Email Segmentation: Send Better Emails

Somewhere in most small businesses there is a list. It started as a spreadsheet, moved into a platform, and has been receiving the same newsletter since 2019. The open rate was respectable once. Now it drifts down a little every month and the owner has quietly decided that email is not what it used to be.

In 13 years of looking at these accounts, I have not once found that the channel stopped working. What I find is one audience being asked to care about everything: the customer who bought twice this year, the person who downloaded a guide in 2021 and never replied, and the supplier who should never have been on the list in the first place. All three get the same message, and two of them have no reason to open it.

This is for the owner or marketing coordinator with a list between a few hundred and a few thousand contacts, one template and a suspicion that the quiet decline is something you can actually fix. It is fixable, and most of the fix is deciding who gets what.

One List, One Message, Falling Opens

The number that quietly declines

Decline is rarely dramatic. A point or two a month, masked by a couple of good sends and a seasonal spike, until the number is half what it was three years ago. The usual cause is not deliverability. It is relevance. When everyone gets everything, the people who would have cared about one message a quarter are trained to ignore the other eleven.

You can also see it in the click data rather than the open data. Opens are increasingly unreliable because of how mail clients handle images, but clicks still tell you who was interested enough to move. If clicks are flat while the list is growing, the message is not landing with a large part of the audience.

Why segmentation is a discipline, not a tool

Every platform has a segmentation builder and almost nobody uses it beyond a date filter. The platform is not the obstacle. The obstacle is that segmentation requires you to accept that some contacts should not receive this email. Most businesses are emotionally attached to their full list, because it represents years of work, and sending to fewer people feels like wasting it.

Sending to fewer people is the point. A smaller send to the right group usually outperforms a large send to everyone, and it does it while protecting the reputation that keeps you out of the promotions tab.

What Segmentation Has to Earn

Better opens are the mechanism, not the goal

Segmentation earns its keep through relevance, and the relevance shows up first in opens and clicks. HubSpot's published statistics state that "Segmented emails drive 30% more opens and 50% more clickthroughs than unsegmented ones (HubSpot State of Marketing Report, 2023)." The same page also reports that "78% of marketers report that subscriber segmentation is the most effective strategy they use for email marketing campaigns (HubSpot State of Marketing Report, 2023)." Both figures are republished on HubSpot's marketing statistics page. Better opens are the mechanism, not the goal. What segmentation is for is enquiries and repeat business from people who are still listening.

Sending less to some people on purpose

The hardest part of the discipline is deciding that a contact does not get this send. The second hardest part is holding that decision when someone in the business asks why the list number looks smaller. The answer is that a list of 4,000 people who mostly ignore you is worth less than a list of 900 who read most of what you send, and that the second number grows revenue while the first grows only the monthly bill.

The Three or Four Segments Worth Building First

Engaged, occasional and never engaged

Start with how people behave, in three groups. Engaged means they opened or clicked in the last 90 days. Occasional means they have interacted in the last year but not recently. Never engaged means they have not opened anything meaningful since they joined. Three groups, one filter each, and you can build all of them this afternoon.

These three groups receive three different things. The engaged group gets your normal sends. The occasional group gets the strongest offers and the most useful content, because they still exist as a chance. The never engaged group stops receiving regular campaigns entirely, which is the least popular piece of advice I give and the most valuable one.

Customers versus enquirers who never bought

The fourth segment sits across the first three: customers, separated from people who enquired and never bought. A customer who has already paid you should not receive the same acquisition message as someone still deciding. Give customers a reason to come back, a reason to buy the next thing, and information that assumes familiarity with how you work. Give enquirers the proof and the objection handling they did not get the first time. Neither group is well served by a generic newsletter.

Keep It Simple Enough to Maintain

How many conditions you really need

Most small business segments need one or two conditions, not eight. Sending a campaign to people who clicked a link in the last 90 days is one condition. Sending to customers in Victoria who bought in the last year is two. The ceiling is higher than people expect, which means the risk is overbuilding rather than running out of room. Mailchimp's help documentation puts the limit plainly: "For anyone with a Free or Essentials marketing plan, each segment can include up to 10 conditions and use one and/or logic operator." You can check that on Mailchimp's guide to segments. Ten conditions is more than almost any small business will use well.

The point where complexity costs you the habit

The real test is whether you can explain the segment out loud without opening the platform. If you cannot, you will not maintain it. Segments that nobody maintains get switched off quietly within a few months, and then the list goes back to receiving everything. Build the version you will still be using in a year, not the version that looks clever the day you build it.

Segment on Behaviour, Not Demographic Guesses

What people clicked and what they bought

What someone clicked is the most reliable signal you have, and it costs nothing to collect. If a third of your engaged contacts clicked the link about commercial solar in the last six months, they have told you what they care about more accurately than any demographic field. Purchase history is the same story with money attached. Both are facts rather than assumptions, and both are already sitting in your platform.

