Facebook Lookalike Audiences help advertisers reach new people who share characteristics with an existing group of customers, leads or engaged users.
The principle is straightforward: provide Meta with a useful source audience, and its systems identify other people who appear similar. However, modern Meta advertising is more automated than it once was. A lookalike may now operate as an audience suggestion rather than a strict targeting boundary, particularly when Advantage+ audience features are enabled.
Lookalikes can still be valuable, but they are not automatically better than broad targeting, interest targeting or Meta’s own predictive delivery. Their effectiveness depends on the quality of the source data, the campaign objective, the conversion signal, the creative and how performance is measured.
The strongest strategy is therefore not simply to create the largest possible lookalike. It is to give Meta a commercially meaningful signal—such as qualified leads, repeat customers or high-value purchasers—and test whether that signal improves profitable customer acquisition.
What Is a Facebook Lookalike Audience?
A Facebook Lookalike Audience is an audience of people Meta considers similar to an existing source audience.
The source audience could contain:
Existing customers
Qualified leads
Repeat purchasers
High-value customers
Website visitors
People who completed a particular website action
App users
People who engaged with a Facebook Page or Instagram account
People who watched a video
People who interacted with an Instant Form
People associated with offline or CRM events
Meta analyses signals associated with the source audience and identifies other users who appear similar. Advertisers are not shown the individual attributes or precise formula used to make these decisions.
This is more sophisticated than simply finding people of the same age, gender or parental status. Meta may use a much wider range of signals available within its advertising system, including behaviours, interests, content interactions and predicted likelihood of taking an action.
Lookalikes are prospecting audiences. They are designed to reach new people rather than simply advertise again to people already in the source. Meta states that people in the selected source are excluded from the resulting lookalike audience, although advertisers should still maintain appropriate customer and lead exclusions at campaign or ad-set level.
For advertising to people who have already visited, enquired or purchased, use a Custom Audience and remarketing strategy instead.
Lookalike Audiences, Custom Audiences and Advantage+ Audience
These audience types are related but serve different purposes.
Custom Audiences
A Custom Audience contains people with an existing connection to the business. It may be created from customer information, website activity, app activity, Meta engagement or another supported first-party source.
Custom Audiences can be used to:
Retarget known prospects
Re-engage previous customers
Exclude existing customers from acquisition campaigns
Separate funnel stages
Create a source for a Lookalike Audience
Lookalike Audiences
A Lookalike Audience uses a source to find new people who appear similar. The advertiser is no longer targeting only the known people in the source.
Advantage+ Audience
Advantage+ audience gives Meta greater freedom to find people likely to produce the campaign’s intended result. A lookalike, customer audience, age range or interest may be added as an audience suggestion, but Meta can deliver outside suggestions if its system predicts that doing so will improve performance.
Audience controls are different from suggestions. Controls such as permitted locations, minimum age and certain exclusions can restrict delivery, while suggestions help guide the system without necessarily creating a hard boundary. Meta explains this distinction in its guidance on Advantage+ audience controls and suggestions.
This means an advertiser selecting a lookalike may not always be running the tightly restricted lookalike test that older Facebook Ads guides describe. Check the audience configuration, Advantage+ settings and estimated delivery rather than relying solely on the name assigned to the ad set.
Are Lookalike Audiences Still Effective?
Lookalikes remain useful, particularly when the source represents a clear, valuable and sufficiently large group. However, they are no longer an essential component of every Facebook Ads account.
Meta’s delivery system already uses conversion history and predicted action rates to find potential converters. With enough reliable conversion data, broad targeting can sometimes equal or outperform a manually selected lookalike.
This is especially plausible when:
The product has widespread consumer appeal
The eligible geographical market is relatively small
Meta receives consistent purchase or qualified-lead events
The campaign has enough budget to learn
The creative clearly communicates who the offer is for
Advantage+ audience expansion is being used
Several audience segments would otherwise overlap
Lookalikes may be more useful when:
The business has a distinctive type of customer
The source contains verified purchasers or qualified leads
Customer value varies considerably
The advertiser is entering a larger or less familiar market
A new campaign has limited conversion history
Different services attract genuinely different customer profiles
The business has strong first-party data but relatively limited website signals
Neither approach should be assumed to win. Test broad, lookalike and other relevant targeting strategies under comparable conditions.
The Benefits of Lookalike Audiences
When implemented well, lookalikes can provide several advantages.
Use First-Party Data for Customer Acquisition
Lookalikes allow a business to use patterns found in its own customer or CRM data to help guide prospecting. This can be more commercially relevant than targeting generic interests loosely associated with the product.
