Facebook Custom Audiences: Complete Guide 2026

A retargeting campaign aimed at website visitors who browsed your pricing page will almost always outperform a cold prospecting campaign shown to people who have never heard of your brand. The numbers back this up consistently: CPA (cost per acquisition) for retargeting typically runs 60-75% lower than prospecting across most industries in the UK and US. The mechanism behind this isn’t complicated. People who already know you are easier to convert.

Facebook Custom Audiences are the tool that makes this possible inside Meta Ads Manager. They let you build targetable segments from your own data: website visitors tracked through the Meta Pixel, customer lists uploaded from your CRM, app users logged via the Facebook SDK, and people who’ve engaged with your content on Facebook or Instagram. From these custom audiences, you can then create Lookalike Audiences to find new customers who share similar characteristics.

If you’re looking for a broader overview of Meta’s advertising platform before diving into audience strategy, our Facebook Ads guide covers account setup, campaign objectives, and ad formats. This piece focuses specifically on custom audiences, retargeting strategies, Lookalike configurations, and the GDPR compliance requirements that UK and EU advertisers need to get right.

Custom Audience Types

Meta offers four source categories for custom audiences. Each pulls from different data, serves different campaign stages, and requires different technical prerequisites. Understanding what each type actually does, and where it falls short, is the first step toward building a retargeting structure that performs.

Website Custom Audiences (Pixel-Based)

If you’ve installed the Meta Pixel on your website (ideally through Google Tag Manager for cleaner implementation), you can create audiences based on visitor behaviour. This is the backbone of most retargeting setups.

The segments you can build from Pixel data:

• All website visitors over a chosen time window (1 to 180 days). The broadest retargeting pool. Useful for brand recall campaigns, less useful for direct response unless your site traffic is already well-qualified.
• Visitors to specific pages. Pricing page visitors, product page viewers, contact page visitors. These segments carry higher intent signals. Someone who looked at your pricing but didn’t convert is a fundamentally different prospect than someone who bounced after three seconds on the homepage.
• Visitors by time spent. Top 5%, 10%, or 25% of visitors ranked by time on site. Meta calculates this automatically. These are the people who actually read your content, scrolled through product details, or compared options. Not drive-by clicks.
• Event-based audiences. If you’ve set up standard events , you can build audiences around specific actions. “Added to cart in the last 14 days but didn’t purchase” is the single most valuable retargeting segment for e-commerce businesses.

Time window selection matters more than most advertisers realise. A 7-day website visitor audience converts at significantly higher rates than a 60-day audience, but the pool is much smaller. The sweet spot depends on your traffic volume. Sites with 500+ daily visitors can run profitable campaigns against 7-day windows. Sites with 50-100 daily visitors may need 30-60 day windows to build a large enough audience for Meta’s algorithm to optimise delivery.

One technical point worth noting: the Meta Pixel alone isn’t enough for reliable tracking in 2026. Apple’s App Tracking Transparency (ATT), browser restrictions on third-party cookies, and ad blockers all reduce Pixel match rates. Implementing the Conversions API (CAPI) alongside the Pixel creates a server-side tracking layer that captures events the Pixel misses. The combination typically improves match rates by 15-30%.

Customer List Audiences

You upload a CSV or TXT file containing customer data, and Meta matches it against its user database. The platform hashes the data (SHA-256 encryption) before matching, so raw personal data never sits on Meta’s servers. Match rates in the UK and US typically fall between 50-75%, depending on data quality.

The fields you can include: email address, phone number (with country code), first name, last name, city, postcode, country, date of birth, gender, Facebook App User ID, and Apple IDFA. Email alone gets you 50-60% match rates. Adding phone number plus name fields pushes it to 65-75%. The more identifiers you provide, the more matches Meta can confirm.

The real power of customer list audiences isn’t in targeting those specific people (though that’s useful for upsell and cross-sell campaigns). It’s in using them as seed audiences for Lookalikes. Your top 500 customers by lifetime value, uploaded as a list and used to generate a 1% Lookalike, will almost certainly outperform any interest-based targeting you could build manually.

Practical segmentation strategies for customer lists:

• High-value customers (top 20% by revenue): Use as Lookalike seed. Also target directly with loyalty offers or premium product launches.
• Lapsed customers (no purchase in 6-12 months): Win-back campaigns with a specific offer or product update announcement.
• Newsletter subscribers who haven’t purchased: These people are interested but haven’t committed. Retarget with social proof, case studies, or a low-friction entry offer.
• Recent purchasers (last 30 days): Exclude from acquisition campaigns. Target separately with cross-sell recommendations.

