Skip to main content
Fill in the workbook for each section below. Each section also has a concept review dropdown that explains the thinking behind it, so you can understand why a field exists before you commit to how you’ll fill it in. Examples below follow the Doughnut Labs brand: a SaaS company that sells Doughnut Technology across Meta, Google, and TikTok. The greyed E.g. text shows you what a filled-in row looks like, so delete it and replace it with your own.
A customer profile name is built by taking one value from each field in order and joining them with your separator. The first fields (customer profile type, funnel stage, the segment itself) apply to almost every customer profile. The later fields (lookalike percentage, geo, demographics, exclusion, platform) only get added when they apply to that customer profile. Lock the formatting rules once, fill in a value for every field you use, then assemble the name. Add new values as your library grows, but don’t rename or delete a value that live customer profiles already use, or the names in your ad account will stop matching this sheet.
1

Lock the formatting rules

Agree on the separator, casing, and abbreviation style in Formatting rules below. These are the same for every customer profile, so decide them once.
2

Fill in a value for every field you use

Work through the field dictionaries: customer profile type, funnel stage, segment, lookalike percentage, geo, demographics, inclusion or exclusion, and platform. Give each value a short form and a plain description.
3

Assemble the name

Concatenate one value from each field, in order, separated by your chosen character, skipping any field that customer profile doesn’t need.
This document is one level of the shared naming conventions framework. See also campaign, ad set / ad group / variant, and ad naming conventions.

​
Formatting rules

These rules apply to every customer profile name regardless of type. Decide each one once, write it down, and don’t mix schemes later. A filter for LAL will not find Lookalike, so a convention that drifts is worse than one that is plain but consistent.
Everything else in this document decides which words describe a customer profile. This section decides the mechanics that make those words sortable, filterable, and readable at a glance, and it is the part that quietly determines whether the convention survives once several people are building customer profiles in the same account.Two choices carry most of the weight. One is the separator: a single consistent character between fields is what lets you split a name back into its parts, whether you are reading it in the customer profile picker or pulling it into a spreadsheet to pivot on. The other is abbreviation discipline. The whole point of a naming convention is that a filter for one value returns every customer profile with that value, and that only works if the value is written the same way every time. The moment someone types Lookalike where the sheet says LAL, that customer profile drops out of the filter and effectively disappears. Picking one scheme and holding to it matters far more than which scheme you pick.

​
Core fields

These fields apply to almost every customer profile you build. Fill in a row for every value you use, give it a short form, and describe what it means so the next person reads the same customer profile the same way. When you assemble a name, you take one value from each field in order.

​
Customer profile type

The kind of customer profile, by how it was built. This is the field that tells someone at a glance whether they are looking at a warm retargeting pool, a modeled lookalike, or a cold interest-based customer profile, and it is the one field that applies to every customer profile you make.
Customer profile type is the field that makes your whole customer profile library readable before anyone opens a single one. It answers the first question anyone has when scanning a list, which is not “who exactly is in this” but “is this cold, warm, or modeled.” Because the answer changes how the customer profile should be used, budgeted, and measured, this field tends to sit at the front of the name where it is easiest to scan and filter on.The main tension is how finely to split it. Some teams keep it coarse, with a few buckets like custom, lookalike, and saved. Others separate a retargeting pool from other custom customer profiles, or a broad Advantage-style customer profile from a manually saved one, because those get different budgets and different creative. Neither is more correct. The split that works is the one that matches the real decisions your team makes, since a distinction you never act on just adds length to every name.

​
Funnel stage

Where this customer profile sits in the customer journey. Tells someone whether a customer profile is meant for cold awareness, mid-funnel consideration, or bottom-funnel conversion, which is usually the single biggest driver of how it gets used.
Funnel stage encodes how warm the customer profile is, which is the thing that most changes how you treat it. A top-of-funnel cold customer profile and a bottom-of-funnel retargeting pool want different budgets, different offers, and different creative, and they get measured against different benchmarks, since a cold customer profile that converts at 0.5 percent and a hot one that converts at 8 percent are both performing normally for what they are. Putting the stage in the name means you can pull every cold customer profile, or every conversion customer profile, in one filter and compare like with like.The judgment here is how many stages you actually run. The three-stage top, middle, and bottom model is the common default, but some teams collapse it to just cold and warm, and others add a separate retention or win-back stage for existing customers. The field works at any resolution. What breaks it is using the labels loosely, so that the same warm customer profile is tagged middle-of-funnel in one place and bottom in another, because then filtering by stage stops telling you anything reliable.

​
Segment

A short, human-readable label for who is actually in the customer profile. This is the part a person reads to know they have the right customer profile, rather than a similar-looking one.
The segment is the one field written for a human rather than a system. Every other field is a code that sorts and filters; this is the plain-language handle that tells a person “yes, this is the customer profile I meant” without opening it to check the definition. A customer profile named only with type and stage codes is findable but not recognizable, and the segment closes that gap.Two details do most of the work here. One is the recency window: site-visitors-30d and site-visitors-180d are completely different customer profiles that behave differently, so folding the window into the label keeps them from blurring together. The other is consistency of phrasing. If the same group is called cart-abandoners in one customer profile and abandoned-cart in another, the segment stops being predictable, and a predictable segment is what lets someone guess the name of a customer profile before they even find it. The tension throughout is length against clarity: too vague (visitors, list) and it tells you nothing, too long and it gets cut off in the customer profile dropdown where you actually pick it.

