Cold Email Follow-Up, Timed by the Signal Not the Calendar

A cold email follow-up exists to give the recipient a reason to answer that the first message did not give them. Not a reminder that the first message happened — they know. If nothing has changed since you sent it, the follow-up has no new reason in it, and a fixed day-3, day-7 calendar guarantees that most of them will not.

The disclosure first. We build Keaz Signals, which sells ecommerce store data with live buying signals attached, and one of the things we sell is a reason to write on a particular day. We are a competitor to plenty of tools in this space, and we have an obvious interest in you concluding that a calendar is a weak trigger. Read the rest with that in mind.

What follows is a reading of Google's own Email sender guidelines, retrieved 13 September 2026, and of our own product behaviour. There is no benchmark study here and no reply-rate table, because we did not run one and we are not going to repeat someone else's.

A fixed cadence assumes that the passage of time is itself informative — that a store which ignored you on Monday is a better prospect on Thursday because three days went by. It is not. The only thing the calendar knows is how long you have been waiting, and that is a fact about you, not about the store.

This is why the second and third steps of most sequences read the way they do. "Just bumping this to the top of your inbox." "Wanted to make sure this did not get buried." "Any thoughts on the below?" Those are not messages. They are the sound a sequence makes when it has been asked to send something and has nothing to send. The writer knew it, which is why the sentence apologises for itself.

The assumption underneath is that non-response means the message was missed. Sometimes it was. More often the store read it, found it irrelevant to what they were dealing with that week, and moved on — and the fix for that is a different message at a different moment, not the same message again on a schedule.

Most follow-up rules that are not pure calendar rules fire on opens: follow up if they opened but did not reply, drop them if they never opened. Google's Email sender guidelines are direct about what that metric is worth, in a section headed Open rate.

Google doesn't track open rates. Google can't verify the accuracy of open rates reported by third parties. Low open rates aren't necessarily an accurate indicator of deliverability or spam classification issues.

Read what that does and does not say. It is not a claim that your sending tool is lying to you. It is Google stating that the number does not come from Google, that Google cannot check it, and that a low reading does not mean what senders usually take it to mean. An open is inferred from a tracking pixel loading, and a pixel loads or fails to load for reasons that have nothing to do with a human reading anything — image proxying, prefetching, a security scanner, a client that blocks remote content by default.

So a branch keyed on opens is a branch keyed on an unverifiable proxy. That is a structural point about the measurement, not an accusation against anyone who built a sequence on it. We have written separately about what email tracking tools measure and what they do not.

The most widely repeated cold email follow-up tactic is to reply into your own thread so the subject line carries a Re: prefix, making the message look like part of a conversation the recipient already joined. Google's sender guidelines address that directly, under message formatting:

don't send messages with subject lines starting with Re: or Fwd: unless the messages are actual replies or forwards

The same page goes further on display names, listing "using characters that imply the mail is part of threaded conversation" among deceptive practices, alongside using the recipient's own name as the display name. Both sit under a broader instruction that message headers "should be accurate, and not misleading or deceptive".

There is a real distinction to keep. Replying inside a thread you genuinely started, to someone you genuinely wrote to, is an actual reply — the Re: is true. Fabricating the appearance of a prior exchange is the thing named. The tactic gets recommended because it works on the eye for about a second, and that second is borrowed against the only asset the whole operation runs on, which is the recipient's willingness to keep opening mail from your domain.

That is a documented guideline from the receiving side, not a legal claim and not a compliance verdict on anyone's sequence.

The alternative trigger is the store itself. A follow-up has a reason when something about the store changed after your first message, and the change is observable from outside: ads that were not running are running now, the active ad count is climbing, a product went live, an email tool got installed, the storefront was rebuilt, the newsletter rhythm shifted. Those events carry a date, and the date is the part of the message that could not have been written last month.

This inverts the question the sequence asks. Instead of "has enough time passed to send step two", it asks "has this store done anything since I wrote". Most weeks, for most stores, the answer is no — and the right follow-up is then none at all, which a calendar can never conclude.