Customer list audiences integrate naturally with your email marketing strategy. The segments you’ve already built in Mailchimp, Klaviyo, or HubSpot can be exported and uploaded to Meta for coordinated messaging across channels.

App Activity Audiences

For businesses with a mobile app, the Facebook SDK tracks in-app behaviour and allows audience creation based on app events. Segments include: users who opened the app in the last X days, users who completed a specific action (level reached, subscription started, item purchased), users who installed but never opened, and high-value users based on in-app purchase history.

This audience type is most relevant for mobile gaming, SaaS with mobile apps, fintech, food delivery, and fitness apps. If your business doesn’t have an app, skip this source entirely. The other three cover your needs.

Engagement Custom Audiences

These audiences come from interactions people have had with your content on Meta’s own platforms. No Pixel required, no data upload needed. Everything is built from platform-native engagement data.

Available engagement sources:

• Facebook Page engagement: People who visited your Page, engaged with any post or ad, clicked a call-to-action button, or sent your Page a message. Time window: up to 365 days.
• Instagram account engagement: Profile visitors, post/reel interactions (likes, comments, saves, shares), ad engagers, and DM senders. Particularly valuable for brands with active organic Instagram presence.
• Video viewers: Segmented by watch depth: 3 seconds, 10 seconds, 25%, 50%, 75%, or 95% completion. This graduated segmentation is incredibly useful. Someone who watched 75% of a 60-second brand video has shown meaningful interest. That’s a warmer prospect than most website visitors.
• Lead form engagers: People who opened a lead form but didn’t submit it. The abandon rate on lead forms runs between 60-80% on average. Retargeting these people with a simpler form or a different value proposition can recover a significant portion of that lost intent.
• Shopping engagers: Users who viewed products, added to wishlist, or initiated checkout through Facebook or Instagram Shops.
• Event RSVPs: People who responded to a Facebook Event you created.

Engagement audiences are the easiest starting point for businesses that haven’t set up the Pixel yet. They require zero technical implementation. If you’re running any organic content on Facebook or Instagram, you already have engagement data accumulating. Creating audiences from this data and running retargeting campaigns against it is one of the fastest paths to demonstrating ROI from social media.

The video viewer funnel deserves specific attention. Running a 60-second educational or brand video to a broad audience (low cost, optimised for ThruPlay), then retargeting 50%+ viewers with a conversion-focused ad, consistently produces CPAs 40-60% lower than going straight for conversion with a cold audience. The video pre-qualifies the audience for you.

Building Custom Audiences Step by Step

The creation process inside Meta Ads Manager is straightforward, but there are decisions at each step that affect campaign performance. Here’s the walkthrough for each audience type.

Website Audience Setup

1. Open Meta Ads Manager. Navigate to Audiences from the left menu (or go to Business Settings > Audiences).
2. Click Create Audience > Custom Audience.
3. Select Website as the source.
4. Choose your Pixel from the dropdown. If you have multiple pixels, make sure you select the one installed on the correct domain.
5. Define the audience rules:
   • All website visitors if you want the broadest retargeting pool.
   • People who visited specific web pages if you want intent-based segments. Enter URL rules (contains, equals, doesn’t contain).
   • Visitors by time spent if you want to target the most engaged visitors.
   • Custom combination if you need AND/OR logic across multiple conditions.
6. Set the retention window (1-180 days). Start with 30 days for general retargeting, 7 days for high-intent pages like pricing or checkout.
7. Name the audience descriptively. “WEB – All Visitors – 30d” is better than “Custom Audience 1.”
8. Click Create Audience. Population usually completes within a few hours. Large audiences may take up to 24 hours.

Customer List Upload

1. Follow steps 1-2 above, then select Customer list as the source.
2. Choose whether to upload a file or import from a connected integration (Mailchimp, Shopify, etc.).
3. Prepare your CSV file. Include as many identifier columns as possible: email, phone (with country code, e.g. +44 or +1 format), first name, last name, city, postcode, country. Clean the data first: lowercase emails, remove special characters from phone numbers, standardise name formatting.
4. Upload the file. Meta will map columns to its identifier fields. Verify the mapping is correct.
5. If you want a Value-Based Lookalike later, include a value column (customer lifetime value, total spend, or average order value).
6. Accept the Custom Audience Terms of Service. This confirms you have proper consent to use this data for advertising .
7. Name and create. Match rates appear after processing, usually within 30 minutes to a few hours.

