Est. reading time: 6 minutes

Open almost any Mailchimp account we inherit and the audience looks familiar. Too many tags stacked on too many contacts, a handful of segments built years ago for a campaign nobody remembers, half the tags overlapping, none of them meaning what the sender thinks they mean, and every campaign going out to “All Subscribers” anyway, because the targeting setup got too messy to trust. This is the most common email problem we see, and it has nothing to do with copy or design. It’s structural: tags and segments do different jobs, and when you treat them as interchangeable, your list slowly becomes unusable.

What tags actually are

A tag is a label you stick on a contact. It’s static, it sits there until you or an automation removes it, and it describes something a contact did, is, or was assigned. Tags don’t update on their own, and that’s the point: tag someone “Black Friday 2024 Buyer” and the tag stays in 2026 unless you take it off. Tags are sticky on purpose. They’re a record of state.

Good uses:

• Source of signup (Meta lead form, popup, checkout, in-store, wholesale application)
• One-time event participation (attended a webinar, claimed a specific promo, downloaded a guide)
• Customer status snapshots you want preserved historically
• Internal flags that don’t exist in your store data (VIP, gifted product, press list, do not discount)

What’s missing from that list matters as much as what’s on it. “Engaged in last 30 days,” “purchased in last 90 days,” “opened more than three campaigns”, those are not tag jobs. People try to solve them with automations that add and remove tags on a schedule, and it works for about six weeks, then the logic drifts and nobody trusts the labels anymore.

What segments actually are

A segment is a live query. You set the rules and Mailchimp rebuilds the audience every time you use it, nothing permanent gets added to the contact, and the segment is the question, not the answer. Segments handle everything dynamic: who clicked in the last 60 days, who engaged with the last three campaigns but hasn’t purchased, who lives in the EU and signed up after a given date, who has spent over $200 and gone quiet for 90 days. The conditions pull from tags, ecommerce data, signup date, location, campaign activity, contact rating, and combinations of all of it, which is where the real targeting happens. Segments reflect who your contacts are right now, not who they were when you first tagged them.

The mental model that fixes this

Tags are facts about a contact. Segments are questions you ask the list. If the answer should never change once it’s true, it’s probably a tag, “this person signed up at our trade show booth in March” will always be true. If the answer changes with time, behavior, or thresholds, it’s a segment, “this person is currently a lapsed buyer” depends entirely on when you ask. The test we run on every audit: if you’d have to manually update this label every week to keep it accurate, it shouldn’t be a tag.

The quick version:

• Use tags for: signup source, one-time actions, historical facts, lifecycle milestones, and internal notes.
• Use segments for: engagement windows, purchase recency, spend thresholds, location, campaign activity, and any behavior that changes over time.

Where most senders go wrong

Tagging engagement. An automation adds an “Engaged” tag on opens and removes it after 90 quiet days. Sounds reasonable, and in practice the rules drift, people get stuck in the wrong state, and the tag stops matching reality, with the extra problem that open-based rules are already unreliable under Apple’s Mail Privacy Protection. Engagement is a segment problem: build it on campaign activity in the last X days, weighted toward clicks, and skip the tag.

Tagging purchase behavior. If your store is connected, Mailchimp already knows who bought what and when, so tagging “Repeat Customer” or “High Spender” duplicates live data and instantly goes stale. Build the segment on purchase count, total spend, or last purchase date, and it updates itself, one of the payoffs of wiring the store connection properly in the first place, per our Mailchimp-Shopify integration guide.

Treating segments like saved lists. Some senders build a segment, send to it once, and never touch it again, the way they’d use a tag. That’s fine, but it misses the value: the segment built for your lapsed-90-day win-back is the segment to use every month, because the people inside it keep changing while the rule stays the same. The audience refreshes for free.

Stacking redundant tags. “Newsletter Subscriber,” “Email Subscriber,” “Signed Up via Website,” and “General List” all tagging roughly the same people, which happens when nobody owns the tagging convention and every new automation invents its own labels. Audit quarterly, merge duplicates, document what each surviving tag means.

How we set this up for clients

When we onboard a Mailchimp account, the first pass maps every existing tag into three buckets: keep, merge, or delete. Anything doing a segment’s job goes in the delete pile and gets rebuilt as a segment, anything redundant merges into a single canonical tag with a written definition. From there the tag system stays narrow on purpose, signup source, lifecycle milestones worth preserving, internal flags, and most accounts run cleanly on fewer than fifteen active tags.

Segments do the heavy lifting: engagement windows, purchase recency, AOV bands, geography, signup cohorts, and combinations. We name them by intent rather than audience size, so “Lapsed buyers 60-180 days, no win-back sent” tells the next person exactly what it’s for. The result is an audience you can actually target, and cleaner targeting means fewer low-intent sends and better engagement signals, which usually translates into stronger revenue per recipient. More importantly, you stop blasting the whole list, because you finally trust the smaller audiences, and the segmentation strategy those audiences enable is the whole engine behind keeping a list alive in the first place.

The shorter version

Tags record, segments query, and if you’re building automations to keep a tag accurate, you’re rebuilding what segments already do for free. Strip the audience back to tags that describe permanent or historical facts, move every behavioral and time-based question into segments, and the list gets simpler, the targeting gets sharper, and the data stops lying to you. This is the kind of cleanup that takes a few hours and makes every future campaign easier to target, easier to explain, and easier to trust, and most accounts we audit have been operating without it for years.