What is a lookalike audience?
A lookalike audience is a group of net-new prospects who share the same characteristics as your best customers, identified algorithmically from a seed list of existing customers. Ad platforms analyze behavioral and demographic signals, and B2B data providers analyze firmographic and technographic signals, then return new prospects who match those patterns.
After your marketing team creates ideal customer profiles, you feed first-party customer data into lookalike tools. The platform searches its users or its database and returns people or accounts that mirror your top performers. If your best customers are mid-market SaaS companies that closed in under 60 days, a lookalike audience finds more companies with that profile that you haven't reached yet.
How lookalike audiences work
Lookalike audience building starts with your seed audience: a selected list of customers with common traits (location, job title, buyer behavior). The platform's algorithm identifies patterns across that data, then finds new prospects who match those patterns.
Take these steps to find and create your lookalike audience:
Gather your first-party data through pixels, tracking, and form fills.
Analyze that data to find the traits your best customers share, and use those traits to select your seed audience.
Export your seed list in the file format your ad platform accepts, and keep a copy for future refreshes.
Upload the seed list to an ad platform or B2B data provider, which matches it against its own users or accounts and returns your lookalike audience.
The process involves three components:
Seed audience: The original customer or lead list you provide
Algorithm matching: Platform analyzes traits and behaviors
Lookalike output: New prospects who share characteristics with your best customers
Building your seed audience with high-intent segments
Not all customers make good seeds. B2B marketers need to be selective about which accounts go into their seed list.
The best seed audiences come from customers who fit your ICP and showed strong buying behavior. That means closed-won accounts, high-LTV customers, SQLs, and engaged prospects who converted quickly.
Start with these high-intent segments for your seed audience:
Closed-won customers from the past 12-24 months
High-LTV accounts showing expansion behavior
SQLs that converted quickly
Website visitors who requested demos or pricing
Choosing your lookalike audience size
Lookalike platforms let you control audience size, and that choice matters. Smaller audiences match your seed more closely. Larger audiences reach more people with a looser match.
The tradeoff is precision versus reach. If you need pipeline precision, go smaller. If you need top-of-funnel reach, go larger.
On Meta, you choose a percentage range when you create the audience. Meta says smaller percentages more closely match your source audience, while larger percentages create a bigger, broader audience. Start at the low end for pipeline campaigns where precision matters, and widen the range for awareness campaigns where reach is the goal. In Demand Gen campaigns, Google offers narrow, balanced, and broad Lookalike segments, which it aims to fill with 2.5%, 5%, and 10% of users in your target location.
Google is also moving the default for Lookalike segments in Demand Gen campaigns toward a suggestion mode. Advertisers who want to keep the original behavior, with the segment as a targeting constraint, can opt out.
Here is how to think about the size decision:
Smaller lookalike: Tighter match, higher relevance, smaller pool
Larger lookalike: Broader reach, lower precision, bigger pool
Let your campaign goal guide the choice.
How to create a lookalike audience in B2B
Lookalike audience building for B2B has its own constraints. B2B seed lists are often small, and match rates drop when CRM records are stale or incomplete. Here is a step-by-step creation workflow.
Step 1: Export your seed list from your CRM. Use your CRM's closed-won or high-value-customer filters to pull accounts from the past 12-24 months, including fast closers, and keep the firmographic fields intact (company name, domain, industry, employee count). Meta and Google match customer lists on contact details such as email and phone, so an account list needs buying-committee contacts attached before it works there. Meta needs a source of at least 100 people and recommends 1,000 to 5,000. Google needs more than 100 active matched people across your seed lists.
Step 2: Clean and enrich the seed data. Deduplicate records, validate business email domains, and confirm firmographic accuracy before uploading. Stale or inaccurate records degrade the lookalike model at the source.
Step 3: Upload to your ad platform. Meta Ads Manager, Google Ads, and LinkedIn Campaign Manager all accept hashed customer lists. Follow each platform's upload format requirements. LinkedIn no longer builds lookalike audiences from them, so on LinkedIn the list works as a source for the options covered below.
Step 4: Set your audience size. On Meta, start with a small percentage for pipeline campaigns and a larger one for awareness. On Google, choose narrow, balanced, or broad based on your campaign objective.
Step 5: Set suppression lists. Exclude current customers and recent churns before the campaign goes live. Running acquisition campaigns against your own customers wastes budget and creates a poor experience.
Step 6: Monitor and refresh your seed list quarterly. Google refreshes Lookalike segments every 1-2 days from the customer data you provide, so a lookalike is only as current as its seed list. New closed-won accounts, churned customers, and role changes in your seed list all affect the model's accuracy. Teams adding customers quickly may need to refresh monthly.
