Open your CRM and export every contact you have not spoken to in six months or more. Sort it by the date of the last touch, oldest first.
Now look at what you are actually holding. Every contact on that export has had six months for their situation, their budget, their job title and their supplier arrangements to change while you were not watching.
That export is the cheapest lead source you own. What follows is the order I work through one.
Why the export exists at all
Before the sorting, the boring part that makes the sorting possible.
The reason you can do any of this is that somebody wrote the contact down. Use a CRM so leads never get lost. Which CRM is close to irrelevant. The expensive one with the pipeline automations and the cheap one both do the job that matters here, which is remembering a name, a date and what was said.
Companies that run outreach out of a sending tool alone lose this entirely. The sequence finishes, the contact drops out of the campaign, and eighteen months of conversations evaporate. Nobody can restart a pipeline from a record that was never kept.
If your answer to "export everything older than six months" is that you cannot, stop reading and fix that first. It is a one afternoon problem that compounds for years.
Row one to row whatever: the four groups
Every old lead falls into one of four groups, and they are worth wildly different amounts. Split the export into four tabs before you write anything.
Group one: never replied. The largest group and the least interesting. These are contacts who received your sequence and said nothing. You learned nothing about them, so there is no relationship to reopen. Handle them exactly as you would a list you have just built, starting with re-verifying every address before you send. Email addresses decay. People leave. Sending eighteen month old addresses at volume is one of the faster ways to spike a bounce rate and damage the domain you send from.
Group two: replied, then went quiet. Smaller and much more valuable. Somebody typed a sentence back to you. Even "not right now, thanks" is information and permission. These get a personal message from a human, never a sequence.
Group three: said "not now" with a reason. The best group in the file. A date, a budget cycle, a project finishing, a person joining. If your notes captured the reason, the reason is your opening line.
Group four: said no. Leave them. A real no is worth respecting, and re-emailing a hard no is how you collect spam complaints.
Why group three is the whole point
Stale leads get misread in one specific way. A lack of reply gets filed as rejection, when most of the time it was timing.
B2B decisions are slower than the people selling into them want to believe, and they have got slower. One 2026 benchmark roundup puts sales cycles 20 to 30% longer than in 2021, with enterprise deals that used to close in four to six months now routinely taking six to twelve.
Read that as a practitioner rather than as a statistic. A "not now" from one person in March was never a company decision. It was one member of a committee, reporting the state of a process that had not started yet.
Some markets are just slower than others. In some spaces nobody is making a fast decision about anything, and a pipeline that looks dead is a pipeline that is early. That is not optimism. It is the reason your six month old list has more live opportunities in it than your last two weeks of new sends.
What to actually send
Two rules, and then it is a normal conversation.
Rule one: do not restart the old sequence. Dropping a reactivated contact back into the same four emails they ignored last year tells them exactly how much attention they are getting. If you want the mechanics of a sequence built for cold contacts, they are set out in how a follow-up should be structured, which is a different job from this one.
Rule two: give the reason you are writing now. Not "just checking in". Something changed, or you think something changed, and you are asking. Their funding round. A new person in the role. The quarter they told you to come back in. The project they said would finish in spring.
Aim the message at a reply, not a booking. A conversation is what produces a call, and a call is what produces a client, in that order. Asking a lead who ignored you for a year to pick a slot on your calendar skips two steps and collects the same silence as last time.
If the contact went quiet after a live conversation rather than after a cold sequence, that is a different situation with a different fix, and I have written it up separately in what to do when a prospect goes silent.
The fastest way to burn a list you already own
The failure mode here is doing all of the above at once, on a thousand contacts, from your main sending domain.
Three guardrails. Verify every address you plan to send to before anything goes out, because an old list is a list with decay baked in. Send groups two and three by hand from a real mailbox, since they are small enough and valuable enough to deserve it. Put group one through your normal cold infrastructure at your normal volume, because as far as the receiving server is concerned, that is exactly what it is.
Reactivation goes wrong when a founder treats a whole stale database as warm. Parts of it are warm. Most of it is a cold list with a nostalgic label on it, and receiving servers will judge it on the data quality, not on the sentiment.
Do the export this week. The group three tab will be the shortest of the four, and I would bet a few of the dates those people gave you to come back have already gone past.