Engagement Audience Setup

1. Follow steps 1-2, then select the engagement source you want: Video, Lead Form, Instagram Account, Facebook Page, Shopping, or Events.
2. Choose the specific engagement criteria. For video: select which videos and the watch depth threshold. For Instagram: choose between all engagers, profile visitors, post/ad engagers, or message senders.
3. Set the time window (up to 365 days for most engagement types).
4. Name and create.

After creation, audiences appear in the Audiences section and can be selected when building ad sets. They refresh automatically as new data comes in, so your retargeting pools stay current without manual intervention. Customer lists are the exception: those need periodic re-uploading to stay fresh.

Lookalike Audiences

A custom audience tells Meta “show ads to these specific people.” A Lookalike Audience tells Meta “find me more people who look like these specific people.” The algorithm analyses the characteristics of your source audience (demographics, interests, online behaviour, purchasing patterns) and identifies users across the UK, US, or any other country who share similar traits.

How Lookalike Sizing Works

You select a percentage from 1% to 10%. In the UK, a 1% Lookalike draws from approximately 450,000 people. In the US, it’s roughly 2.4 million. The percentage represents how closely the generated audience matches your source. 1% is the tightest match, and typically delivers the lowest CPA. As you expand to 3%, 5%, 10%, you get larger reach but weaker signal.

Most advertisers should start with 1% and only expand if they’re running out of delivery volume or scaling spend beyond what the 1% audience can absorb. In practice, a 1% UK Lookalike of 450,000 people can support daily budgets of £50-200 comfortably. You’d only need to go wider if you’re spending significantly more than that or if frequency starts climbing above 3.

Source Audience Quality Determines Everything

The Lookalike is only as good as the seed you feed it. A Lookalike built from your best 200 customers will outperform one built from 50,000 random website visitors, every time. The algorithm doesn’t need volume. It needs clear signal.

Meta’s recommended minimum source size is 100 people, but audiences of 1,000-50,000 tend to produce the most reliable Lookalikes. Below 1,000, the algorithm doesn’t have enough data points to identify patterns confidently. Above 50,000, the source becomes so broad that the “similarity” signal weakens.

Value-Based Lookalikes

When you upload a customer list with a value column (total spend, customer lifetime value, average order value), Meta weights its Lookalike modelling toward high-value customers. Instead of finding people who look like your average customer, it finds people who look like your best customers.

For e-commerce brands with meaningful differences between customer cohorts, this is a significant upgrade. If your top 20% of customers generate 60% of your revenue (a common pattern), a standard Lookalike treats every customer equally. A Value-Based Lookalike focuses on the traits that distinguish your high-spenders from everyone else. Advertisers running Value-Based Lookalikes typically report 15-25% higher ROAS (return on ad spend) compared to standard Lookalikes from the same source data.

Testing Multiple Lookalikes

Don’t guess which source audience produces the best Lookalike. Test it. Set up three ad sets with identical budget, creative, and placement settings. Only the audience differs:

• Ad Set A: 1% Lookalike from purchasers
• Ad Set B: 1% Lookalike from lead form completions
• Ad Set C: 1% Lookalike from website visitors (top 25% by time spent)

Run all three for 7-14 days with enough budget to generate at least 50 conversions per ad set (for statistical significance). The winner gets scaled. The losers get paused. This test often produces surprising results. In some accounts, the lead form Lookalike outperforms the purchaser Lookalike because the lead form data is fresher and the conversion event fires more frequently, giving Meta’s algorithm more signal to work with.

One caution about Lookalike overlap: if you’re running both a Lookalike audience and interest-based targeting in separate ad sets within the same campaign, check the Audience Overlap tool in Meta Ads Manager. If overlap exceeds 30%, you’re bidding against yourself and inflating your own costs. In 2026, the recommended approach is to use either Advantage+ Audience or Lookalikes. Layering interests on top of a Lookalike rarely improves performance and often hurts it.

Retargeting Strategy by Industry

Retargeting is not one strategy. The approach that works for an e-commerce fashion brand has almost nothing in common with the approach that works for a B2B SaaS company. The product, the buying cycle, the price point, and the decision-making process all shape how retargeting should be structured.