Attribution takes setup: connecting lookalike audience exposure to closed-won revenue requires closed-loop measurement between your ad platform, MAP, and CRM. Set that up before you judge a lookalike campaign on pipeline.
Lookalike audiences vs. custom audiences vs. retargeting
Custom audiences target people you already know: existing customers, website visitors, or email subscribers. Retargeting re-engages warm prospects who already interacted with you. Lookalike audiences find net-new prospects algorithmically similar to your best customers.
Here is when to use each:
Audience type | Source | Use case |
|---|---|---|
Custom audience | Your CRM, pixel, or list uploads | Target known contacts and accounts |
Retargeting | Website visitors, app users, ad engagers | Re-engage warm prospects |
Lookalike audience | Modeled from seed audience | Find net-new prospects similar to best customers |
LinkedIn groups list-based and retargeting audiences under Matched Audiences. LinkedIn discontinued its lookalike audiences on February 29, 2024. In their place it recommends Predictive audiences, built from sources such as contact lists, company lists, and conversions, and Audience Expansion for Matched Audiences.
B2B use cases for lookalike audiences
B2B teams use lookalike audiences at several points in the GTM motion.
Common use cases include:
ABM list expansion: Find accounts that match your Tier 1 target profile
New market entry: Identify prospects in verticals where you've had early wins
Demand gen scale: Grow your targetable audience beyond known contacts
Expansion lookalikes: Find new accounts similar to customers that expanded
Lookalike audience targeting helps you find more accounts that fit, whether you're expanding into adjacent verticals or scaling ABM programs beyond your existing database.
Layering intent signals for better timing
Lookalike audiences tell you who looks like your best customers. Intent signals tell you when they're actively researching. Combining both tells you which targets to prioritize.
ZoomInfo's intent data shows which companies are researching topics related to what you sell. Use it to rank lookalike accounts, so the ones that match your profile and are researching now get budget first.
Lookalike: Right profile (matches your ICP)
Intent: Right timing (actively researching)
Combined: Highest-priority targets (fit + in-market)
Why B2B marketers use lookalike audiences
Lookalike targeting addresses a core B2B problem: finding qualified accounts at scale. You use your customer data to find accounts that share the traits of the ones you've already won.
ZoomInfo is an all-in-one AI GTM Platform. The foundation is data at scale: 500M contacts, 100M companies, and 1.5B+ data points processed daily. The more complete the record behind each account, the more a lookalike model has to match on.
The reasoning layer on top of that data is the GTM Context Graph. It reasons across ZoomInfo data, your CRM, conversations, and buying signals to show why accounts are moving. For lookalike work, that can help separate the accounts that only resemble your customers from the ones also showing buying activity.
GTM Studio, currently in an early access program, lets sales, marketing, and revenue teams use natural language to build workflows, run searches, and activate campaigns. For teams building AI agent workflows, ZoomInfo MCP includes a Find similar companies tool, and one of its published workflows expands a closed deal into a lookalike account list with matched personas and enriched contacts.
B2B teams use lookalike audiences to:
Reach new accounts that mirror your best customers
Spend less of the ad budget on accounts outside your ICP
Put pipeline budget on accounts that match your closed-won profile
Scale ABM campaigns beyond your known account list
Smartsheet saw a 3X increase in its average audience match rate using ZoomInfo Marketing for advertising campaigns.
How ZoomInfo powers B2B lookalike audience building
The manual B2B workflow is to export a CRM list, upload it to an ad platform, and hope the match rate holds. The weak points are stale data, low match rates, and no connection between the lookalike model and the intent signals that show which matched accounts are in-market.
ZoomInfo Marketing builds the model on ZoomInfo's side. Its Account Fit Score learns the firmographic, technographic, and geographic patterns of your closed-won deals from your CRM or an uploaded customer list, then scores accounts in ZoomInfo's database by their similarity to your ideal customer profile. You can build audiences from hundreds of company insights, including firmographic, technographic, intent, and engagement data, then run them through ZoomInfo Advertising or send them to Meta, LinkedIn, and Google for retargeting.
Score fit first, then layer intent. An account that matches your ICP and is researching your category belongs at the top of the list. An account that matches on fit alone can wait for a broader awareness campaign.
Redwood Logistics cut cost per click by nearly 99% with ZoomInfo Marketing display campaigns that target on job titles, intent signals, website behavior, and account lists.
Request a demo to see how ZoomInfo Marketing builds and activates B2B lookalike audiences.
Best practices for building effective lookalike audiences
Follow these practices to get more from your lookalike audiences:
Use your highest-quality seed: Closed-won accounts make a tighter seed than broad website visitor lists
Refresh regularly: Refresh the seed list quarterly, or monthly if you add customers quickly
Set suppression lists: Exclude current customers from acquisition campaigns
Test multiple seeds: Compare lookalikes from different customer segments
Align seed to goal: Use high-LTV customers when deal value matters most, fast-closers for pipeline velocity
Start with a clean, ICP-aligned seed list
The quality of your CRM data shapes the quality of your lookalike audiences: enriched, accurate firmographic records produce tighter seed lists and better algorithmic matches.
