-point consent rows keyed on contact point + purpose + topic + channel at send time. A missing row means not configured, and the compliance profile default for that purpose decides the outcome. That gives three competing explanations for the 1,400, and you confirm which one applies before fixing anything:
- Purpose/topic mismatch. The preference-center form collects consent for topics A and B; the March newsletter was bound to purpose Commercial + topic C (or to Transactional). The unsubscribe wrote rows the message never reads. Confirm by opening the March message and comparing its purpose/topic with the rows the form writes.
- Stamped 'Allow' rows from migration. The checkbox-to-consent mapping created an Allow row that the preference-center form later overwrote only for the topics it displays.
- Wrong classification. A newsletter sent under the Transactional purpose bypasses commercial consent entirely.
First three actions
- Suspend the next commercial send and audit one message's purpose/topic against the forms and preference center that feed it. Every message, form and journey must reference the same purpose/topic tuple.
- Bulk-write explicit deny rows for all purposes and topics for the 1,400, then reconcile the migration mapping so a legacy 'Yes' never silently becomes an Allow.
- Leave the 62% unconfigured and keep the Commercial default at Block. Unlocking them by default converts silence into complaints and domain reputation damage; the answer is a form-based re-permission campaign, not a default flip.
2. Marketing rejects the 'block by default' advice. Build the argument for the consent design you would actually ship, and say what happens to the 62%.
Read the worked answer
Design
- Purposes stay broad and few: Commercial, Transactional, Tracking. Purposes are legal categories, not marketing categories.
- Topics carry the granularity: Newsletter, Product offers, Events, under Commercial. The preference center surfaces topics; recipients rarely understand purposes.
- Every asset is bound explicitly: each email, form and journey message declares purpose + topic. An unbound asset is a defect, not a default.
- Preference center is the single opt-out surface and must display every topic you mail. A form that shows three of your seven topics is how opt-outs leak.
- Defaults: Commercial = Block when not configured; Transactional is typically configured to send and is used only for non-promotional content.
Why not flip the default to Allow
The 62% are not opt-outs, they are unknowns. Treating unknown as yes is what produced the complaint and bounce profile that gets a sending domain throttled. The cost of a re-permission campaign is a smaller list; the cost of a complaint spike is a smaller list and worse inbox placement for the contacts who did consent.
What happens to the 62%
They stay unconfigured and unmailed until they submit a form or update the preference center, which writes an Allow row. Segmentation should separate 'unconfigured' from 'denied' so you can target a re-permission journey at the unknowns only. Expect a low single-digit response and plan the next quarter's revenue on the consented base, not the database size.
3. You must hand the campaign team a segment called 'Engaged and consenting' for a quarterly send. Outline the logic and name one trap that will make the segment counts disagree with actual sends.
Read the worked answer
Logic
Real-time segment on Contact, joined to contact-point consent and to interaction history:
- Consent for purpose = Commercial, topic = Newsletter, channel = Email, status = Allow (require the row to exist; do not treat 'no row' as eligible).
- Email address present and not blocked/bounced.
- Engagement: at least one open or click in the last 180 days.
- Exclusions: suppression list members, and contacts with an open service complaint.
Trap
Consent is re-checked at send time, not at segment evaluation time. A contact can be in the segment when it refreshes, withdraw consent ten minutes later, and be skipped by the journey - so segment size will always exceed send volume by some margin. The mirror-image trap also bites: contacts who grant consent after the segment was last evaluated are not in it yet. Do not use segment count as the denominator for open rate; use actual sends from the email insights view. If the campaign needs a stable, auditable list, snapshot the segment members into a static list at a defined cut-off and report against that.