The Dormant List Audit SOP

Purpose

To determine what a dormant customer list actually contains before any outreach is sent to it. A dormant list is not one thing. It is a pile containing dead records, properties that changed hands, people who no longer need the service, customers who left angry, and customers who simply drifted, and each of those requires a different response or no response at all. Sending one message to the whole pile wastes the majority of the effort, irritates a minority, and produces a response rate so low that shops conclude reactivation does not work when what failed was the targeting.

This SOP produces a segmented, reachable list with a disposition on every record. It stops before the outreach itself.

Scope

Applies to every customer record classified Dormant or Cold in the quarterly list review, meaning more than two service intervals since last contact. Runs once a year, or once before any planned reactivation campaign, whichever comes first. Requires roughly one working day of records time plus a small sample of discovery calls.

Does not apply to the Slipping bucket, which is worked directly without an audit because those customers are recent enough that the records are still accurate.

Roles and responsibilities

Procedure

Phase A: reduce the raw list to what is reachable

Nothing in this phase requires a phone call. It is entirely a records exercise and it typically removes a large fraction of the list.

A1. Deduplicate. Same property under two spellings, same person under two phone numbers, a household entered once by each spouse. Merge, keeping the earliest first-job date and the most recent contact date. Duplicates inflate every downstream count and cause the same household to receive two messages, which is the most visible possible signal that you do not know who they are.

A2. Verify contact data. Remove records with no phone and no email and no mailing address. Flag rather than remove records with one channel only, because they are still reachable but the campaign design has to account for it.

A3. Confirm property status where you can. A property that has visibly changed hands or changed use is a different customer at the same address. Techs and public records are both usable here. Do not assume; a wrong assumption removes a live customer.

A4. Route open balances out. Any record with money owed goes to collections, not to outreach. A reactivation message to someone who owes you is read as a collections message with extra steps, and it damages both efforts.

A5. Apply do-not-contact. Anyone who asked not to be contacted, anyone the shop decided not to serve again, and anyone with a documented legal or safety incident. Each of these needs a written reason in the record, because a flag with no reason will be overridden by the next person working the list.

Phase B: segment the survivors on record evidence only

Read each surviving record for evidence of why the customer went quiet. Assign a segment only where the evidence exists; guessing here is worse than leaving it blank.

The unresolved-issue segment is the one to handle first even though it is the smallest, because those records represent a live reputational exposure that has been sitting untouched, and because they are the only segment where the shop can plausibly fix the cause.

Phase C: sample before you campaign

Do not extrapolate from assumption. Take a random sample of the no-evidence group, around 30 records, and make discovery calls with no offer and no pitch. The only question is what happened. Log the answer into the same four segments.

Thirty is chosen to be small enough to complete in a day and large enough to distinguish a segment that is a quarter of the list from one that is a twentieth. It cannot distinguish 20% from 25%, and you should not treat it as if it can.

The discovery call is deliberately offer-free. The moment it carries an offer, the answers change: customers who no longer need the service will not say so to someone who is selling, and the sample stops measuring what you need it to measure.

Phase D: assign dispositions and size the campaign

Every record now carries a segment. Assign each segment a disposition, a channel, and a priority, and size the total against the hours available. If the reachable, campaignable count exceeds capacity, cut by priority rather than sending a thinner message to everyone.

Phase E: write the dispositions back permanently

A "need ended" or "moved" disposition must be permanent, not just excluded from this campaign. The value of this audit compounds only if next year's audit starts from a list that has already been cleaned. Shops that run the audit as a campaign preparation and discard the findings will do the same work again in twelve months on the same records.

The worked example

A shop's dormant and cold buckets together contained 412 records.

Phase A removals came to 172 records: 61 duplicates or records with no usable contact information, 44 confirmed property sales or changes of occupant, 38 records that on reading were out-of-area or wrong-fit calls the shop had taken once and would not take again, and 29 already flagged do-not-serve or routed to collections. Those four figures sum to 172, and 412 minus 172 leaves 240 reachable records.

That is the first result and it is the one shops are least prepared for: 42% of the dormant list was never a campaign audience at all. A shop that had sent to all 412 would have measured its response rate against a denominator that was four tenths fiction, and concluded reactivation performs worse than it does.

Phase B found record evidence on 53 of the 240: 22 with a documented complaint, dispute, or unclosed callback, and 31 where the record showed the equipment had been replaced or the property's use had changed. That left 187 records with no evidence of why the customer went quiet.

Phase C sampled 30 of the 187. The results: 9 said the need had ended, 7 said they had used someone else, 2 raised an issue with the shop that was not documented anywhere in the file, and 12 gave no particular reason beyond having not thought about it. Those sum to 30.

Extrapolating those proportions across the 187: roughly 56 need-ended, 44 gone elsewhere, 12 with an undocumented issue, and 75 drift. Those four sum back to 187, which is worth checking because rounding four extrapolated figures independently will usually not close and you have to reconcile the remainder into the largest group rather than leaving the arithmetic open.

Totals across the 240 reachable records: 87 need-ended, 34 with an issue, 44 gone elsewhere, and 75 drift. Those sum to 240.

What the shop did with each. The 87 need-ended records were permanently marked for that service and excluded from every future reminder, removing more than a third of the reachable list from all future work. The 34 issue records went to the owner as individual calls, not a campaign; the important detail is that 19 of the 22 documented ones had a note in the file that nobody had read since it was written, some of them years old. The 44 gone-elsewhere records got the lowest priority and a short, specific message with no pretence that the shop did not know. The 75 drift records were the actual campaign audience.

The proportion that matters. Seventy-five campaignable records out of the 412 the shop started with is 18%. The audit's entire value is in that ratio: the same outreach hours applied to 75 records instead of 412 land on customers who have no reason not to come back, and the response rate measured against 75 tells you something true about your message. Measured against 412, it tells you about the list.

The undocumented issues. The two out of 30 in the sample who raised a problem the shop had no record of extrapolate to roughly 12 across the no-evidence group, which the shop had no way to find except by asking. That finding is the strongest single argument for the discovery-call phase, and it is also a finding about the intake process rather than about the dormant list: complaints were being resolved verbally and never written down.

Verification

Check that the removal reasons sum to the removal count. Every record that leaves the list must leave with a reason. Records that disappear without one are the shop's blind spot and they tend to be the oldest ones.

Check that Phase C carried no offer. Read the actual script used. If it contains a discount, a service reminder, or anything the customer could interpret as a sale, the segment proportions are contaminated and should not be used for sizing.

Check that dispositions were written back to the permanent record, not to a campaign spreadsheet. The test is simple: next year, does the audit start from 412 again? If it does, Phase E did not happen and the entire day's work has to be repeated.

Check that the issue segment actually got called by the owner. This is the segment that reliably gets deferred, because those calls are uncomfortable and produce no bookings. It is also the segment where an unaddressed complaint is still visible to the customer and, potentially, to anyone who asks them for a recommendation.

Check that the campaign was sized against the audited count. A campaign plan built on the raw list count before the audit will have set expectations for a list that does not exist.

References

• U.S. Federal Trade Commission, CAN-SPAM and TCPA requirements for commercial contact and honouring opt-out requests
• Trade-standard practice for customer-record hygiene and service-history documentation
• See related: The Quarterly Customer List Review SOP
• See related: How to Write a Win-Back Message That Gets Answered
• See related: The Lost-Customer Win-Back Campaign