Our side of this is a pool and a schedule, not a clever rule. As of September 2026 we watch 4.1 million Shopify and WooCommerce stores across Europe and North America, re-scraped weekly, roughly 41,200 of them carrying a fresh signal in any seven-day window. Those figures move, which is why they are dated. The events are listed in the buying signal catalogue, and stores arrive with contacts in the ecommerce leads database.

A follow-up is an extra message to somebody who did not ask for the first one and did not answer it. That is not free, and the price is paid in a budget most sequences never look at. Google asks senders to keep the spam rate reported in Postmaster Tools below 0.10%, and to avoid ever reaching 0.30% or higher.

Those are small numbers. At a 0.30% ceiling, three complaints in a thousand delivered messages is the line you are told never to reach. A four-step sequence sends four times as many messages to the same non-responders as a one-step one, and every step is another chance for somebody to press the button rather than ignore it.

The same guidelines are blunt about the underlying practice — do not send to people who did not sign up, because they may mark it as spam and future messages to them will be marked as spam — and they tell senders to always give an easy way to unsubscribe, with one-click unsubscribe required above 5,000 messages a day. Reading those numbers on your own domain is covered in reading Google Postmaster Tools for a cold email domain.

A signal-timed follow-up names the change and stops. It does not reference the earlier message, does not ask whether it was received, and does not apologise for existing. It reads like somebody noticed something, because somebody did.

The shape is short: what changed, when, why that makes your offer relevant now, one ask. If the change cannot be tied to the offer in a sentence, it was not the right signal for you, and the honest move is to let it go rather than write around it.

Our AI sales agent writes that per store from the signal plus your own knowledge base, and sending runs through your own Instantly workspace, where replies sync back. We do not export: leads move into campaigns and stay there. Patterns per event type are in cold email templates by signal type, and the opening line in subject lines driven by a buying signal.

We are not going to tell you the optimal number of follow-ups. It is the number you would most want from an article with this title, every sending vendor publishes one, and they disagree with each other because each is measuring its own customers' offers on its own customers' lists. We did not run that study, so we do not have an answer, and quoting somebody else's would be borrowing authority we have not earned.

We are also not going to tell you the reply rate for signal-timed follow-ups, or how much better they perform than calendar-timed ones. That is the claim this entire article points toward, and we have not measured it. A reply rate is a property of your offer, your segment and your domains together, and a vendor controls at most one of those.

And nothing here is legal advice about what you may send to whom. We are GDPR-conscious by design and we will not claim more than that; responsibility for what you send stays with you. The Google guidelines quoted above are a receiving provider's operational requirements, which is a different thing from the law in your market.

A fixed cadence wins whenever there is no signal to wait for, and that covers more cases than we would like. If your prospects are not ecommerce stores, nothing we watch applies to you at all. If they are stores on a platform other than Shopify or WooCommerce, or outside Europe and North America, they are invisible to us whatever they are doing this week.

The rest of where we lose belongs in the same paragraph. We have no LinkedIn signals and no LinkedIn sending, and none are planned. Registry and firmographic triggers and hiring signals are marked planned, not shipped. We do not export, which for some teams is disqualifying and should be. And a signal-timed follow-up is slower by construction: it waits, and some weeks it sends nothing, which is difficult to defend to anyone whose target is messages out.

If you want to know whether the signals fire often enough in your segment to be worth reshaping a sequence around, that is a question a list answers better than an article does. Signups start with 1,000 free leads, and access is capped per market.

• Email sender guidelines, Gmail Help, Google. Every quoted line above — the open-rate passage, the Re: and Fwd: instruction, the threaded-display-name practice, the 0.10% and 0.30% spam-rate figures and the one-click unsubscribe threshold — comes from this page, retrieved 13 September 2026.
• Our own product behaviour as published on the buying signal catalogue, retrieved 13 September 2026.
• Own pool figures: 4.1 million Shopify and WooCommerce stores under watch across Europe and North America, re-scraped weekly, approximately 41,200 carrying a fresh signal in a seven-day window. As of September 2026. These figures move and are not permanent.
• No reply rates, no follow-up benchmark studies and no competitor figures are cited here. That is deliberate, and the reason is two sections above.