How to Build Ad Audiences From Your Customers
How to build ad audiences from your customers: a seed list of best buyers, a warm list of visitors and an exclusion list of people already in your pipeline.
To build ad audiences from your customers, make three lists from your own data: a seed audience of your best actual customers, a warm audience of people who visited or started a form, and an exclusion audience of current customers and open opportunities. Together they usually beat the default interest targeting inside an ad manager.
Interest targeting is offered first in every tool and is usually the weakest choice for a small company. It describes people by what they are said to like. Your customers describe them by what they actually did, which is to pay you. This post shows how to turn that into audiences, and what has to be true of your records before you can.
How to build ad audiences from your customers: the three lists
Each list does a different job, and each one needs you to be able to produce a list of people on demand.
| Audience | Who is in it | What it is for |
|---|---|---|
| Seed | Your actual best customers, uploaded as a list | The platform builds a lookalike audience from it, finding more people like them |
| Warm | People who visited, engaged or started a form and did not finish | A reminder, at a fraction of the cost of reaching a stranger |
| Exclusion | Current customers and open opportunities | Stops you paying to show acquisition ads to people already in your pipeline |
You can start with one and add the others. But all three come from the same source, which is a clean record of who your people are and what they did.
The seed audience: quality beats quantity
A lookalike is only as good as its seed. The platform studies the people you give it and goes looking for more of them, so whatever pattern is in the list is the pattern it will copy.
That is why the seed should be your best customers rather than every address you have ever collected. A few hundred people who paid are worth more than thousands who once downloaded something. A list of everyone teaches the platform to find more people who download things and leave. It does that very efficiently, and you pay for each of them.
Deciding who counts as a best customer is your judgement to make. Use the same measure you used for what a customer is worth, so the people the platform learns from are the people the economics were built around. That arithmetic is explained in how to set an ad cost ceiling.
The warm audience: remind, do not recruit
Some of the cheapest attention you can buy is attention from people who have already met you. Visitors who did not finish a form, readers who engaged and then left, and people who looked at what you offer have all shown a little interest already.
A warm audience lets you remind them at a fraction of the cost of reaching a stranger. It is also the right place to test an offer before spending on cold audiences, because warm people forgive more and tell you sooner whether the offer is any good. The reasoning is set out in how to write a paid ads offer.
Treat warm audiences with some care, though. Reminders can be very efficient, and they can also take credit for sales that were going to happen anyway. The post on how to judge ads by deals not clicks explains how to check.
The exclusion audience: stop paying to talk to your own pipeline
This is the audience people forget, and it is the one that saves money immediately. Current customers and open opportunities do not need an acquisition ad. Showing them one wastes budget and can confuse people you are in the middle of a conversation with.
Excluding them also keeps your numbers honest. If existing customers click your ads, the platform reports a result, and your cost per result looks better than the truth. The people you were trying to reach were new, and they were not the ones who clicked.
You need a list of your customers on demand
All three audiences depend on one capability: being able to say who your customers are, in a list, whenever you need to. Many founders discover at this point that they cannot. Customers are scattered across invoices, a mailbox and memory, and there is no single place to ask the question.
If that is you, that is the first piece of work, and it is worth more than any campaign. Make a single record of each person with a clear status, where they came from and what they bought. Once that exists, the three audiences are an export away.
| Question you should be able to answer | Feeds which audience |
|---|---|
| Who has paid us? | Seed, and exclusion |
| Who has visited or started a form without finishing? | Warm |
| Who is in conversation with us right now? | Exclusion |
| Where did each person come from? | Checking that the seed reflects your real best customers |
Why interest lists lose
An interest list describes a category of people and leaves the rest to chance. It cannot know that the people who actually buy from you share some pattern, because it was not built from them. Your own customer list was.
That is the quiet advantage of paid acquisition for a small company. Every other channel reaches whoever happens to be there. Ads let you point at a group, and the most powerful group to point at is described by the people who already bought. The wider picture of how ads fit into a small company is in what is paid advertising for founders.
Interest targeting against your own lists
| Question | Interest targeting | Audiences from your customers |
|---|---|---|
| Where does it come from? | A list of interests offered by the ad manager | People who actually bought, visited or are in your pipeline |
| What does it teach the platform? | A broad category | The pattern in your real customers |
| Does it need work from you? | No, which is why it is the default | Yes, you must be able to list your customers |
| What is the risk? | Reaching people who only look like prospects | A seed that is too broad, which copies the wrong pattern |
The table shows why interest targeting is the default and why it is weaker. It asks nothing of you, and it gives the platform little to learn from. Your own data takes effort, and that effort is the advantage.
What to leave out of the seed
The discipline of a good seed is mostly subtraction. Every address you add that is not a good customer dilutes the pattern the platform is trying to copy. Be strict about what stays out.
- People who downloaded something once and never replied.
- Contacts collected for other reasons that never became customers.
- Anyone whose only action was to be on a list.
- Customers you would not want more of, if you can say so honestly.
The last item deserves thought. A paying customer who was costly to serve, or who left quickly, is still a customer, but if you would not choose to attract a hundred more like them, they do not belong in the seed. The seed is a description of who you want, not a record of who you got.
A smaller, cleaner seed will feel risky because it looks like less. It is not. It is a sharper instruction, and the platform is a fast student that learns exactly what the list shows it.
A routine for building the lists
- Produce a list of everyone who has paid you, with what they paid and when.
- Pick out your best customers, using the same measure of worth as your cost ceiling.
- Save that smaller list as the seed. Leave out addresses that simply downloaded something.
- Define the warm audience: visitors, engaged people and unfinished forms.
- Define the exclusion audience: current customers and open opportunities.
- Apply the exclusion to every acquisition campaign before it launches.
- Refresh the lists on a schedule, so the seed reflects the customers you have now.
The lists go stale. New customers should enter the seed and the exclusion, and people who have become customers should leave the warm audience, otherwise you will keep reminding people who have already bought.
Frequently asked questions
Why not use every contact as the seed?
Because a lookalike copies whatever pattern is in the seed. A list of everyone teaches the platform to find people who download things and leave. Use the people who paid.
Do I need a large list?
Quality matters more than size. A few hundred people who paid are worth more than thousands who once downloaded something.
What if I cannot produce a customer list?
Then building one is the first task. Without it you cannot make a seed, a warm audience or an exclusion, and it is worth more than any campaign.
Should I still ever use interest targeting?
It is the default and usually the weakest option for a small company. Start from your own data, and treat interests as something to test against it, not a replacement.
The full method, with the checklists, is in The Paid Ads Playbook, a free 31-page playbook.