How To Segment Email Contacts By Tags, Lists And Custom Fields

Email segmentation starts with a simple question: what do we know about this person, and what does that allow us to send responsibly? Tags, lists and custom fields are three common ways of recording the answer. They are often used interchangeably, which leads to muddled data, accidental over-mailing and segments that nobody can confidently explain.

The useful distinction is this:

• Lists usually record membership or subscription status: for example, people who signed up for the monthly product newsletter, attended a webinar, or became customers.
• Tags are flexible labels, normally applied internally, that mark a source, status or operational fact.
• Custom fields store a specific value about a contact, such as country, preferred product category, renewal date or account type.

A segment is the audience created by filtering one or more of those data points, often alongside behavioural data such as a recent purchase or a click. Most email platforms use slightly different terms, but the underlying model is consistent. For example, Mailchimp describes tags as internal labels, fields as stored contact data and segments as filters built from data such as tags, signup source, location and activity. Its audience-management guidance is a helpful illustration of the distinction.

The aim is not to collect every imaginable attribute. It is to maintain enough accurate, permissioned data to decide who should receive a particular message, who should not, and why.

Start with the job each data type is meant to do

Before creating a new tag or field, decide whether the information is a category, a durable fact, a changing state, a permission or a one-off campaign cohort. That choice determines where it belongs.

Use this	Best for	Example	Common mistake
List	Membership, subscription route or a defined programme	“UK product updates” or “Webinar registrants – May 2026”	Creating a new list for every campaign, which can duplicate contacts and fragment unsubscribes
Tag	Flexible internal labels and event/source markers	source:webinar-q2, customer:enterprise, review-needed	Using tags to hold values that should be searchable or reportable fields
Custom field	A value that has a defined format and may change	country=GB, plan=professional, renewal_date=2026-10-31	Saving several unrelated facts in one free-text field
Segment	A live or fixed audience selected for a particular communication	Subscribed UK customers on annual plans renewing in the next 45 days	Treating a saved segment as the source of truth rather than the result of rules

Platform terminology can complicate matters. One system may call a list an “audience”; another may use “list” for a static collection and “segment” for a rule-based collection. HubSpot, for instance, distinguishes active segments, which update as records meet or cease to meet rules, from static segments, which are fixed at the time they are saved. Its documentation on active and static segments explains the practical difference.

So document the meaning in your own data model rather than assuming a label used by a platform will settle it.

Use lists for membership and subscription boundaries

A list is most valuable when it expresses a relationship that is meaningful on its own. A person may be a member of a customer community, a subscriber to a regional newsletter, a registrant for an event or an applicant for a partner programme. Those are legitimate list memberships because they tell you something about the person’s connection to that stream.

Lists also matter operationally. They can provide an audit trail for where consent was collected, control enrolment in an automated journey, or separate genuinely distinct brands and purposes. However, splitting a single marketing database into many near-identical lists creates work and risk:

• the same person may be stored more than once;
• one unsubscribe may not reliably suppress every duplicate;
• purchase and engagement history becomes harder to view as one record;
• frequency controls may fail because each list sends independently; and
• reporting can make reach and performance look better or worse than they really are.

Many platforms therefore encourage a primary audience with tags and segments for organisation. Mailchimp notes that contacts in separate audiences do not share data and may increase contact counts when duplicated. Read its guidance on organising audiences. The right architecture depends on the platform and business, but the underlying principle is sound: keep one contact record where the same person, brand purpose and suppression rules apply.

Do not use a list simply because a campaign needs a recipient selection. Create the segment from your main permissioned audience instead. Reserve a static list for a cohort that needs to remain fixed, such as attendees at a specific in-person event, a controlled test group, or a historical campaign population that must be reported on later.

Use tags for labels, not a second database

Tags are deliberately flexible. They work well when you need to annotate a contact with an internal fact that is not naturally one field with one value. A contact can carry multiple tags at once: source:partner-webinar, interest:analytics and customer:active, for example.

Good tag use is usually event- or process-oriented:

• source:conference-2026 identifies the acquisition route;
• content:pricing-guide records a known content request;
• migration:complete supports a temporary operational workflow;
• sales:do-not-enrol prevents a particular automation from starting; and
• review:permission isolates records needing verification before marketing resumes.

