Reactivating Old Clients When Your CRM Has Been Asleep for Two Years

Reactivating old clients: where do you start?

Short answer: start by sorting, not writing. Isolate the contacts with no activity for six to twelve months, split them into three groups (former clients, lost prospects, never converted), write one message per group, then send to a small first batch that you follow by hand. Volume comes later, once the first batch has answered.

A CRM that has been asleep for two years nearly always holds the same things: a few hundred names, half-filled records, old quotes nobody followed up and clients who bought once and vanished. Nobody dares touch it, because the work looks huge and writing to people who did not ask you anything feels awkward. This article gives a concrete starting point, a four-step calendar, the mistakes that cost the most, and the share an AI can take. It does not replace the processes the studio automates: it covers the problem as you live it, before anything is automated.

What the published guides say, and what they leave out

We read several guides on reactivating dormant customers, including those from La Poste (business section), Cleverly and The Pedowitz Group. They agree on the order of the moves: define what a dormant contact is, segment, understand why the contact went quiet, write a fitted message, then measure. They differ on the numbers, which is telling.

One figure comes back in nearly every guide: reactivating a customer would cost five to seven times less than acquiring a new one. The ratio is quoted everywhere and sourced almost nowhere. Keep the meaning (a former client already knows your work) without turning it into a sales argument. None of these guides describes the case we care about either: a ten-person business with no marketing team, whose CRM is a spreadsheet or a half-filled tool.

Step 1: build a clean list before writing a word

A reactivation list is built from what you already know: the date of the last contact, the last invoice, the last quote, and the probable reason for the silence. Most of the work happens here, and it pays the most.

1. Gather the sources. CRM, old quotes, newsletter list, invoicing history, mailbox contacts. One client often appears three times under three spellings.
2. Remove duplicates. An AI does this well on names and addresses, provided a person checks a sample.
3. Date the last real exchange. Not the last email opened: the last quote, call or payment.
4. Set aside who must not be contacted. Clients in dispute, open unpaid invoices, people who asked not to be contacted, people who left their company.
5. Note the historical value. A former client at 12,000 francs a year does not get the same treatment as a prospect who asked for a price once.

La Poste offers a simple inactivity score, given as an example: a last purchase over twelve months ago is worth three points, no email opened in six months two points, no site visit in nine months one point, and six points make a contact very inactive. It is an illustration, not a standard. It shows that sorting on three criteria is enough to start: you do not need a statistical model to rank a few hundred records.

Step 2: three groups, three messages

A dormant contact is not a cold contact. The reason for the silence changes the message, and a single message for everyone reads like a circular.

• Former clients. They know you. The message recalls what you did together, says what has changed on your side since, and asks one simple question: does the need still exist?
• Lost prospects. A quote went out, nothing followed. The message picks up the context of the offer, offers to update it, and leaves a clean way out.
• Never converted. A contact met at a trade fair, a price request with no follow-up. The message is the shortest and the most careful, because the relationship never existed.

A useful message fits in five sentences: who you are to this contact, why you write now, one new and verifiable piece of information, one closed question, and an easy exit (“if this is no longer relevant, tell me and I will stop writing”). That last sentence matters more than expected: it reduces blunt unsubscribes and tells you which contacts to remove.

Step 3: a calendar in four steps

Here is a calendar we recommend to a small business with 200 to 2,000 contacts. It follows the ranges in the guides (three to five messages, five to seven days apart) and tightens them.

After this sequence, a contact who never reacted leaves the active list. La Poste suggests sending a last satisfaction survey before removing them. Pedowitz recommends making reactivation a recurring programme rather than a one-off: a small batch each month or quarter, as new contacts cross the dormancy threshold. That is also the setting where automation becomes useful.

What an AI can prepare, and what stays with you

A well-configured AI prepares the list, sorts contacts into groups, drafts one message per group from each contact’s history, and ranks replies by level of interest. It sends nothing without your approval. This is the “AI prepares, you approve” principle.

The main benefit is not saving an hour of writing. It is making the operation repeatable: the same sequence, at the same pace, without depending on a quiet week. It plugs into your existing tools, with no change of CRM or mailbox. To size the time currently spent on this kind of task, the express diagnostic in five questions (available in French on the studio site) gives an order of magnitude.

The mistakes that cost the most

• Writing to the whole base on the same day. A mass send after two years of silence damages your domain’s sender reputation, and you have learned nothing. The pilot batch exists for this.
• Opening with a discount. You devalue your offer and attract bargain hunters. An incentive arrives at the third message, not before.
• Ignoring the reason for the silence. A client who left over a quality problem does not answer “we are thinking of you”. If your history holds a complaint, the contact leaves the sequence and gets a call.
• Measuring the open rate. It tells you whether the subject line worked, not whether a deal is moving again. Count open conversations, reactivated quotes, and clients still active 90 days later, as Cleverly proposes.
• Handing sensitive data to a tool without checking where it goes. The content of old quotes and client details do not belong in just any service. Our article on the Swiss FADP and sending data to an LLM lists the points to check. It is a good-practice pointer, not legal advice.

When reactivation is not the right choice

It is better to say it: reactivating a dormant base does not always pay, and automating the operation pays even less often.

• Fewer than a hundred contacts. One person can call them all in a week. A spreadsheet and an afternoon are enough, with no automation.
• An unusable base. Without reliable last-exchange dates, the sequence will write to clients who signed last month. Clean first, as we explain in the article on AI automations that fail on real-world data.
• An offer that lost its reason to exist. If clients left because the price or scope no longer fit, following up fixes nothing. Call five former clients and ask why before sending anything.
• Very few, very important clients. A dormant large account is reactivated by a call from the owner, not by a prepared message.

Where to start this week

Export your CRM to a spreadsheet, add a “last real exchange” column, sort, and look at the thirty oldest high-value contacts. Write to ten of them, by hand, in five sentences. The replies will tell you whether the subject deserves a process, and which one. If it does, reactivating dormant contacts can then run on its own, under your control. Browse all our articles for related topics.