Operated function · reactivation
Somebody who left because the product lacked one thing that now exists is a different prospect from somebody whose business changed, who is under contract elsewhere, or who had a bad experience nobody resolved. Mailed the same offer, four of those five learn that your prices are negotiable.
It is large, it grows on its own, and nobody owns it — which is precisely why it survives every planning cycle untouched. Working it well is a quarter of somebody’s attention with an uncertain payoff, and working it badly is one afternoon and a campaign send, so the badly version is what tends to happen.
The reason each person stopped is mostly unrecorded. Where a reason exists it was captured at the last touch, which means it is either "price" or blank, and neither is the actual cause. So the list gets segmented by recency and value, because those are the only fields that are reliably populated — and neither has much to do with whether somebody would come back.
Then the offer goes out. It produces a spike, which is reported as success. What it also produces is slower and never attributed: a share of the active base learns that leaving quietly for a while is how you get a better price, and the next renewal cycle is a little harder for reasons nobody connects to the campaign.
And a meaningful part of the list should not be contacted at all. People who left angry, people whose business closed, people who unsubscribed, people under contract with a competitor for another eighteen months. Mailing them costs deliverability, costs goodwill, and in some jurisdictions costs more than that.
Meanwhile the group that would genuinely return — because the specific thing that made them leave is fixed — is invisible inside the list, and gets the same generic message as everyone else.
Reactivation depends on whether the specific cause of departure has changed, and most businesses hold recency and value but not cause — so the list can only be segmented on fields that do not predict return.
This is why the discount is the default. When cause is unknown, price is the only lever that works on everybody a little, and a lever that works on everybody a little is exactly the lever that is most expensive over time.
Cause is more recoverable than it looks. Support history, the last conversations before silence, what the account did and stopped doing, whether a competitor was named, whether an issue was raised and never closed, whether the contact simply left the company — these already exist in your systems, unjoined, and they separate the list far better than recency does.
Once separated, most of it becomes a do-not-contact decision, which is a real outcome even though it produces no campaign. Suppressing the people who should not be reached protects deliverability and reputation, and it makes the remaining group small enough to approach properly rather than in bulk.
And for the group whose blocker is genuinely resolved, the honest message is not an offer. It is that the specific thing they left over is different now, said plainly, with no discount attached — which is both cheaper and considerably more likely to work.
How much of the dormant list has a known cause — measured by records with a reconstructed departure cause, against a starting point where the field was blank or said price.
How much of it should never be contacted — measured by the suppressed count and the reason for each, which is an outcome even though it produces no campaign.
Reactivation rate within the resolvable segment — measured by returns from the blocker-resolved segment specifically, rather than a blended rate across the whole list.
Whether returns needed a discount — measured by reactivations closed at standard terms, as a share of all reactivations.
Deliverability of your sending domain — measured by bounce and complaint rates before and after suppression, on the same infrastructure.
Whether the active base was affected — measured by discount requests and concession rates in the active book, watched across the campaign period rather than assumed unaffected.
that a large dormant list contains a large recoverable population. Frequently it does not, and the honest output of the first phase is a much smaller list plus a much larger suppression set. Nothing here decides pricing, approves an offer, or contacts anybody who has opted out.
Cause reconstruction is mostly a joining exercise across systems that already exist: the CRM, the support desk, the product’s own usage records, and the communication history. Nothing new is purchased and no third-party enrichment is required to do the first phase.
Suppression state is authoritative wherever you already keep it. An unsubscribe or a do-not-contact flag in your marketing platform is honoured as the source of truth rather than re-derived, because re-deriving consent is how a business ends up mailing somebody who opted out.
Outbound uses your own domain and sending infrastructure, so reputation effects land on the sender they belong to and remain visible to you.
Consent state is read from your system of record and is never inferred. A dormant contact is not a consenting one, and time passing does not restore a permission somebody withdrew.
Suppression runs before segmentation rather than after, so the people who should not be contacted are removed before anybody designs a message. Doing it in the other order is how a suppressed contact ends up in a test send.
Every contact carries a receipt naming the segment, the reason for inclusion, the message and the outcome — which is what answers a complaint, and in some jurisdictions is what a regulator will ask for.
Operational access is not permission to train. Your customer and departure data does not become material improving anything serving another organisation, including a competitor who would value a list of your lapsed customers more than you do.
Legal will ask about consent, and the answer is that consent is read rather than reconstructed, and that suppression is applied first. Marketing communication rules differ by jurisdiction and by how the contact was originally acquired, so both are established at scoping rather than assumed.
Whoever owns pricing should be in the room before any message is written. The position this page argues for — reach the resolvable segment without an offer — is a commercial decision, and it is worth making deliberately rather than discovering that a campaign shipped with a discount in it.
Where an obligation attaches through a jurisdiction or an acquisition source, it is marked applicability-gated rather than presented as standing.
The dormant population as you define it — read-only, nothing sent — segmented by reconstructed departure cause, with the suppression set identified first.
The first phase produces two numbers and no campaign: how much of the list should never be contacted, and how much of it left for a reason that has since been resolved. Both are useful and the first is usually larger than expected.
A common and legitimate outcome is that the resolvable segment is small. That is worth knowing before a quarter is committed to it, and it usually redirects attention to the active book, where the same analysis tends to be more valuable anyway.
If the resolvable segment is meaningful, the first outreach goes to it alone, about the specific blocker, without an offer, with responses routed to a person rather than into a sequence.
Then the useful question is what your last one did to the active base, because that cost is the one nobody measures. If it was a blanket offer to a list segmented by recency, the spike was real and so was the lesson it taught. The first phase here sends nothing and produces the cause breakdown, which lets you judge whether the campaign was reaching a resolvable group or buying back people who would have returned anyway.
Some usually is, and it is normally a much smaller fraction than the list size suggests — which is exactly why the size is misleading as a business case. The honest first output is a large suppression set and a small resolvable one. If that resolvable segment is too small to be worth working, that is a finding rather than a failure, and it cost a read-only exercise.
It is rarely in the field labelled "reason", and it is usually recoverable from things you do have: an unresolved ticket before silence, a contact whose email began bouncing, usage that narrowed to one feature, a competitor named in a conversation. That reconstruction is the first phase, and where a cause genuinely cannot be established the record is marked unknown rather than assigned a plausible one.
For most of the list, correct — which is why most of the list is not contacted. The segment this argues for is the group whose specific blocker is now resolved, and for them the message that works is that the thing they left over is different, said plainly. If your testing shows otherwise for your market, the offer decision belongs to whoever owns pricing, made deliberately, with the cost to the active base counted rather than assumed away.