How to Reactivate Old Customers Without Damaging Email Deliverability
A brand with 600,000 historical email addresses does not necessarily have a 600,000-person marketing opportunity.
It may have:
- 80,000 valuable former customers,
- 150,000 weak one-time buyers,
- 40,000 people who unsubscribed,
- thousands of dead addresses,
- years of disengagement,
- and a sender reputation that could be damaged very quickly.
This is why I get nervous when someone says:
“We have a huge old list. Let’s wake it up.”
The size of the list is not the exciting part.
The quality of the audience is.
And the person who bought from you four months ago should not be sent alongside the address that has not opened, clicked or bought anything in four years.
Those are different commercial opportunities.
They are also different deliverability risks.
Inbox providers are paying attention to whether people actually want your mail
Google’s requirements for senders to Gmail now include authentication, valid DNS, TLS and low spam rates. For senders delivering more than 5,000 messages per day to Gmail accounts, Google requires SPF, DKIM and DMARC, along with one-click unsubscribe for promotional mail. (support.google.com)
Google recommends keeping user-reported spam below 0.1% and preventing it from reaching 0.3% or higher. (support.google.com)
Yahoo’s bulk-sender requirements similarly require SPF, DKIM, DMARC, easy unsubscribe and complaint rates below 0.3%. Yahoo also explicitly tells senders to monitor inactive recipients and remove invalid addresses. (senders.yahooinc.com)
Those numbers are small.
A 0.3% complaint rate means:
3 complaints per 1,000 inboxed messages.
So when someone proposes emailing 300,000 cold historical contacts “just to see what happens,” understand what they are risking.
Deliverability is not simply whether the email leaves your ESP
A lot of people confuse:
delivery
with:
deliverability.
If the message is accepted by the recipient’s mail server, it may count as delivered.
That does not mean it landed in the inbox.
Validity makes the distinction clearly: delivery rate measures whether mail was technically delivered, while deliverability is about whether it reaches the inbox rather than spam or another filtered location. (validity.com)
For customer reactivation, that matters because you are often dealing with people who have become less engaged over time.
Sending more of the wrong mail can make future mail harder to place.
Old customers are not the same thing as inactive subscribers
This distinction is important.
A previous buyer may have stopped purchasing but still:
- open emails,
- visit the site,
- browse products.
Another may have stopped engaging entirely.
Those are different risk profiles.
You should not define send safety from purchase lapse alone.
A useful reactivation decision combines:
commercial value
with:
mailbox engagement
and:
permission status.
Start with the strongest customer cohort, not the oldest one
If I were reactivating a large historical base, I would almost never start with the deepest-lapsed segment.
I would begin with customers who are:
- clearly beyond expected purchase timing,
- previously valuable,
- relatively recent,
- valid and appropriately marketable,
- showing some evidence of engagement if available.
The first wave should be the group with the best combination of:
recovery probability + commercial value + send safety.
Progressive rollout is your friend
Imagine you identify:
45,000 plausible recovery customers.
That does not mean you send to all 45,000 tomorrow.
A controlled rollout might look like:
Wave 1
2,500–5,000 highest-confidence customers
Measure:
- bounce behaviour,
- complaint rate,
- unsubscribe,
- click,
- purchase,
- revenue per recipient.
If healthy:
Wave 2
10,000–15,000
Then:
25,000+
Only if the audience and economics continue to support expansion.
This is not merely about deliverability.
It is also about learning.
I would much rather discover that the Return Reason is weak at 3,000 sends than at 300,000.
Authentication is now basic infrastructure
Before you worry about a clever win-back sequence, the sending domain should be technically sound.
For bulk senders, Google and Yahoo both now expect:
- SPF,
- DKIM,
- DMARC,
- compliant message formatting,
- easy unsubscribe.
Google also requires TLS and proper forward/reverse DNS. (senders.yahooinc.com)
This is not optional polish.
It is infrastructure.
Keep transactional mail protected
Yahoo explicitly recommends segregating marketing/bulk traffic from user, transactional and alert traffic by IP or DKIM domain. (senders.yahooinc.com)
That makes sense.
You do not want an aggressive reactivation test damaging the reputation behind:
- password resets,
- order confirmations,
- shipping notifications,
- receipts.
Transactional communication is mission-critical.
Marketing experimentation should not casually put it at risk.
Suppression is not lost revenue
Sometimes marketers resist suppressing inactive contacts because the list gets smaller.
That is the wrong way to think about it.