Reach Beyond Existing Remarketing Lists
Remarketing is limited by the number of people who already know the business. Lookalikes extend first-party signals into a larger acquisition audience.
Focus on Valuable Customer Traits
A lookalike based on paying customers is usually a stronger commercial starting point than one based on every website visitor. A source built from repeat or high-value customers may be stronger again, provided it contains enough people to be useful.
Reduce Dependence on Visible Interests
Not every valuable buyer can be described accurately through the interests available in Ads Manager. Lookalikes can identify similarities that are not available as manual targeting options.
Support Market Expansion
Lookalikes can be useful when expanding into a new region, provided the original customers are relevant to that region and the business can serve it effectively.
However, similarity does not remove differences in language, income, competition, culture, regulation or buying behaviour. A successful UK customer source does not guarantee equivalent performance in another country.
How Meta Builds a Lookalike Audience
Meta requires a source containing at least 100 people from a single country and generally recommends a source of between 1,000 and 5,000 people. These are platform guidelines rather than guarantees of performance. The number of successfully matched and eligible users matters more than the number of rows originally uploaded.
Meta also allows advertisers to select a percentage representing audience similarity.
A 1% lookalike contains the people Meta considers most similar to the source within the chosen audience location. Increasing the percentage creates a larger audience but progressively reduces the required degree of similarity.
For example:
A 1% audience prioritises similarity and has the smallest reach.
A 1%–3% audience provides additional scale while remaining relatively close to the source.
A 3%–5% audience trades more similarity for greater reach.
Wider ranges can support further scaling but may begin to behave similarly to broad targeting.
A 1% audience is not automatically the best. A highly similar audience can be too small, more competitive or unable to absorb the available budget efficiently. A broader audience may provide Meta with more conversion opportunities.
Choosing the Best Lookalike Source
The source audience is the most important strategic decision. A lookalike can only learn from the group it is given.
Paying Customers
A customer audience is generally more valuable than an audience containing all website visitors or unqualified leads because it represents a confirmed commercial outcome.
For ecommerce, useful sources include:
All purchasers
Repeat customers
Customers above a minimum order value
Customers from a particular product category
Customers with a high predicted lifetime value
Customers who purchased recently
For lead generation, sources might include:
Sales-qualified leads
Booked appointments that attended
Accepted quotations
Sales opportunities
Closed customers
Customers above a minimum contract value
A lookalike based on every form submission may reproduce the characteristics of spam, job seekers, accidental enquiries and poor-fit prospects. This is one reason why improving Facebook Ads lead quality requires CRM feedback rather than relying only on the number of leads reported by Meta.
High-Value Customers
Value-based lookalikes allow customer values to be included with an eligible customer list. Meta can then use the relative values when finding similar people.
The value column must be meaningful and consistent. Depending on the business, this could represent:
Total customer revenue
Gross profit
Lifetime value
Subscription value
Average order value
A carefully designed customer score
Revenue is not always the best input. A high-revenue customer can still be unprofitable if fulfilment costs, refunds or sales effort are excessive. Where possible, use the value that most closely reflects the outcome the business wants to reproduce.
Meta provides specific guidance for creating a value-based Lookalike Audience.
Qualified Leads and CRM Stages
Lead-generation advertisers should distinguish between a submitted lead and a valuable lead.
A useful CRM sequence might include:
New lead
Valid lead
Qualified lead
Appointment attended
Sales opportunity
Closed customer
Revenue received
Separate Custom Audiences can be created from meaningful stages and used as lookalike sources where sufficient volume exists.
Connecting Meta with the CRM can also return downstream events for measurement and optimisation. This is different from merely uploading a customer list. Our Facebook Ads CRM integration guide explains the difference between lead retrieval, audience synchronisation, attribution and conversion-event feedback.
Website Events
Website audiences can be useful when customer volume is insufficient, but the selected action should represent genuine intent.
Possible sources include people who:
Purchased
Submitted a lead form
Reached a high-intent product or service page
Added a product to their basket
Initiated checkout
Used an important calculator or configurator
Spent meaningful time engaging with key content
An audience of all website visitors is easier to build but contains a wide mixture of customers, researchers, existing staff, suppliers, accidental visitors and low-intent users.
Reliable website audiences depend on an appropriately configured Meta Pixel, consent controls and, where appropriate, the Conversions API. Browser and server events must be deduplicated correctly to avoid inflating activity.
Facebook and Instagram Engagement
Engagement audiences can be created from activities such as video viewing, profile engagement, Page interaction or Instant Form engagement.
