Most website visitors don't buy a product or submit a request during their first visit. They might explore the product's features, check prices, compare offers, close the page, and then return to the website a few days later. Facebook retargeting helps you continue reaching this audience and show them ads after their previous interaction with the business.
Meta Ads lets you target website visitors, Instagram followers and active audiences, video viewers, lead form users, your customer base, and other segments. Website data is typically transmitted via the Meta Pixel and Conversions API, and audiences are built based on specific user actions. This approach is often referred to as Facebook retargeting.
Retargeting shows ads to people who have already interacted with a company and left a digital footprint. This could include visiting a website, viewing a product page, adding an item to a cart, interacting with an Instagram post, watching a video, opening a form, or making a previous purchase. The user can then be assigned to the corresponding Custom Audience.
Next, the advertiser selects the desired segment at the ad group level and displays a suitable message. A catalog visitor can be reminded of a category they viewed, someone who abandoned a shopping cart can be shown a product again, and a potential customer can be offered a case study or consultation. The logic depends on the user's previous action.
The work scheme is as follows:
Data source → user action → Custom Audience → ad → return visit → request or purchase.
This sequence helps leverage existing demand. However, the mere fact that a person is added to an audience doesn't guarantee anything: the result depends on the quality of the initial data, proper segmentation, offer, creative, frequency, and configured analytics.
Standard targeted advertising often reaches users who may not have previously been familiar with the company. The audience is selected based on parameters available to the advertising system, or the algorithm searches for people who are highly likely to perform the desired action. Therefore, the first advertising exposure often occurs during the product discovery phase.
Facebook retargeting begins after a previous interaction. The advertiser knows that a person visited a specific page, interacted with a profile, watched a video, or performed another action. A message can be tailored to that user's previous behavior instead of showing the company's generic ad again.
The difference is especially noticeable in niches with long decision-making cycles. Potential clients of legal, medical, educational, or B2B services rarely submit a request after a single page view. A repeat touchpoint helps them return to a specific service, review a case study, find answers to questions, or discuss the terms of cooperation.
The terms "retargeting" and "remarketing" are often used interchangeably. Both describe repeated communication with users who have previously interacted with a business. In practice, advertising professionals typically infer the meaning from the specific platform, audience source, and chosen scenario.
Historically, the term "remarketing" has been particularly common in Google Ads materials, while "retargeting" is widely used when discussing Meta Ads, social media, and advertising platforms. For businesses, the underlying logic is more important: identifying a person's previous action, building a suitable audience, and displaying a message that's relevant to their current decision stage.
Therefore, Facebook retargeting doesn't require a separate logic just because of the term itself. The setup is built around data sources, Custom Audiences, events, exclusions, and ad scenarios that match audience behavior.
Retargeting is used in situations where one ad exposure isn't enough to elicit a purchase or lead. It helps re-engage interested visitors, continue communication with potential customers, and segment users based on their readiness to take the desired action.
Most often, retargeting is used to solve the following problems:
Each task requires a separate scenario. It's generally not a good idea to show the same message to a user who's just read an article as to someone with an uncompleted order. The more closely the advertising communication matches the previous action, the easier it is to assess the campaign's true impact.
A retargeting audience member has already seen the company, so it's pointless to start each communication with the same general introduction. Creative should take into account previous contact and help the user obtain the information they might have lacked for their next action.
For different segments, case studies, reviews, objection responses, product reminders, delivery terms, consultations, and other messages are used. The choice depends on the product and the reason why the user may have stopped at an earlier step.
A case study helps users understand how a company solved a similar problem for another client. For complex services, this format is often more useful than a repeat ad with a general list of benefits.
A good case study provides concrete context: the original task, the work performed, and the actual results that can be verified. If there are no numerical data, you can't invent them for the sake of persuasiveness.
Social proof can also include real testimonials, project examples, certificates, or other verifiable materials. Ads should link to a page where users can explore the details themselves.