Build seeds carefully using ICP criteria (firmographics, technographics). Filter for positive outcomes like closed-won accounts, high NPS, and expansion accounts. Remove bad-fit customers who converted despite being wrong for your product. Working with an all-in-one AI GTM Platform like ZoomInfo makes it easier to verify firmographic and technographic attributes before they enter your seed list.
Use this checklist for seed list hygiene:
Remove duplicates and outdated records
Filter for ICP-fit accounts only
Prioritize customers with strong outcomes (retention, expansion, referrals)
Exclude accounts that churned quickly or required heavy support
Common lookalike audience mistakes (and how to avoid them)
Watch for these common pitfalls:
Using all website visitors as seed: Too broad. Use converters or high-intent visitors instead.
Never refreshing the seed: Lookalikes decay as your customer base evolves.
Audience overlap: If your lookalike overlaps heavily with retargeting, two campaigns split budget on the same people.
Defaulting to the largest size: Start smaller for precision, and expand only if volume is the goal.
Privacy and compliance for lookalike audiences
GDPR and CCPA can apply to the personal data you upload to build lookalike audiences. The steps below cover operational hygiene. Check legal requirements with your own counsel:
Use first-party data you have a lawful basis to use: Upload contacts whose consent or other basis you have documented
Document data sources: Track where seed data originated
Hash identifiers: Use hashed emails where platforms support it
Maintain suppression lists: Exclude opt-outs and do-not-contact records
Establish governance: Define who can upload and maintain lists
Meta says changes made to align its products with evolving privacy rules in Europe may affect campaigns advertising to people in the European Region. If you run lookalike campaigns that reach people in the European Region, verify your seed data consent status and watch Meta's policy updates.
ZoomInfo holds ISO 27001, ISO 27701, SOC 2 Type II, TRUSTe GDPR, and TRUSTe CCPA certifications.
To see how ZoomInfo helps B2B teams build accurate audiences, start for free.
Frequently asked questions
What is a lookalike audience?
A lookalike audience is a group of net-new prospects who share the same characteristics as your best customers, identified algorithmically from a seed list of existing customers. Ad platforms analyze behavioral and demographic signals, and B2B data providers analyze firmographic and technographic signals, then return new prospects who match those patterns.
What is lookalike modeling?
Lookalike modeling is the method behind a lookalike audience. A model studies the shared traits of a seed list, such as your closed-won customers, and selects or scores new people or accounts that share those traits. Ad platforms like Meta and Google run the model on their own users. B2B data providers like ZoomInfo run it on company and contact data, which suits account-based targeting.
What's the difference between a custom audience and a lookalike audience?
A custom audience targets people you already know, such as existing customers and email subscribers uploaded from your CRM, or website visitors tracked by your pixel. A lookalike audience uses that custom audience as a seed list to find algorithmically similar net-new prospects outside that list. Use custom audiences to reach known contacts and lookalike audiences to add new ones.
How do I build a seed audience for lookalike modeling?
Start with your highest-quality customer segments: closed-won accounts from the past 12-24 months, high-LTV customers, and SQLs that converted quickly. Filter for ICP-fit accounts only and remove bad-fit customers and accounts that churned quickly. Meta needs at least 100 people in a source audience and recommends 1,000 to 5,000. Clean and enrich the list before uploading because stale or inaccurate records degrade the lookalike model at the source.
How do intent signals improve lookalike audience targeting?
Lookalike audiences tell you who looks like your best customers. Intent signals tell you which of those accounts are researching a solution now. Combining both tells you which targets to prioritize: accounts with the right profile and active buying behavior. ZoomInfo Marketing lets you build audiences that combine firmographic, technographic, intent, and engagement data.
Does LinkedIn still offer lookalike audiences?
No. LinkedIn discontinued lookalike audiences on February 29, 2024. It now recommends Predictive audiences, built from sources such as contact lists, company lists, and conversions, and Audience Expansion for Matched Audiences.
Can I use lookalike audiences for ABM campaigns?
Yes. Lookalike audiences can help ABM programs expand beyond known account lists. Build your seed list from Tier 1 closed-won accounts, then use the lookalike output to identify net-new accounts that fit your top-tier ICP. Layer intent signals on top to prioritize which new accounts to target first.
How often should I refresh my lookalike audience?
Refresh the seed list behind your lookalike audiences at least quarterly. Lookalike models decay as your customer base evolves: new closed-won accounts, churned customers, and role changes in your seed list all affect the model's accuracy. Teams adding customers quickly may need to refresh monthly.