​
Conditional fields

Add these only to the customer profiles they apply to. Skip the field entirely for a customer profile that doesn’t need it, rather than inserting a placeholder like na.
Some fields only make sense for some customer profiles. A lookalike percentage belongs on a lookalike and nothing else. A geography or demographic belongs on a customer profile that is actually split that way. Fill in the ones your library needs, and leave the rest out of the names they don’t apply to.

​
Lookalike percentage

The size and closeness of a lookalike customer profile. Only for lookalikes, where the same seed produces very different customer profiles at 1 percent versus 10 percent and you need to tell them apart.
A lookalike is defined by two things: the seed it was modeled from and how wide you cast the net. The seed usually lives in the segment field (a 1 percent lookalike of purchasers is not the same customer profile as a 1 percent lookalike of site visitors). The percentage is what this field captures, and it matters because the same seed at 1 percent and at 10 percent are genuinely different customer profiles: one is small and tightly matched, the other is large and loose, and they rarely perform the same way.This field earns its place only on lookalikes, which is why it sits among the conditional fields rather than the core ones. Putting it on a custom or interest customer profile would be meaningless. Where it does apply, a consistent way of writing the value matters, since 1pct, 1%, and 1-percent are three different strings even though they mean the same size, and only the one that matches this sheet will come back when you filter for it.

​
Geography

The location a customer profile is limited to. Only for customer profiles that are actually split by region, so the versions of the same segment for different markets don’t collide.
When the same segment is run separately for different markets, the customer profiles are otherwise identical: same type, same funnel stage, same recency window. The geography code is often the only thing keeping the US and Canada versions of a customer profile from looking like duplicates. It also lets someone pull every customer profile for one market in a single filter, which is how a regional budget or a localized campaign gets audited.The field only belongs on customer profiles that are genuinely split by location. If an account only ever runs in one country, adding the country to every name just makes them longer without telling anyone anything, because a value that is always present stops signaling anything. Where geography does vary, the usual approach is a short standard region code, and as with every field the value is only as useful as its consistency: us, usa, and united-states will not filter as the same market.

​
Demographics

An age, gender, or other demographic split. Only for customer profiles you deliberately break out this way, so the segments you test against each other stay separate and comparable.
This field exists for the case where you deliberately split one segment into demographic slices to test them against each other, for example running the same interest customer profile separately for two age bands to see which converts. When you do that, the slices are identical on every other field, and the demographic code is what keeps them distinct and lets you compare them cleanly later.It belongs only on customer profiles that are actually broken out this way. Many customer profiles aren’t, and letting the platform’s algorithm handle age and gender is often the better choice, in which case this field should simply be absent rather than filled with an “all” placeholder. As with geography, the value is only useful if it is written consistently: an age band entered as 18-24 in one name and 18to24 in another will not group together when you try to compare the two.

​
Inclusion or exclusion

Whether a customer profile is being targeted or suppressed. Only worth marking when the same customer profile gets used both ways, so a targeting pool and its matching exclusion don’t get confused for each other.
Optional: many teams keep two clearly named customer profiles (a target and an exclusion) rather than tagging one. Add this field only if your naming would otherwise be ambiguous.
The same list of people can play two opposite roles. A customer list might be the customer profile you target for an upsell in one campaign and the customer profile you suppress from a new-customer acquisition campaign in another. Because the roles are opposite, mixing them up is costly: an exclusion list accidentally used as a target shows acquisition ads to people who already bought, and a target list used as an exclusion silently blocks the people you meant to reach. This field marks which job a customer profile is doing.It is optional because many teams avoid the ambiguity a different way, by keeping the target and the exclusion as two separate, clearly named customer profiles rather than one customer profile tagged with a role. That works just as well. The field is worth adding only when your names would otherwise leave someone guessing whether a customer profile is meant to be included or held out, which is the exact situation where a wrong guess wastes budget.

​
Platform

The ad platform the customer profile lives in. Only needed if you keep customer profiles for different platforms in one shared library or naming sheet, so a Meta customer profile and its Google equivalent stay distinct.
Every ad platform holds its own customer profiles, so within a single platform the name never needs to say which one it is. This field matters only when you track customer profiles across platforms in one place, like a shared planning sheet or a cross-channel library, where a Meta lookalike and its Google equivalent would otherwise look like the same customer profile and get confused in reporting.Because most people build and pick customer profiles inside one platform at a time, many teams never need this field at all, and adding it to every name just makes them longer. Where you do run the same customer profile logic across Meta, Google, and TikTok and compare them side by side, a short platform code keeps them separate. As always, the code only works if it is consistent, so meta, fb, and facebook should not all appear for the same platform.

​
Assemble the name

Take one value from each field, in order, joined by your field separator. Skip any conditional field a customer profile doesn’t use, rather than leaving a gap or a placeholder. Read left to right, the name should tell you the type, the stage, and exactly who is in the customer profile. A warm retargeting customer profile of recent cart abandoners uses only the core fields: CA_MOF_cart-abandoners-7d A cold lookalike needs its percentage, so that field is filled in: LAL_TOF_purchasers-180d_1pct A lookalike split by market and age adds geography and a demographic: LAL_TOF_hi-ltv-crm_1pct_us_25-34 A customer list used as a suppression customer profile marks the exclusion: CA_RET_purchasers-180d_EXCL An interest customer profile tracked in a cross-platform library adds the platform: SAV_TOF_int-doughnut-lovers_us_meta
Last modified on July 21, 2026