What to do when your email list stops responding
If the list has gone quiet, pause the next broad send. Check whether messages can reach people, which contacts still have a reason to hear from you, and what happens after they act. The first repair depends on where that path breaks, so test each handoff before changing the copy.
Francisco Contreras · Founder, Machina
5 min read

Find where the response stopped
The list is large, the campaign shows as sent, and the replies have thinned out. Pause the next broad send. We compare the send report with what customers actually did, then repair the first broken handoff. When mail is rejected or filtered, we fix sending before testing another subject line.
Pull the last several comparable sends and record the intended audience, attempted sends, failed deliveries, complaints, unsubscribes, clicks, replies and the customer action you wanted. Compare the same kind of campaign over similar periods. If the list grew or the offer changed, mark that change before treating a rate as a trend.
| What you find | Check next | Decision |
|---|---|---|
| Failed deliveries or spam complaints rise across comparable sends | Open the provider’s rejection reasons and complaint report. Check the sending domain and where those addresses joined the list. | Repair sending and exclude invalid or unwilling recipients before another broad send. |
| Mail is accepted while clicks and replies fall | Compare recent clickers, repliers and customers with quiet contacts. Check what each group signed up to receive. | Test one relevant message with the active group. Give eligible quiet contacts one reason to stay, then suppress nonresponders. |
| Clicks are recorded without completed forms or orders | Follow the email link on a phone. Complete the form or purchase path and check what the customer record receives. | Repair the path after the click before increasing email volume. |
| A reminder arrives after a booking or accepted estimate | Trace the inquiry, estimate and job status that started the message. Check whether the completed action stopped the sequence. | Correct the event and stop rule before changing the copy. |
Machina diagnostic framework. The list and suppression decisions draw on the 3sixteen case. The event and stop decisions draw on the Burge Pest Control case.
Use each rate to choose a test, then confirm the cause with the underlying send or customer record. A send report alone cannot tell you where a message landed.
Make sure the message can arrive
We start with the provider’s identity-check status, rejection reasons and complaint report. We send from the same setup to mailboxes the team controls and inspect where the message lands, whether its links work and whether an unsubscribe takes effect. One test inbox has a narrow view, but it can expose a broken setup before the campaign changes.
For the sending domain, we check which services are authorized to send through SPF and whether DKIM gives the receiver a verifiable signature. Gmail requires at least one of these checks for mail to personal Gmail accounts. Senders of more than 5,000 messages a day to those accounts need both and a published DMARC policy. The visible From domain must align with either the SPF or DKIM domain, and marketing mail needs one-click unsubscribe. We inspect complaints and placement reports after these checks pass because authentication alone does not promise an inbox placement.
When we rebuilt email for 3sixteen, campaigns were going to all 290,000 contacts although about 40,000 were active. We fixed delivery first, then stopped sending every campaign to the full list. Mailing dead addresses put the next send at risk, so the audience rule belonged beside the sending repair.
Decide who still has a reason to hear from you
We look at what the person asked for, when they joined, what they bought or inquired about, and whether they clicked, replied or took another relevant action. Product arrival news belongs with people who asked for it. An open estimate calls for a different message. Write the audience rule before changing the template.
At 3sixteen, our active rule required an unsuppressed contact who had opened an email at least once in the last 90 days, placed no order in the last seven days, and had fewer than five lifetime bounces. That contact also had to have clicked an email in the last 90 days, been active on the site in the last 90 days, or placed an order in the last 180 days. An open alone did not qualify anyone. The separate win-back flow gave contacts with 120 days of no open and no click one invitation to click and stay. We suppressed those who did nothing.
We treat opens with care because Apple’s Mail Privacy Protection can download remote email content in the background even when a person does not engage. A click or reply is a more direct action. We keep orders separate from those message signals in the customer record.
Keep quiet contacts stored while you check their permission record. Send an eligible inactive group one useful invitation to continue, tied to what they originally requested. Make the choice to leave easy, then suppress those who do not answer. Set aside addresses whose signup source or expected topic you cannot establish.
We unsubscribe a test address, then check that the sending list excludes it from the next marketing send. For US commercial email, the FTC requires a clear opt-out path and honoring requests within ten business days. The test confirms that the request reaches the system that sends the mail.
Send when a customer action gives you a reason
Once delivery and eligibility are sound, we write the message around a recorded customer action. We identify what happened, who qualifies, which question the email answers, which action it invites and what stops the sequence. Then we test what happens after the customer acts early. A reminder that arrives after an estimate is accepted points to a faulty stop rule.
For Burge Pest Control, we separated a new inquiry from an estimate awaiting a decision. The new-inquiry sequence acknowledged the request and set expectations. Higher-value or unusually urgent opportunities generated owner alerts for a person to follow up.
A completed inspection or quote without a booking started the separate estimate follow-up. We stopped it when the customer booked, declined the quote or unsubscribed. That message addressed the questions a homeowner had before approving work, including timing, disruption and what would happen next.
A store can write to someone who asked when a product returns. A service firm can follow up on an estimate that remains open. We check the event record before building either message, then stop the sequence as soon as the customer takes the intended action. The email flows guide shows how to define those triggers and stop conditions.
Judge the repair by the action it was meant to support
Choose one segment and one message purpose for the next test. Record how many were eligible, attempted, delivered or rejected, then count clicks, replies, unsubscribes and the intended customer action. A customer reply can be a successful outcome for an estimate follow-up. If clicks rose and the form failed, repair the form next.
Compare the result with a similar earlier send and keep the differences visible. We compare email-attributed sales with the customer record to see what was purchased. The platform’s reporting window does not isolate sales caused by the message.
When the checks point to a sending or sequence problem, our email marketing work starts with a deliverability audit of the list and the flows already running. We fix inbox placement before raising volume, then build around the customer action that should prompt the next message. In the application to work with Machina, describe when response changed, who received the emails and what action the next message should support.
Keep reading
Related guides
Explore related work
Next steps
Let's talk about your project
Tell us what you're working on. We'll tell you if we can help — and exactly what we'd do.