These sources are useful when a business has limited customer data, but engagement is not equivalent to buying intent. People may watch a video or interact with a post without being plausible customers.
Use stronger thresholds where available. For example, people who watched most of a product demonstration may represent a more relevant seed than everyone who viewed the opening seconds.
App and Offline Activity
Businesses with an app can create audiences from tracked app events. Offline or CRM activity can also support audiences where the necessary data sources and matching information have been configured.
An integrated dataset can combine relevant web, app, CRM or offline events for measurement and activation. Read our guide to Facebook Ads datasets for the wider tracking architecture.
How to Create a Facebook Lookalike Audience
Meta changes menu labels and campaign workflows regularly, but the typical process is:
Open Ads Manager or Meta Business Suite.
Navigate to Audiences.
Select Create Audience.
Choose Lookalike Audience.
Select an eligible source audience or data source.
Choose the audience location or locations.
Select the required percentage or percentage range.
Name the audience clearly.
Create the audience and allow Meta to populate it.
Select it within the relevant campaign or ad set.
Meta’s current setup requirements include an eligible source containing at least 100 people from one country. It generally recommends between 1,000 and 5,000 source members, although a smaller high-quality source can sometimes be more useful than a much larger low-quality one. See Meta’s current Lookalike Audience creation guidance.
Use a naming convention that makes the audience understandable without opening its settings. For example:
LAL 1% | Closed Customers | UK | 180DLAL 1–3% | Qualified Leads | UKLAL 1% | High-LTV Customers | UKLAL 3–5% | Purchasers | UK
Include the source, location, similarity range and relevant retention period. If the source is static, include the upload month in the internal name so that outdated lists can be identified.
A Practical Lookalike Testing Framework
Testing every possible source and percentage at once usually fragments the budget. Start with a small number of commercially distinct hypotheses.
Test 1: Broad Versus Lookalike
Compare:
Broad or Advantage+ audience
A 1% customer or qualified-lead lookalike
Keep the objective, conversion location, adverts, offer, attribution settings and geographical controls consistent.
Test 2: Source Quality
Compare audiences created from different funnel stages:
All leads
Qualified leads
Closed customers
This determines whether deeper commercial data produces better new-customer outcomes.
Test 3: Similarity Versus Scale
Where budget and market size justify it, compare:
1%
1%–3%
3%–5%
Use non-overlapping ranges if the purpose is to understand incremental performance. Otherwise, overlapping audiences make it harder to determine whether similarity level caused the difference.
Test 4: Standard Versus Value-Based
For businesses with significant variation in customer value, compare:
A standard customer lookalike
A value-based customer lookalike
The value-based version should be judged by customer value or profit—not merely its cost per conversion.
Measure Commercial Results, Not Just Meta Leads
A low cost per lead does not prove that a lookalike is successful.
For lead-generation campaigns, compare:
Cost per valid lead
Cost per qualified lead
Qualification rate
Appointment booking and attendance rate
Cost per sales opportunity
Customer acquisition cost
Revenue per lead
Closed revenue
Return on advertising spend
For ecommerce, compare:
New-customer acquisition cost
Purchase conversion rate
Average order value
Gross margin
Refund or cancellation rate
Repeat purchase rate
Customer lifetime value
New-customer ROAS
Meta’s attributed results should also be reconciled with CRM, ecommerce and finance data. Platform attribution is modelled and can claim conversions that another platform also reports. Treat Ads Manager as an optimisation and diagnostic system—not the company’s only financial record.
Our guide to Facebook Ads offline conversion tracking explains how qualified leads, sales opportunities, customers and revenue can be returned from the CRM.
Lookalike Audience Mistakes to Avoid
Using Every Lead as the Source
A large lead list may appear attractive, but it can include poor-fit prospects and invalid enquiries. Create the source from the deepest reliable stage with enough volume.
Choosing Size Over Quality
Meta’s source-size recommendation should not be interpreted as a reason to dilute a valuable audience. Five thousand mixed leads are not necessarily a better source than 1,000 confirmed customers.
Creating Too Many Small Ad Sets
Splitting limited spend across numerous lookalikes prevents each ad set from collecting enough conversion evidence. Consolidation often gives Meta more room to learn.
Assuming 1% Must Perform Best
Similarity is only one factor. Audience size, auction competition, budget and delivery flexibility also affect performance.
Layering Too Many Interests
Adding interests, behaviours and demographics on top of a lookalike can reduce reach and obscure what is being tested. It may also prevent Meta from finding converters who do not fit the advertiser’s assumptions.
