Value-Based Lookalikes

Paid Media

Also: Value-Based Lookalike Audiences · VBL

What it targetsPeople who resemble your best customers
Input neededCustomer list with value data
Watch forSmall seed lists give weak matches
Judge againstStandard lookalikes, not just reach

Quick definition

Value-based lookalikes are an audience-targeting method where an ad platform builds a new audience by matching the traits of your highest-value existing customers, not just any customer. Instead of finding people similar to everyone who ever bought, it prioritises similarity to the ones who spent the most.

How it varies across Australia

Value-based lookalikes tend to outperform standard lookalikes on return on ad spend (ROAS) across most Australian ecommerce accounts, though the gap narrows once a customer file lacks enough high-value buyers to seed a strong match. Adoption still lags in smaller Australian advertisers who default to standard lookalikes out of habit.

See acquisition benchmarks across Australian industries

What it actually means

Imagine asking a friend to introduce you to more people at a party. A standard introduction just points you at anyone who showed up. A value-based introduction points you at the guests who actually buy rounds. Value-based lookalikes work the same way inside ad platforms like Meta and Google.

You upload a customer list with a value attached to each person, usually total spend or lifetime value (LTV). The platform's algorithm studies the traits of your highest spenders specifically, not your full customer base, and finds new people who share those traits. The resulting audience skews toward people who are statistically more likely to become high-value customers, not just any customer.

The catch is data quality. The model is only as good as the seed list. A file of two hundred customers with messy value data will produce a weaker lookalike than a clean file of two thousand. Most Australian advertisers running value-based lookalikes underestimate how much seed-list hygiene affects the output. Garbage value data in means a garbage audience out, dressed up in confident-looking targeting language.

A standard lookalike finds people who look like your customers. A value-based lookalike finds people who look like your best customers. That difference is the whole point.

How it shows up

Value-based lookalikes show up as an audience type inside Meta Ads Manager and Google Ads customer match campaigns. In reporting, they usually show up as a lift in average order value (AOV) or ROAS compared with standard lookalike audiences run in parallel, even when cost per click (CPC) looks similar or slightly higher. The signal to watch is quality of conversion, not volume of conversion.

The Australian context

Australian advertisers often run value-based lookalikes off customer lists that are too small to seed a strong model, simply because the local market is smaller than the US or UK. Meta generally recommends a seed audience in the thousands for the model to find reliable patterns. Smaller Australian ecommerce brands sometimes need to combine multiple value tiers or extend the lookback window on their customer export to hit a workable seed size.

Where people get this wrong

Uploading a seed list with unclean or outdated value data.The algorithm can only match on the values it's given. Refunds not deducted, old customers still counted, or currency errors all quietly poison the resulting audience.
Running a value-based lookalike from a tiny seed list.Small seed lists don't give the platform enough pattern to work with, so the audience ends up behaving like a standard lookalike anyway while looking more sophisticated in reporting.
Never comparing performance against a standard lookalike control.Without a side-by-side test, you're assuming the value-based version works better instead of confirming it. Sometimes a standard lookalike still wins on a thin customer file.

Value-Based Lookalikes vs Lookalike Audience

Value-Based LookalikesLookalike Audience
What it matches onTraits of your highest-value customersTraits of any customer in the seed list
Data requiredCustomer list with value attachedCustomer list, no value needed
Typical outcomeHigher average order value, similar reachLarger reach, more variable quality
Best used whenYou have a clean, sizeable value fileYou need scale fast and value data is thin

Related terms

Common questions

How big does my customer list need to be for value-based lookalikes?

Most platforms recommend a seed list in the low thousands for reliable pattern matching. Smaller Australian advertisers sometimes need to widen the lookback window or combine value tiers to reach a workable size. Below a few hundred, the model has too little signal to differentiate from a standard lookalike.

What value should I use to seed the audience?

Total historical spend or lifetime value (LTV) are the most common choices. Some advertisers use average order value (AOV) instead if they want to bias toward big single purchases rather than repeat buyers. Pick whichever value best reflects the customer behaviour you're trying to replicate.

Do value-based lookalikes always outperform standard lookalikes?

Not always. If your value data is thin, outdated or inconsistent, a value-based lookalike can perform the same as or worse than a standard one. It only earns its keep when the underlying customer file is clean and large enough to give the algorithm something real to learn from.

Can I use value-based lookalikes for lead generation, not just ecommerce?

Yes, provided you have a way to assign value to leads, such as deal size or conversion likelihood from your CRM. It requires exporting that value alongside the contact list, which most CRM and ad platform integrations support directly.

Debrief

Get the next one

No spam. No fluff. Just the next article, straight to your inbox.

Keep exploring

About New Rebellion

New Rebellion is a marketing intelligence consultancy. We build tools, score Australian businesses on how their marketing actually performs, and publish Debrief every day. This dictionary is part of how we work in the open.

How we think →