Key takeaways
- A lookalike audience starts from a seed list, so the seed decides the quality.
- On ad platforms the matching happens inside the platform, out of your sight.
- In B2B, lookalike modelling usually means finding companies with the same traits as your best customers.
- Every lookalike list needs checking by hand and excluding existing customers.
How lookalike audiences work on ad platforms
Lookalike audiences on ad platforms work by taking a seed audience you supply, such as a customer list or the people who converted on your site, and finding other users whose profiles and behaviour resemble it. The platform does the matching with its own data and shows your ads to the result.
Most platforms let you trade size for similarity: a tight audience resembles the seed closely but is small, while a broad one reaches more people with a weaker match. Uploading a customer list needs a lawful basis and must meet the platform's terms.
Lookalike modelling for B2B accounts
Lookalike modelling in B2B finds companies, rather than individuals, that share the traits of your best customers. The traits are usually firmographic, such as industry, headcount, location and growth, and sometimes technographic, such as the tools a company uses.
The model can be a simple set of rules written from your ideal customer profile or a statistical model trained on past wins. Either way the output is a list of candidate accounts to review, not a list of buyers.
Lookalike audience vs ideal customer profile
An ideal customer profile is the written description of the companies you serve best, and a lookalike audience is a list of real people or companies built to match it or to match your customers directly. The profile explains why; the audience is the who.
The two keep each other honest. If a lookalike list keeps surfacing companies your team would never sell to, the seed or the profile needs work.
How to build a good seed list
A good seed list contains your best customers, not all of them. Choose accounts that bought for the reasons you want to repeat, stayed and got value, and leave out one-off deals, heavily discounted wins and customers who left.
- Use a consistent definition of best, such as retained and growing accounts.
- Include enough accounts for patterns to show, but never pad the list with weak fits.
- Refresh the seed as your product and market change.
- Exclude current customers and open deals from the resulting audience.
Common mistakes with lookalike targeting in B2B
Lookalike targeting in B2B usually fails through a seed of every customer, which teaches the model to find average accounts. Another is matching only on surface traits such as industry and size, missing the situation that actually made your customers buy.
Teams also forget to look at the results. Ten minutes reading the first page of a lookalike list will show whether it resembles your customers in ways that matter.
Where Pebbles Ai fits
- CapabilityProspecting & customer recordsDescribe your ideal customer and an assistant finds matching companies on LinkedIn by keyword, industry, headcount and HQ country or region, 25 by default and up to 50 a search. It then finds the right people at up to ten of them per run.
- FoundationCompany IntelCompany Intel records your company, offerings and audience, so the assistants start from what you sell and whom it is for.
Common questions
How large should a lookalike seed list be?
Large enough for shared traits to show, and small enough that every entry is a customer you would happily clone. Ad platforms state their own minimum sizes, so check theirs before uploading.
What is the difference between a lookalike audience and retargeting?
Retargeting reaches people who have already interacted with you, such as past site visitors. A lookalike audience reaches new people who resemble your customers but have not met you yet.
Can I upload a customer list to build a lookalike audience under GDPR?
It can be lawful when your privacy notice and lawful basis cover it and the platform's terms are met. Check with whoever handles data protection before you upload.
About this page. Written by the Pebbles Ai team for people who run sales, marketing and growth. Last updated .