After the first visit, people may delay making a decision due to price, uncertainty about the results, unclear terms, or the need to compare several companies. A retargeting ad can answer precisely this question.
For example, instead of repeating a general slogan, you can demonstrate the process, timelines, payment terms, or an explanation of a complex step in the service. For a high-priced product, a detailed case study or consultation is useful, while for an online store, clear delivery and return policies are more effective.
To identify real objections, it's worth using sales department data, correspondence, and customer questions. Imaginary pain points usually result in less accurate communication.
This scenario is especially understandable for e-commerce, where a user might view several products and then postpone a purchase. Advertising reminds the user of their previous interest and returns them to a specific category or product.
This approach is also applicable to services. Visitors to a specific service's page can be shown a case study on that service or an answer to a common question, rather than a general advertisement for the agency or company.
Reminders shouldn't become endless harassment. If users don't respond for a long time, frequency increases, and performance declines, the campaign and timeframe should be reconsidered.
An additional incentive can sometimes help bring people back to the decision, but the offer must be tangible. This could be a relevant discount, a free estimate, a consultation, a bonus, or another condition that the company is truly willing to fulfill.
Artificial scarcity, a non-existent timer, and a constant "last chance" quickly undermine trust. These tactics are especially ineffective when users regularly see the same ad for weeks at a time.
For expensive services, a discount isn't always the primary selling point. Sometimes it's more helpful to demonstrate the process, guarantees within the actual commitments, a case study, or the team's qualifications.
A single advertising message can be adapted to multiple placements, but the visual presentation must take into account the viewing environment. A short vertical video and a card in the feed require different compositions, even if they lead to the same page.
For retargeting, it's useful to test multiple formats, as audiences quickly become accustomed to a single creative. However, the test must maintain a clear logic: if you simultaneously change the audience, offer, copy, and visuals, it will be more difficult to determine the cause of the result.
The creative should also match the landing page. If the ad promises a specific service or product, the user should see it immediately after clicking.
Stories and Reels are typically consumed vertically, so the key message needs to be quickly grasped on a small screen. Cluttered text, small details, and horizontal layouts are less effective due to viewing characteristics.
For a warm audience, you can jump straight to a familiar topic: remind them about the product, show a short case study, or outline the terms of the offer. The user has already interacted with the company, so there's no need to spend the entire video reintroducing the brand.
When creating multiple variations, you should check the readability of the text, safe areas of the interface, and the correctness of the link or CTA.
You can use a static image, carousel, video, or other available format in your feed that matches your campaign's objectives. This format makes it easier to show several arguments, product cards, or a sequence of elements.
For e-commerce, a carousel can help showcase several items, and for services, it can highlight stages of work or parts of a case study. However, the amount of information should be limited so that the ad remains understandable without extensive study.
The ad copy should continue the previous contact. If the user has already visited the service page, it makes more sense to answer the next question than to repeat the content of the first screen of the website.
I don't recommend retargeting all website visitors as a single 180-day audience. A person who added a product to their cart yesterday and a user who read an article two months ago are at different stages of the decision process. I first separate audiences by action and recency of contact, excluding those who have already completed the target action, and then select the offer, creative, and budget. This makes it easier to control frequency and understand which segments are actually driving leads or sales.
The process begins not with an ad, but with defining a specific segment and the action considered a sign of interest. After that, the audience source is selected, inclusion conditions are set, exclusions are created, and only then is the segment included in the advertising campaign.
Step-by-step configuration helps avoid audience overlap and accidentally showing the same ad to people with different motivations. The Meta Ads interface changes over time, so it's best to use the logic of the entities rather than memorizing the location of a specific button.
Dental clinic · Kyiv and Chernihiv
A domain with no history on a website builder. We built the semantic core for both cities, reworked the landing pages and built the link profile from zero. Four months: 34.8k clicks, impressions 1.32 → 1.76M, DR 0 → 41.
E-commerce · international
A catalogue of digital 3D models. We clustered the semantics, rebuilt the hub pages and closed duplicates and indexing errors. Google users 247 → 532, CTR 2.4% → 4%.