E-Commerce

The highest-value retargeting segment in e-commerce is abandoned cart. These people selected a product, added it to their basket, reached the checkout page, and then left. They were seconds away from buying. Dynamic product ads (DPAs) that show the exact items left in the basket, sometimes with a time-limited incentive, are the highest-converting ad format in Meta’s entire system for this segment.

Setting up DPAs requires a product catalogue connected to Meta . Once the catalogue syncs, Meta automatically pulls product images, prices, and availability into the ad creative. No manual ad creation needed for each product.

Beyond abandoned cart, the layered time-window approach works well for e-commerce:

Budget allocation for e-commerce: roughly 70% to prospecting and 30% to retargeting. That 30% will often generate 50%+ of your revenue because it targets people who already showed purchase intent.

Professional Services (Law Firms, Accountants, Consultancies)

Service businesses face a different dynamic. There’s no shopping cart to abandon. The conversion action is typically a contact form submission, phone call, or consultation booking. The buying decision involves higher perceived risk (“am I choosing the right solicitor?”) and often takes multiple touchpoints to close.

The most effective retargeting segment for services: visitors to pricing, services, or “about us” pages who didn’t convert. These people are actively evaluating. Retarget them with client testimonials, case study highlights, professional accreditations, and content that builds trust. A video of a partner discussing a relevant legal topic builds more confidence than a discount offer ever could.

For lead generation campaigns, consider retargeting lead form abandoners. Between 60-80% of people who open a Meta lead form close it without submitting. That’s not because they weren’t interested. They got distracted, thought the form was too long, or wanted more information before committing. A retargeting ad that addresses the most common objection (“No obligation. 15-minute call. We’ll review your situation and give honest advice.”) can recover a meaningful portion of these drop-offs.

B2B and SaaS

B2B buying cycles stretch across weeks or months. Multiple stakeholders review options. Nobody signs a five-figure contract because they saw one ad. This means retargeting in B2B isn’t about pushing for immediate conversion. It’s about staying visible throughout the evaluation period while progressively building credibility.

A sequential retargeting approach works well:

1. Week 1-2: Show thought leadership content (blog posts, industry insights, data-driven analysis) to website visitors. The goal is positioning, not conversion.
2. Week 3-4: Show case studies and client results to the people who engaged with the thought leadership. They know your name now. Show them proof.
3. Week 5+: Show a direct CTA (demo request, consultation booking, trial signup) to the people who engaged with case studies. By now, they’ve had multiple positive touchpoints with your brand.

This sequential approach requires setting up separate custom audiences for each stage and using exclusion logic to prevent people from seeing the wrong message at the wrong time. More on exclusion strategies in the next section.

For B2B retargeting, consider combining Meta campaigns with LinkedIn Ads. LinkedIn’s professional targeting is unmatched for B2B, and running retargeting across both platforms ensures your brand appears wherever decision-makers spend their time.

Local Businesses (Restaurants, Gyms, Salons, Clinics)

Local businesses need short retargeting windows. The decision cycle for “where should I eat tonight” or “which gym should I join” is hours to days, not weeks. A 7-14 day retargeting window is usually sufficient. Anything beyond 30 days is typically wasted budget for local services.

Engagement audiences work particularly well here because local businesses often have strong organic social presence but limited website traffic. Retargeting Instagram profile visitors, post engagers, and Reel viewers requires no Pixel setup and reaches an audience that already has some familiarity with the brand.

Geo-targeting layered on top of custom audiences (retarget website visitors within a 10-mile radius) prevents budget waste on people who are geographically too far away to visit.

Layered Funnels and Exclusion Logic

Running all your retargeting through a single audience (“all website visitors, last 30 days”) is the most common mistake in custom audience strategy. It means your newest prospect who visited the pricing page yesterday sees the same ad as someone who glanced at a blog post 28 days ago. Their intent levels are completely different. The message should be different too.

Time-Based Layering

Split your website visitors into time-based cohorts and assign different messages and budgets to each:

• 1-7 days: Hottest audience. Highest conversion rate. Allocate the largest share of your retargeting budget here. Messaging should be direct: product benefits, social proof, strong CTA.
• 8-21 days: Warm audience. Conversion rate drops but the pool is larger. Messaging shifts to reminders and alternative value propositions.
• 22-60 days: Cooling audience. Lower budget allocation. Messaging focuses on brand recall, educational content, or a fresh angle they haven’t seen before.