That flexibility can create tag sprawl. “VIP”, “Vip”, “VIP customer” and “high value” may all be applied to mean roughly the same thing. Over time, nobody knows which one is authoritative. Prevent this with a small tag taxonomy:

• Use lower-case, predictable names, ideally with a namespace such as source:, event:, interest: or workflow:.
• State who or what can apply each tag: form, integration, import or staff action.
• Give temporary tags an expiry or review date.
• Do not make a tag the sole record of consent, subscription status or an opt-out.

That last point is important. A manually removable tag such as newsletter-opt-in is a poor substitute for platform-level subscription status and an auditable consent record. Preferences and suppressions need to be enforced wherever an email might be sent, including automations, support tools and future platform migrations.

Use custom fields when the value needs structure

A custom field, sometimes called a profile property or merge field, holds a specific attribute. It should have a clear name, an expected format and an identifiable source. That structure makes filtering more dependable and can support personalisation in email content.

For example:

• preferred_topic: a controlled set of values such as security, operations or product;
• country: use an agreed country-code standard rather than a mixture of “UK”, “United Kingdom” and “Great Britain”;
• customer_since: a date field, not free text;
• plan_tier: a single current value such as starter, professional or enterprise;
• renewal_date: a date that can drive a timely renewal journey; and
• marketing_frequency: a preference such as weekly, monthly or product announcements only, where your platform can enforce it.

If a property can hold more than one value, be explicit about the data type. A multi-select interest field might be an array or a group membership in one platform, while another needs separate yes/no fields. Klaviyo’s guidance, for example, distinguishes text and list property types and specifies formats for imported date and list values. Its profile-properties reference shows why imports should be mapped and checked rather than treated as a simple spreadsheet upload.

Avoid turning one field into a dumping ground. A field called notes containing “Interested in security; met at Manchester event; likely annual plan; asked not to call” cannot be reliably segmented, may become stale, and may expose colleagues to more information than they need. Split genuinely useful, governed facts into structured fields. Keep free-text operational notes out of marketing selection wherever possible.

Build segments by combining evidence, exclusions and timing

The most useful segments rarely rely on one condition. They answer three questions:

1. Eligibility: who is permitted and appropriate to receive this?
2. Relevance: what known need, preference, lifecycle stage or action makes the message useful?
3. Exclusions: who should be held back because they have unsubscribed, recently received a similar message, converted, complained, bounced or entered an incompatible journey?

Consider a software company promoting an advanced reporting webinar. A reasonable segment might be:

Marketing-email subscribers in the UK or Ireland, with plan_tier=professional or plan_tier=enterprise, who have used reporting in the past 90 days or selected reporting as an interest, excluding people who registered already, cancelled their subscription, have an unresolved delivery issue, or received a marketing email in the last three days.

The list membership supplies permission and programme scope. Fields establish plan and geography. A behavioural condition or self-declared preference supplies relevance. Tags can mark the registration source or a temporary exclusion. The finished segment is substantially safer than “all customers”.

Use AND conditions to narrow to contacts who meet every necessary criterion. Use OR carefully where either of two signals genuinely qualifies somebody. Parentheses matter in complex segments. For example:

(plan_tier = professional OR plan_tier = enterprise)
AND marketing_status = subscribed
AND (interest contains reporting OR used_reporting_feature within 90 days)
AND NOT webinar_registered = true
AND NOT contacted_marketing within 3 days

Always preview the contact count and inspect a sample of records before sending. A segment that returns zero people is often a field-format problem. A segment that unexpectedly returns almost everyone is often an OR condition that is too broad, an unpopulated field being handled incorrectly, or a missing exclusion.

Separate preferences from inferred engagement

There is a major difference between what a person has told you and what you infer from their activity. A preference selected in a form, such as “monthly product updates”, should be retained and honoured as a preference. A tag based on opening three emails is an internal engagement signal, not consent to receive more frequent mail.

Use behavioural data cautiously. A recent purchase, event registration, click on a product category or logged-in account action can be useful relevance signals. But open data is not a complete measure of attention, and a lack of opens does not prove lack of interest. Make important audience decisions from multiple signals where possible: recent clicks, site or product activity, purchase history, explicit preferences and recency of permission.

For permission-based marketing, suppressions must override every other segment rule. The UK Information Commissioner’s Office explains that people have an absolute right to object to direct marketing and recommends retaining suppression details so organisations can avoid contacting them by mistake. Read the ICO’s guidance on respecting marketing preferences.

Be equally careful with profiling. Segmenting on purchase history or declared interests can make messages more useful, but organisations should be transparent about profiling, keep the data accurate and no more extensive than necessary, and consider the risk to individuals. The ICO’s guidance on collecting information and profiling sets out these expectations. Sensitive information, including health, financial or other special-category data, needs particular caution and should not be casually imported into a marketing platform.