Medical centre · Ukraine
Narrow visibility and a small semantic core at the start. Semantics, landing page structure, metadata and internal linking, then gradual link building.
Facebook advertising for business: Meta Ads setup, audiences, creatives, analytics, and optimization. Find out the cost and schedule your campaign launch with Seo-Gen.
Order Instagram ad setup for your business. Audience analysis, targeting, creatives, analytics, and campaign optimization. Find out the cost of Seo-Gen's services.
LinkedIn Ads setup for B2B: strategy, targeting, launch, analytics, and campaign optimization. We help you generate targeted leads and control your advertising costs.
Setting up Meta Ads for business: strategy, audiences, creatives, Pixels, analytics, launching and optimizing ads on Facebook and Instagram.
Ecommerce Facebook Ads for online stores: product catalog, Meta Pixel and CAPI, retargeting, creatives, analytics, optimization, and sales scaling.
Lead generation in Meta Ads: Meta Lead Ads setup, forms and website, lead qualification, CRM, analytics, and ways to improve lead quality on Facebook and Instagram.
Turnkey Telegram Ads setup: audience and channel selection, competitor channel targeting, launch, optimization, and analytics of Telegram advertising campaigns.
Advertising in thematic Telegram channels: platform selection, audience verification, posting, and results monitoring. Order your campaign from Seo-Gen.
Setting up TikTok Ads for business: strategy, Ads Manager, Pixel, audiences, In-Feed and TopView, launch, optimization, and reporting. We'll calculate promotion costs.
Full-service Spark Ads setup on TikTok: we connect accounts, authorize videos, launch campaigns in TikTok Ads Manager, test creatives, and optimize ads against KPIs.
Facebook retargeting is a repeat advertising campaign aimed at users who have previously interacted with a business. They may have visited a website, viewed products, read service pages, interacted with Instagram, watched videos, or opened a lead form.
For each scenario, you can create a separate Custom Audience and show it a suitable ad. This approach helps you maintain contact with interested users and avoid lumping them in with new, cold audiences.
Results depend on the quality of data, segmentation, and advertising offer, so simply launching retargeting is not enough to ensure stable sales.
First, you need to determine the data source and user action that will be used to create the audience. For a website, this typically involves checking the Meta Pixel, key events, and, if necessary, the Conversions API.
Next, create a Custom Audience, set a time window, create exclusions, and connect the audience to the desired Ad Set. After launch, monitor Frequency, CTR, CPA or CPL, and actual conversions.
When setting up, it is better to start with a few understandable segments and gradually complicate the structure once a sufficient amount of data has accumulated.
The Meta Pixel is needed for scenarios based on user actions directly on the website. Without it, the advertising system won't be able to use the standard browser method to determine whether a specific visitor has viewed a product, submitted a request, or added an item to the cart.
However, retargeting is possible without a website. You can work with Instagram and Facebook audiences, video views, lead forms, and customer lists, if these sources are available to the project.
For websites, Pixel is often used in conjunction with other event delivery methods, including the Conversions API.
Yes, some retargeting scenarios don't require a website. A company can build audiences of people who have interacted with Instagram or Facebook, watched videos, opened lead forms, or are already in its customer base.
This option is suitable for local businesses and projects where the primary communication occurs within social networks. However, segmentation capabilities will depend on the available data sources.
If a site exists, its events usually provide additional signals about the user's commercial interest and help to more accurately segment the audience.
Targeted advertising can reach new audiences who haven't previously interacted with a company. Retargeting leverages a user's previous contact and tailors advertising to their existing interest.
For example, a regular campaign brings a person to an online store for the first time, while retargeting brings back a user who has viewed a product or left it in their shopping cart.
Both approaches can operate simultaneously. The first replenishes the top of the funnel, while the second continues communication with the existing audience.
This could be due to a missing Purchase event, a delay in data transfer, incorrect exclusion of buyers, or an error in audience settings. It's also worth checking whether the ad is part of a customer recovery campaign.