Interest targeting can still be tested separately, as explained in our Facebook interest-targeting guide, but it should not automatically be stacked onto every lookalike.
Ignoring Advantage+ Expansion
If Advantage+ audience treats the lookalike as a suggestion, results cannot be interpreted as delivery exclusively to that lookalike. Review the active audience controls before drawing conclusions.
Failing to Exclude Existing Customers
The selected source is excluded from the lookalike itself, but it may not contain every existing customer. Maintain a more complete and regularly updated customer exclusion where acquisition reporting or offer eligibility requires it.
Using Stale Customer Data
Customer behaviour, product mix and business positioning can change. A static list created years ago may no longer represent the customers the business wants to acquire.
Where possible, synchronise audiences from a reliable CRM or refresh static sources on a documented schedule.
Comparing Audiences with Different Creative
If each audience receives different adverts, the test is measuring targeting and creative simultaneously. Use consistent creative for the initial audience comparison, then test audience-specific messaging separately.
Should You Combine Lookalikes with Demographic Targeting?
Demographic controls should be used when they reflect genuine commercial or legal requirements, such as:
A service available only within specific locations
A product subject to minimum-age restrictions
A language requirement
A service that is objectively relevant only to a defined group
They should not be added merely because the current customer list appears to contain more people from one demographic. Existing customer patterns can reflect previous advertising, pricing, store locations or creative choices rather than the true size of the wider opportunity.
Over-restricting a lookalike can remove potentially valuable customers and reduce Meta’s ability to optimise. Our Facebook demographic-targeting guide covers when demographic settings provide useful control and when they become unnecessary constraints.
Lookalikes are also limited or unavailable for some advertising involving housing, employment and financial products or services. Check the applicable Meta requirements before planning the audience strategy.
Data Protection and Customer Lists
Customer information used for Custom Audiences and lookalikes remains personal data even when it is hashed for matching. Hashing is a security measure; it does not automatically anonymise the information.
UK advertisers should:
Establish an appropriate lawful basis
Explain the advertising use clearly in their privacy information
Honour direct-marketing objections and withdrawals where applicable
Avoid uploading data collected for an incompatible purpose
Use only the fields necessary for legitimate matching
Restrict access to customer audiences
Maintain appropriate retention and suppression processes
Review contracts and controller responsibilities
Obtain appropriate consent for non-essential tracking technologies where required
The ICO notes that organisations should be transparent when customer information is used for social-media matching and when it contributes to similar-audience targeting. Do not assume Meta’s acceptance of an uploaded list proves that the advertiser’s use is lawful.
How Lookalikes Fit into a Complete Facebook Ads Strategy
A mature audience strategy can include:
Broad or Advantage+ prospecting
Customer and qualified-lead lookalikes
Relevant interest tests
Website and engagement remarketing
CRM-stage retargeting
Customer exclusions
Existing-customer cross-sell or retention campaigns
CRM conversion feedback
Creative tailored to awareness and buying stage
Audience targeting is only one component. Strong creative, a credible offer, accurate tracking, suitable campaign objectives, effective landing pages and responsive sales follow-up can have a greater impact than changing from one lookalike percentage to another.
Meta increasingly relies on automation, but automation still depends on the information supplied to it. Weak conversion tracking trains delivery towards weak outcomes. A campaign optimised for cheap form submissions may find more cheap form submitters; it will not necessarily find more profitable customers.
Final Thoughts
Facebook Lookalike Audiences remain a useful way to convert first-party customer knowledge into prospecting signals. Their strongest application is not finding people who merely resemble website visitors, but helping Meta find more people who resemble valuable commercial outcomes.
Start with the best source available. Prefer customers, qualified leads or meaningful values over superficial engagement. Keep the structure consolidated, compare lookalikes with broad targeting, and measure results through the CRM or ecommerce system.
Most importantly, remember that a lookalike is a hypothesis—not a guarantee. Meta’s systems decide how similarity is calculated, Advantage+ features may expand beyond the suggested audience, and platform-reported conversions do not automatically equal profitable growth.
The best audience is the one that produces the strongest verified business result after advertising cost, lead quality, sales performance and customer value are taken into account.
One PPC provides Facebook Ads management covering campaign strategy, audience testing, creative, Meta Pixel and Conversions API tracking, CRM integration, lead-quality optimisation and revenue reporting. Book a consultation if you would like us to review your audience sources and identify whether lookalike, broad or Advantage+ targeting offers the strongest testing opportunity.