Make frequency control part of segmentation

Relevant targeting does not excuse repeated sending. A contact can qualify for a newsletter, an event invite, a product launch and a renewal reminder on the same day. Without a cross-campaign frequency rule, each team may see a defensible segment while the recipient sees an unwelcome pile-up.

Build contact-pressure controls into the selection process. Depending on your system, that may mean excluding contacts sent any promotional campaign within the last 48 or 72 hours, limiting marketing sends per rolling week, or applying category-specific limits. Transactional messages, such as a receipt or password reset, should normally sit outside promotional frequency caps, but should still be correctly classified and sent only where necessary.

This is both a customer-experience and deliverability concern. Google’s sender guidance says bulk senders should avoid unwanted email, make promotional messages easy to unsubscribe from, monitor spam reports and keep user-reported spam rates below 0.1%, avoiding 0.3% or higher. Google’s sender guidelines FAQ also distinguishes promotional messages from transactional messages for one-click unsubscribe requirements.

When a sender is repairing a reputation problem, tighter engagement and contact-pressure segmentation is often central to recovery. Email Foundry’s work in that setting can include audience verification, suppression review and likely-human engagement filtering alongside infrastructure and provider-level analysis. The purpose is not to find a workaround for mailbox-provider protections; it is to reduce unwanted mail and rebuild sending from demonstrably permissioned, appropriate audiences.

Create a practical governance routine

Segmentation stays useful only when the underlying data is maintained. Give each field, list and tag an owner, a definition and a source. A lightweight data dictionary can be enough:

Data item	Meaning	Allowed values or format	Source	Review
preferred_topic	Topics the contact chose to receive	Controlled multi-select values	Preference centre	On each update
source:webinar-q2	Registered through the Q2 webinar form	Tag; automatically applied	Form integration	Archive after reporting period
renewal_date	Current contract renewal date	ISO date	Billing system	Daily sync
Customer newsletter list	Permissioned subscribers to customer marketing	Subscription status held by platform	Preference centre and approved imports	Continuous

Review saved segments before major sends, especially when a CRM migration, form change or new integration has occurred. Check that an import updated existing records rather than creating duplicates, that automation triggers will not fire unexpectedly, and that a changed field name has not silently broken a rule. Test with internal seed records that deliberately represent key cases: subscribed, unsubscribed, newly registered, existing customer, lapsed customer and recently contacted.

Finally, retire data that no longer has a defined use. Every extra field and tag makes the audience harder to understand. A smaller, well-governed set of permission, preference, lifecycle and relevance data is more valuable than a large catalogue of ambiguous labels.

The decision rule to remember

Use a list when you need to represent membership or a subscription boundary. Use a tag when you need a flexible internal label. Use a custom field when you need to store a defined, reusable value. Then use a segment to combine those facts with behaviour, timing and exclusions for a particular message.

When every data item has a clear job, targeting becomes easier to audit, recipients receive fewer irrelevant messages, and deliverability decisions rest on evidence rather than guesswork.

Frequently asked questions

Should I use tags or custom fields for customer interests?

Use a structured custom field or a preference/group feature when people select interests themselves and you need to honour those choices consistently. Use tags for internal labels, campaign source markers or temporary workflow states. The exact implementation depends on whether your platform supports multi-select fields cleanly.

Are lists and segments the same thing?

Usually no. A list commonly represents membership or subscription status, while a segment is a filtered selection. However, platform terminology varies: some systems call a fixed audience a list and a rule-based audience a segment. Define the purpose and update behaviour rather than relying on the label alone.

Should unsubscribed contacts be deleted from the database?

Normally, keep enough information in a suppression record to ensure they are not accidentally marketed to again. The ICO specifically recommends a suppression or do-not-contact list rather than simple deletion for this purpose. Your retention policy and legal obligations should determine what other information is retained and for how long.

How many tags should a contact have?

There is no useful universal number. The better test is whether each tag has a clear definition, reliable application method and active use in reporting, automation or segmentation. Remove temporary or duplicate tags once they no longer serve a purpose.

Sources and further reading

• Create a Mailchimp Audience — Mailchimp
• Create active or static segments — HubSpot
• Profile properties reference — Klaviyo
• Respect people’s preferences — Information Commissioner’s Office
• Collect information and generate leads — Information Commissioner’s Office
• Email sender guidelines FAQ — Google