If a Purchase isn't reported to the advertising system, the user may still be listed in the AddToCart or InitiateCheckout segment. The campaign then treats them as a person with an unfinished purchase.
The fix begins with checking the event, buyer audience, and exclusion rules at the relevant Ad Set level.
These data transfer methods can work together, so choosing one as a universal replacement for the other is usually unnecessary. Meta Pixel transmits browser events, while the Conversions API is used for server-side transmission of configured events.
When working together, deduplication must be configured correctly to prevent the same action from being counted multiple times. The architecture depends on the website, CMS, server side, and integrations used.
Before launching, you should check the actual transmission of events and ensure that the numbers correspond to real user actions.
A custom audience is a segment of people created based on a selected data source. It can include website visitors, customers from your own database, Instagram and Facebook users, video viewers, or participants in other available interactions.
When creating an audience, the advertiser defines the inclusion conditions and the period during which the contact is considered relevant. The segment is then used in the advertising campaign.
For ease of use, audiences should be given clear names indicating the source, action, and time window.
Retargeting can be used by small businesses if they have a sufficiently large, suitable audience and a clear campaign goal. With very small traffic, excessive segmentation will create groups with insufficient users.
In this situation, it's helpful to start with a broader structure, grouping related actions together, and closely monitoring frequency. As traffic grows, audiences can be further subdivided.
Results should be assessed based on conversions, sales, and project economics, not on the mere existence of a retargeting campaign.
Retargeting in Meta Ads is best built around specific user behavior. Website visitors, video viewers, Instagram users, form openers, and buyers all have different levels of interest, so they require different audiences, time windows, and ad messages.
Before scaling campaigns, check the Meta Pixel, events, Conversions API, Custom Audiences, and exclusions. After launch, results are assessed based on leads, sales, CPA, CPL, ROAS, and CRM data, while Frequency and CTR are used as additional signals to assess the quality of the ad mix.
If Facebook retargeting is already running, but audiences overlap, events are being reported incorrectly, or ads continue to display after a conversion, start by auditing your ad account and data sources. A revised structure provides a clear foundation for further testing of creatives, budgets, and ad scenarios.
We reply within one business day. No newsletters, no “just a reminder” calls.
He will look at the site himself instead of passing it to a manager.
First, determine the source from which the custom audience will be built. This could be a website, Instagram or Facebook activity, video views, a lead form, a customer list, or another available source.
It's best to make the audience name clear to your team. It's helpful to include the source, action, and time window, for example, "Website – AddToCart – 7 days". This system simplifies work when you have dozens of segments in your advertising account.
Before saving, make sure the selected source is actually receiving data. Creating a beautiful audience structure won't help if your website events are configured incorrectly or if the selected Instagram profile is generating virtually no interactions.
A condition defines what action qualifies a user for a specific segment. For a website, this could be a URL visit or a Pixel event; for Instagram, it could be a specific interaction; for videos, it could be a specified view depth; and for a customer base, it could be a match against an uploaded list.
The condition must match the advertising message. If the audience includes visitors to a specific service page, the ad can continue the conversation specifically about that service. If all website visitors are collected without distinction, personalization will be significantly weaker.
With a complex structure, it's better to create a few clear segments first, rather than dozens of nearly identical audiences. Once you've accumulated statistics, you can refine the structure where it truly impacts results.
The window shows how far back a user's contact is taken into account when building an audience. Different periods are appropriate for different products. A user who abandoned a shopping cart may be most relevant in the first few days, while a customer of a complex B2B service may take several weeks to make a decision.
You can test intervals of 1-3, 7, 14, 30, or 60-90 days as working segments. These values are not universally standard. The choice depends on traffic volume, product cost, deal duration, and the number of intermediate touchpoints.
The lower the site's traffic, the harder it is to split audiences finely. Sometimes a seven-day segment turns out to be too small for stable performance, so the structure has to be expanded and the results verified using actual data.
