LinkedIn predictive audiences: Requirements and setup
LinkedIn predictive audiences combine source data with LinkedIn's AI to reach people predicted to take similar actions. The source can include customers or conversions, making its definition a key decision. Official overview
This guide covers requirements and preparation before setup. It reflects the official help available in October 2026, rather than claiming the feature was newly announced.
Check requirements by source type
Choose one source type; multiple sources within that type can be used.
| Source | Requirement in the overview |
|---|---|
| Contact or company list | 300–300,000 rows |
| Conversions, Lead Gen Forms, or retargeting | More than 300 member accounts |
The creation guide also uses “300 or more.” Check the detailed requirements and your interface instead of depending on the boundary alone. Requirements, creation guide
Meeting a size threshold does not establish lead quality. Avoid combining unrelated products or purposes merely to increase the count.
Select the people you want more of
Use this internal decision table when selecting source data.
| Candidate source | Check first |
|---|---|
| Content downloads | Relevant industries and uses; information-only interest |
| Demo or consultation requests | Sales qualification and service eligibility |
| Customers who purchased | Product, contract period, and market bias |
| Website visitors | Recruitment, support, and existing customer activity |
For a new business service, recruitment visits and demo requests should not automatically represent the same desired action. This is a planning example, not a performance guarantee.
Record the source definition, period, count, and update owner. Explain your ideal prospect before asking AI to expand an audience.
Create the audience in Campaign Manager
Open the ad account, choose “Plan” → “Audiences,” then “Create audience” → “Predictive audience.” Set the name, source, location, and size before confirming creation. Official steps
Prepare this worksheet first.
Advertising objective and market:
Desired prospect definition:
Source type, conditions, period, and count:
Records to exclude from the source and why:
Predictive audience location and size:
Whether existing customers should receive ads:
Budget, creative, and destination:
Sales qualification after an inquiry:
Reviewer and review date:
After creation, check the current status and size before using the audience in an ad set. Avoid declaring failure from the initial size alone; verify readiness in your interface.
Check seed exclusions for acquisition
LinkedIn recommends excluding the seed contact or company list from the ad set if you do not want it included in targeting. Using a predictive audience does not automatically exclude existing customers. Seed list treatment
Define “new” consistently. A person's first inquiry and a company's first contract require different records. Reconcile platform exclusions with your sales definition.
Evaluate lead cost and sales quality separately
A lower cost per lead may not help if more inquiries are outside your service criteria.
| Measure | Question |
|---|---|
| Spend and leads | Are outcome definitions and periods consistent? |
| Qualified inquiries | Do industry, use, and eligibility match? |
| Sales opportunities | Are duplicates and existing customers treated consistently? |
| Response workload | Can sales handle the volume? |
Changing source data, creative, location, and budget simultaneously obscures the reason for a result. Record changes and review periods with consistent qualification. See incrementality measurement for a separate evaluation question.
Start with source and exclusion definitions. Shoryaku AI's chat can help choose evaluation criteria using summaries and business conditions you supply. It does not connect directly to LinkedIn Ads; manage settings and verify records in Campaign Manager and your own systems.