To prevent overlap, each ad set should exclude the audiences above it in the funnel. The 8-21 day ad set excludes the 1-7 day audience. The 22-60 day ad set excludes both the 1-7 and 8-21 day audiences. This ensures each person sees the appropriate message for their recency level.

Behaviour-Based Layering

Beyond time windows, segment by behaviour. Combine conditions with AND/OR logic:

• Visited pricing page AND viewed at least 2 product pages: Very high intent. These people are comparing options and checking costs.
• Watched 75%+ of a brand video OR visited the case studies page: Strong interest. They’ve invested time in understanding your brand.
• Visited any page EXCLUDE purchasers: Basic exclusion to prevent showing acquisition messaging to existing customers.

Exclusion Rules Every Account Should Have

Exclusion logic is as important as targeting logic. Without exclusions, you waste budget showing irrelevant messages to the wrong people.

• Exclude purchasers from prospecting and retargeting campaigns. Create a separate campaign for existing customers with upsell, cross-sell, or loyalty messaging.
• Exclude lead form submissions from lead gen campaigns. Someone who already submitted a form doesn’t need to see the same form again.
• Exclude existing retargeting audiences from prospecting campaigns. If someone is already in your retargeting funnel, you don’t want to also reach them through your cold audience campaign (where CPC is typically higher).
• For e-commerce: exclude recent purchasers (last 7-14 days) from all campaigns except post-purchase flows. Nobody wants to see ads for the shoes they bought three days ago.

Check audience overlap regularly using the Audience Overlap tool in Ads Manager . If two audiences in separate ad sets overlap by more than 30%, consolidate them or add exclusion rules.

GDPR and Data Privacy Compliance

If you’re running ads targeting the UK or EU, GDPR (General Data Protection Regulation) and the UK GDPR aren’t optional considerations. They’re legal requirements with potential fines of up to 4% of annual global turnover or 17.5 million pounds (UK) / 20 million euros (EU), whichever is greater. Getting custom audiences wrong from a data protection perspective creates real business risk.

The areas that require attention:

Customer List Uploads: Lawful Basis

Under GDPR, you need a lawful basis to process personal data. For uploading customer data to Meta for advertising purposes, the two most relevant bases are:

• Consent (Article 6(1)(a)): The individual has given explicit consent for their data to be used for targeted advertising. This is the safest approach. Tick-box opt-in at point of collection (“I agree to receive marketing communications including targeted social media advertising”) that is separate from terms and conditions, not pre-ticked, and clearly worded.
• Legitimate Interest (Article 6(1)(f)): The processing is necessary for a legitimate interest and doesn’t override the individual’s rights. Some businesses use this for customer list uploads, but it requires a documented Legitimate Interest Assessment (LIA) and the interest must be balanced against the individual’s reasonable expectations. If a customer wouldn’t expect their email to be uploaded to Facebook, legitimate interest probably doesn’t hold up.

In practice, the ICO (Information Commissioner’s Office) in the UK and DPAs across Europe increasingly expect explicit consent for sharing customer data with third-party advertising platforms. Building robust consent mechanisms from the start avoids costly retroactive fixes.

Meta Pixel and Cookie Consent

The Meta Pixel uses cookies and tracking technologies that fall under the Privacy and Electronic Communications Regulations (PECR) in the UK and the ePrivacy Directive in the EU. You must obtain consent before the Pixel fires. A compliant cookie consent banner needs to:

• Present marketing/analytics cookies as opt-in, not opt-out
• Not use dark patterns (e.g. making “Accept All” prominent while hiding “Reject”)
• Allow granular control (users can accept analytics cookies but reject marketing cookies)
• Not pre-select any optional cookie categories
• Store and be able to demonstrate consent records

Technically, this means implementing Consent Mode. Meta supports this through its Conversions API and Pixel. When a user declines marketing cookies, the Pixel doesn’t fire, but server-side events via CAPI can still send aggregated, non-identifying conversion data. This preserves some tracking capability while respecting user choices.

Privacy Policy Requirements

Your privacy policy must explicitly mention:

• That you use Meta Pixel and/or Conversions API for tracking
• That you share customer data with Meta for custom audience targeting
• The categories of data shared (email, phone, purchase history)
• The lawful basis for processing
• How long custom audience data is retained
• How individuals can exercise their right to opt out or request deletion

Data Subject Rights

Under GDPR, individuals have the right to request deletion of their data. If a customer exercises this right, you must remove their data from your CRM and also remove them from any custom audiences uploaded to Meta. Meta provides an option to remove individual records from uploaded customer lists, but this requires maintaining a process internally to action these requests within the mandatory 30-day response window.