Where stated preferences beat assumptions

Where you genuinely need to know something you cannot infer, ask. A one question preference link in a welcome email, or a simple interest picker on the signup form, is honest and it works. Guessing that a postcode means a lifestyle, or that a job title means a budget, is how lists end up full of segments nobody trusts. Ask for what you need, then act on the answer.

Behaviour based segments can also be predicted rather than hand built, and the platforms market this hard. Mailchimp, for example, states on its segmentation page that you can "See up to 88% more revenue by targeting your most valuable customers with predicted segments." That figure is Mailchimp's claim about its own product rather than independent research, and I would treat it as a reason to test predicted segments on a slice of your list rather than a number to plan a year around.

The Hygiene Job Nobody Wants

Suppressing the never engaged

After a defined period, usually six to twelve months with no meaningful interaction, remove contacts from your regular sends and run a reactivation sequence instead: two or three emails asking directly whether they still want to hear from you, with an obvious way to say no. Anyone who does not respond comes off. Anyone who does moves into the engaged group and starts again.

Doing this feels like throwing away years of list building. It is the opposite. Unengaged recipients are the ones who report messages as spam, and, as Gmail's sender guidelines note, "Over time, user spam reports can lower your domain's reputation." That is a problem for every email you send afterwards, including the ones to your best customers. The full guidance sits on Google's page about email sender guidelines if you want the detail.

The spam rate line you cannot cross

The number to watch is your spam complaint rate, and the published threshold is low: "Keep spam rates reported in Postmaster Tools below 0.3%." On a list of 3,000, that is nine complaints before you are in territory that starts damaging the domain. Nine out of 3,000 sounds generous until you remember that complaints come disproportionately from people who never wanted your emails, which is exactly the group the suppression rule above removes.

Seven Ideas for the Next Three Sends

Send the engaged group a short piece of genuinely useful information with no offer attached. Send the customers a straightforward loyalty or repeat purchase message. Send the occasional group your single strongest offer, without the rest of the newsletter around it. Send the never engaged group the first reactivation email and nothing else. Send everyone in one state or city something local, if you serve more than one market. Send a follow-up to anyone who clicked the last send but did not convert. Then send nothing to the contacts who turned out not to want this, and note how much cleaner the next report looks.

Two of those sends go to subsets of the same people. That is fine. Sending three relevant emails to one group beats one irrelevant email to three groups, and the platform does not charge you more for it.

What to Measure Each Month

Check four things, once a month, in this order. The size of each segment, so you notice when one is growing because a filter stopped working. The open and click rate per segment, so you can see whether the message or the audience is the problem. Enquiries or revenue attributed to email, which is what decides whether the programme deserves more time. And the complaint rate, which is the number that protects everything else.

If the revenue side is unclear, that is a measurement problem worth solving rather than a reason to keep sending blind. Working out what email actually returns is the same exercise as working out what any channel returns, and we do that kind of review through ROI analysis work.

A One Hour Segmentation Plan for This Week

Half an hour to build the three behaviour groups and the customer split, using the open and click data you already have. Ten minutes to write the reactivation email for the never engaged group. Ten minutes to decide what the other groups will receive over the next three sends, in one sentence each. Ten minutes to book the review in next month's calendar, because a segment nobody checks stops being true after a quarter.

That is the whole job for a small list. If you would rather have someone set the structure up with you and then hand it over, an email marketing engagement that starts with the list you already have is usually a short piece of work rather than a retainer. The content side matters too, since segments only help if there is content worth sending to each group, and you can read common questions about engaging a consultant before you decide anything.

Email Segmentation FAQ

How many segments should a small business start with?

Three or four. Engaged, occasional, never engaged, and customers kept separate from people who enquired but never bought. Start with a small set of core segments and add complexity later, because a structure you can maintain beats an elaborate one you abandon in a month.

What is the difference between a segment and a list?

A list is everyone you may contact. A segment is the slice of that list you are sending to this time. You keep one audience and cut it many ways, which is why platform limits usually matter less than people expect: on Mailchimp's Free and Essentials plans a segment can carry up to 10 conditions with one and/or logic operator, which is more than most small businesses use.

Will segmenting actually improve results, or is it hype?

The mechanism is straightforward: different people get relevance, and relevance lifts opens and clicks. HubSpot's published statistics put segmented campaigns at 30% more opens and 50% more clickthroughs than unsegmented ones. The caveat is that segmentation only works if the message actually differs. Sending the same email to a new segment and calling it personalisation changes nothing.

What should I do with contacts who never open?

Suppress them from regular sends after a defined period, then run a reactivation sequence to see if anything lands. This is the least popular recommendation in email marketing and the most useful, because unengaged recipients are the ones who report messages as spam, and Gmail's own guidance is that spam reports lower your domain's reputation over time.

Is there a spam rate I need to stay under?

Yes. Google's published sender guidelines say to keep spam rates reported in Postmaster Tools below 0.3%. On a list of a few thousand that is a small number of complaints, which tells you how quickly a disengaged send can matter.

How often should I review my segments?

Once a month is enough for most small businesses. Check the size of each segment, its open and click rates, and whether anyone should have moved between groups since last month. A segment that never changes size is usually a condition that stopped being true.