Custom Audience

Narrowing an audience with the conditions a platform offers only goes so far. Age and interests are values the platform inferred, so there's no way to tell how many of your own customers ended up inside the segment you thought you had narrowed.

Pagedeck discovered the limit in an unexpected place. In a sales meeting, contacts at companies already paying for the product mentioned that they kept seeing ads inviting them to sign up. The interest-based segment had paying customers sitting inside it. The list is something only you know, and the platform can't act on it until you tell it.

Upload the list in a form the platform can match, then either include or exclude it.

1. Decide the purpose first. Targeting the people on the list and excluding them are opposite settings. Paid customer lists usually go to exclusion; accounts sitting on a free plan usually go to inclusion.
2. Pull the list. Signup email addresses or phone numbers are the key, cut to purpose — accounts on a free plan for three months or more, accounts that paid and then cancelled.
3. Convert the values into something unreadable before uploading. Send a transformed version of each email address rather than the address itself, and all the platform can do is check it against its own equivalent values. What information goes where falls under marketing privacy rules.
4. Check the match rate. Upload 5,000 records, have the platform find 3,200 people, and the match rate is 64 percent. Products where customers sign up with work email see lower rates, because platform accounts are usually registered with personal addresses.
5. Too small an audience won't run. A list of a few hundred hits the platform's minimum, or it runs and burns out because the same people see the ad repeatedly.
6. Set a refresh cadence. A customer who paid yesterday and isn't on the exclusion list sees the signup ad again today. Weekly re-uploads are typical.

"Exclude paying customers. They're currently getting signup ads."

Budget is being spent on people who already pay.

"Uploaded 5,000 records and the match rate came back at 64 percent."

Reporting how many of the list the platform could find.

"This list is 300 people, so it can't run alone — seed only."

Too small to target directly, so it feeds a lookalike audience instead.

Ravi split the lists three ways and the setup got simpler: paying customers excluded from every campaign, accounts on a free plan for three months or more included in the conversion campaign, and cancelled accounts included with a message of their own. The exclusion alone freed 11 percent of that month's budget, which went to reaching people who hadn't seen the product yet.