For website builds Bravery delivers, GDPR-compliant cookie management and privacy policy frameworks are included as standard. Getting these foundations right during site development is far cheaper than retrofitting compliance later.

GDPR Compliance Checklist for Custom Audiences

1. Collect explicit opt-in consent for marketing communications, including targeted advertising, at every data collection point (forms, checkout, newsletter signup).
2. Document your lawful basis for uploading customer data to Meta (consent or legitimate interest with a completed LIA).
3. Implement a PECR/ePrivacy-compliant cookie consent banner with Consent Mode integration for the Meta Pixel.
4. Update your privacy policy to cover Meta data sharing, Pixel tracking, and custom audience creation.
5. Establish an internal process for handling data subject access and deletion requests, including removal from Meta custom audiences.
6. Review and re-upload customer lists every 90 days to ensure deleted records aren’t re-introduced.
7. Maintain records of consent and processing activities as required by Article 30.

If your business processes data in regulated sectors (healthcare, financial services, legal), additional sector-specific regulations may apply. Consulting with a data protection specialist or solicitor is advisable before launching custom audience campaigns in these industries.

Naming Conventions and Maintenance

This sounds like a minor administrative detail. It isn’t. Within three months of actively building custom audiences, most accounts accumulate 20-40 audiences. Without a consistent naming system, finding the right audience during campaign setup becomes a time-consuming guessing game, and misconfigured targeting (using the wrong audience in an ad set) is one of the most expensive mistakes you can make.

Recommended Naming Format

[Source] – [Segment] – [Window] – [Date Created]

Examples:

• WEB – All Visitors – 30d – Apr26
• WEB – Pricing Page – 7d – Apr26
• WEB – Add to Cart No Purchase – 14d – Apr26
• LIST – High Value Customers Top 20% – Mar26
• LIST – Newsletter Subs No Purchase – Apr26
• IG – Profile Visitors – 90d – Apr26
• VID – Brand Video 50% Viewers – 30d – Apr26
• LEAD – Form Openers Not Submitted – 30d – Apr26
• LAL – 1% UK – Purchasers – Apr26
• LAL – 1% US – High Value Customers – Apr26

The prefix (WEB, LIST, IG, VID, LEAD, LAL) lets you sort audiences by type at a glance. The date helps identify stale audiences that need refreshing or archiving.

Maintenance Schedule

Custom audiences aren’t set-and-forget. They need periodic maintenance:

• Customer lists: Re-upload every 60-90 days. Customer data changes. People switch email providers, change phone numbers, or need to be removed due to data deletion requests. Stale lists lose match accuracy over time.
• Website and engagement audiences: These refresh automatically as long as the Pixel is active and your Facebook/Instagram content continues generating engagement. No manual action needed, but review them quarterly to ensure the Pixel is tracking correctly and audience sizes are consistent with your traffic levels.
• Lookalike audiences: Recreate quarterly using updated source audiences. Meta doesn’t auto-update Lookalikes when the source audience changes. A Lookalike created from January purchasers won’t include February and March purchaser data in its modelling unless you recreate it.
• Archive inactive audiences: Audiences you’re no longer using should be archived (not deleted, in case you need them later). This keeps your Audiences dashboard manageable.

The combination of a clean naming convention and quarterly maintenance reviews prevents the two most common operational problems: using outdated audiences that hurt campaign performance, and accidentally overlapping audiences that inflate costs.

Audience Size and Budget Alignment

A common error is allocating too much budget to too small an audience. Retargeting pools are naturally smaller than prospecting audiences. If your website gets 200 visitors per day, your 30-day retargeting audience is roughly 6,000 people. Spending 50 pounds per day against 6,000 people will push frequency (the average number of times each person sees your ad) above 3-4 within days. At that point, ad fatigue sets in, CTR drops, CPA rises, and people start hiding your ads.

A practical rule: keep daily retargeting spend at roughly 1 penny per person in your audience. 6,000-person audience = roughly 60 pounds per day maximum. 20,000-person audience = roughly 200 pounds per day maximum. Adjust based on actual frequency data in Ads Manager, but this gives you a starting framework.

Frequently Asked Questions