The longer it takes a client to compare options and agree on a purchase, the longer the follow-up period can be. For example, when choosing a complex B2B service or expensive equipment, a potential client may return to the offer several times before committing.
However, a long window doesn't mean that the same person should be constantly shown the same ad. The message can be varied as the distance from the initial contact increases: first, remind them of the service, then show a case study, answer a common objection, or offer a consultation.
The decision on duration is made after analyzing the actual transaction cycle. If most customers convert within two weeks, maintaining the same retargeting campaign for several months without a separate scenario is usually unnecessary.
In e-commerce, short scenarios are especially relevant for viewed products and abandoned carts. User interest in a specific item often changes more quickly than when choosing a complex service, so the first few days after a visit are especially important.
For category visitors, a wider window can be used than for incomplete checkouts. Shoppers with a Purchase event should be excluded from the cart abandonment campaign and, if necessary, transferred to a separate repeat sales segment.
Periods need to be tested based on the number of users. A store with high traffic can afford a detailed breakdown, while a smaller project is sometimes forced to combine similar time periods to ensure a sufficient audience size.
Exclusions help remove users from your campaign who are no longer relevant to a specific ad message. The most obvious example is a customer who has placed an order but continues to see an abandoned cart ad.
You can also exclude people who submitted a request from your initial acquisition campaign, existing customers from new user ads, or a warmer segment from your general website visitor base.
Without such a structure, a single user could be placed in multiple ad groups at once. Then, campaigns begin competing for the same user, resulting in unnecessary impressions and less clear statistics for the advertiser.
After changing events or site logic, exclusions should be re-checked. If the Purchase event stops being transmitted, the customer audience may no longer update correctly.
After audience preparation, a campaign is created for a specific business objective. Its structure typically includes a campaign level, an Ad Set ad group, and the ads themselves. The roles of these levels differ, so it's best to assign settings sequentially.
At the campaign level, the overall promotion logic is defined, at the Ad Set level, audiences and associated display parameters are selected, and at the ad level, creative, text, and link are placed.
Goal names and individual Meta interface elements can be updated, so a long-term setup is built around a business outcome. It's important to define in advance what constitutes success: a purchase, a request, a message, a registration, or another measurable action.
The prepared Custom Audience is selected at the ad group level. Here, you should also check the exclusions and ensure that the selected segment matches the campaign logic.
For example, an abandoned cart group might include AddToCart users for the past seven days and exclude the Purchase audience for the same or a longer period. A group for a service page would be assembled using a different criterion.
For analysis, it's best not to mix too many different segments if your budget and audience size allow for separation. This way, you can see which user groups are generating leads or purchases, and which are simply wasting impressions.
After launching, you should monitor your audience size and frequency. A small segment can quickly receive too many repeat contacts.
Some automated Meta settings can expand their reach beyond the initial parameters when the system believes this will help drive results. While this expansion can be useful for acquisition campaigns, it can also change what a rigorous retargeting segment test actually measures.
Before launching, you should check how the selected audience settings work in the current version of your advertising account. If your goal is to analyze specific website visitors, you should understand whether the system can show ads to people outside of that group.
Campaign results should also be interpreted taking into account the chosen automation. Otherwise, the advertiser may believe that sales came from retargeting, even though some of the reach actually came from a new audience.
Meta Ads allows you to collect audiences from various sources, so retargeting isn't limited to website visitors. Businesses can work with their own customer base, Facebook and Instagram activity, video views, lead forms, and events transmitted from the website via Meta Pixel or other data sources.
When setting up, it's important not to lump all interactions into one large segment. Different users have different levels of interest, time of contact, and context of their previous interaction. Blog page audiences, pricing page visitors, and AddToCart users may require completely different advertising messages.
Website visitors remain one of the primary sources for retargeting. After properly configuring the data source, advertisers can create audiences based on website visits and specific actions. For example, they can separately target people who visited a specific service page, a catalog category, a product card, or a pricing page.