Validity recommends creating suppression lists for people who remain inactive so that ongoing low engagement does not damage sending reputation. (validity.com)
A person you should not be emailing is not an asset just because their address still exists in the database.
Database size is vanity if most of the file cannot be reached safely.
Define stop conditions before you send
This is one of the most practical things you can do.
Before Wave 1, agree:
What would make us stop?
Possible stop conditions include:
- complaint rate deterioration,
- unusual unsubscribe rate,
- bounce spike,
- sender reputation decline,
- weak conversion,
- poor revenue per recipient,
- margin economics not supporting expansion.
Do not wait until the send is live to decide what “too bad” means.
Spam complaints are especially dangerous
Google recommends staying below 0.1% user-reported spam and never reaching 0.3% or higher.
Yahoo says bulk senders should stay below 0.3%. (senders.yahooinc.com)
These are not targets.
They are ceilings.
You want to be comfortably below them.
That means a reactivation campaign generating acceptable sales but unusually high complaints may still be a bad campaign.
Be careful with long-lapsed addresses
The longer someone has been disengaged, the more questions I want answered.
- Is the address still valid?
- Does the person remember the brand?
- Are they still appropriately permissioned?
- Has there been any recent site or email activity?
- Is the relationship valuable enough to justify the risk?
Bluecore has case studies showing that long-lapsed customers can sometimes be commercially valuable.
That does not mean every old address should be reactivated.
The category, customer history and send health matter.
Do not overvalue open rates
Privacy features have made opens a noisier engagement signal.
More importantly, opens were never the objective.
For Revenue Recapture I care much more about:
- clicks,
- site activity,
- purchase,
- revenue per recipient,
- incremental lift,
- unsubscribe,
- complaint rate.
An open is useful context.
A purchase is behaviour.
A commercial-send-risk matrix
| Customer | Commercial opportunity | Send risk | Suggested action |
|---|---|---|---|
| Recent repeat lapse | High | Low | Test early |
| Recent one-time buyer | Medium | Low | Test selectively |
| Former VIP | High | Medium | Tailored treatment |
| Long-lapsed repeat buyer | Medium | Medium/High | Smaller controlled test |
| Deep-lapsed one-time buyer | Low | High | Deprioritize |
| Unsubscribed | N/A | High | Suppress |
| Hard bounce | N/A | High | Suppress |
This is a much more useful model than:
“How many old emails do we have?”
Case study lesson: scale does not require saturation
Omnisend’s recent Dukier case study reports that the brand grew heavily through automation while maintaining spam complaints below 0.01% and unsubscribes at 0.36% in the period described. The company segmented customers by type and market rather than treating the entire file identically. Those figures are vendor-reported, but the principle is worth noting: scaling lifecycle revenue does not require sending indiscriminately. (omnisend.com)
Relevance is a deliverability strategy.
The first reactivation message should make sense
A customer who has been gone for a long time may not need:
“WE MISS YOU!!! 25% OFF.”
They may need:
- what changed,
- what is new,
- why this is relevant,
- a product that matches their history.
The more the customer has forgotten you, the more the message needs a credible reason to exist.
Reconfirming interest can be better than endless sending
Yahoo recommends periodically considering reconfirmation for inactive subscribers, while Validity similarly recommends re-engagement campaigns followed by suppression when contacts remain unresponsive. (senders.yahooinc.com)
That is an important mindset.
The goal is not:
never lose an address.
It is:
maintain a healthy audience that wants to hear from you.
A safe first-wave checklist
Before a meaningful customer-reactivation send:
Infrastructure
- SPF working
- DKIM working
- DMARC published
- TLS enabled
- DNS correct
- unsubscribe working
Audience
- unsubscribed contacts removed
- invalids suppressed
- hard bounces suppressed
- audience permission reviewed
- strongest cohort selected
Commercial logic
- customer is genuinely lapsed
- relationship previously had value
- Return Reason makes sense
- offer does not unnecessarily destroy margin
Measurement
- purchase tracked
- complaints monitored
- unsubscribes monitored
- holdout considered
- stop conditions agreed
The bigger point
Do not treat email reputation like an unlimited resource.
It is an asset.
The customer database is also an asset.
A weak reactivation strategy can damage both at the same time.
So the goal is not to email as many old customers as possible.
It is to find the customers where:
the commercial opportunity is strong enough
and
the sending risk is sensible enough
to justify the test.
That is a much more disciplined way to recover revenue.
Related reading
How to Prioritize Lapsed Customers
How to Measure Customer Reactivation
Want to know which customers we would test first?
The Revenue Recapture Analysis starts with the strongest recovery opportunity—not the largest list.