You can also segment audiences by events. A user who viewed a page is different from someone who submitted a form or initiated an order. This granularity helps build sequential advertising communication and avoid mixing different stages within a single Ad Set.
Website visitors can have completely different motivations. One person accidentally opened an informational article a few weeks ago, another researched the service's pricing today, and a third added an item to their cart and closed the checkout page. Combining them into a single audience would inevitably result in a generic advertising message.
In addition to the depth of action, the time elapsed since the visit should be considered. A user who visited the site yesterday typically remembers the company better than someone who last visited two months ago. Therefore, separate segments based on actions and contact time are often used in the operational structure.
Retargeting can leverage interactions that occur directly within Meta services. A user may not visit the website, but they may regularly view posts, visit a profile, engage with content, send messages, or respond to advertising.
These audiences are especially useful for companies that conduct a significant portion of their communications through social media. For example, a local business could target active Instagram users and show them an ad for booking a service. An online store could specifically target people who have engaged with its content but haven't yet visited the website.
When creating a campaign, consider the strength of previous engagement. A single post view and a conversation with the company show different levels of interest, so it's best to analyze and test them separately.
Video views allow you to segment users based on the depth of their engagement with the content. Someone who watched a few seconds might have accidentally lingered, whereas someone who watched the majority of the video typically indicates a more conscious interest in the topic.
The advertising structure can use audiences based on different viewership percentages, such as 25%, 50%, 75%, or 95% of the video. The specific option depends on the video's length, audience size, and campaign objectives. A high viewership percentage may be significantly more common for a short video than for a long, expert-driven video.
After viewing, you can continue communicating with other content: offer a case study, demonstrate a service, invite the user to a consultation, or direct them to a product page. Repeating the same video without changing the message is usually less informative for the audience.
Meta's built-in lead forms allow you to collect leads without requiring the user to visit your website. Form interactions can also be used for further audience segmentation. Of particular interest are people who opened the form but didn't complete their submission.
For such users, you can create an ad that addresses a potential objection or clarifies the offer. For example, you can highlight the consultation terms, cost, timeframe, a case study, or a description of the work process. Users who have already submitted a request should generally be excluded from the initial acquisition campaign unless re-display is specifically provided for by the campaign logic.
This separation reduces the risk of a situation where a person has already contacted a company, but the advertising system continues to persistently ask them to fill out the same form.
A company may use its own customer base to build a user audience, provided the processing and transfer of such data is legal and complies with applicable requirements. Typically, readily available identifiers, such as phone numbers or email addresses, are used, which the advertising system attempts to match with its users.
This segment is suitable for repeat sales, additional services, recapturing inactive customers, and excluding existing buyers from acquisition campaigns. In B2B projects, the customer base also helps separate existing customers from new potential leads.
Before loading the database, you should convert the data to the correct format and check whether the chosen scenario truly requires using a list. The smaller the initial audience, the more limited the available reach.
For online stores, one of the most obvious scenarios involves users who viewed a specific product or added it to their cart. These actions demonstrate a more significant commercial interest than simply visiting the homepage or reading informational material.
It usually makes sense to separate the AddToCart audience from regular catalog visitors. For them, you can use ads with product reminders, delivery information, availability information, or other incentives to help them complete their purchase.
However, customers with a Purchase event must be properly excluded from such a campaign. Otherwise, a person may continue to see ads for a product they've already ordered for several days, and the advertiser will waste budget on useless re-impressions.
Lookalike Audiences are created based on a source audience and are designed to find new users with similar characteristics. The source can be existing clients, customers, leads, or other high-quality segments that provide the advertising system with a clear targeting point.
This audience is linked to the retargeting infrastructure via source data, but Lookalike itself is aimed at attracting new people. Therefore, it makes more sense to distinguish it from classic retargeting, where ads are shown to users after their previous contact with the company.
Lookalike is often used in parallel: retargeting segments target existing audiences, while lookalike audiences expand the reach of potential customers. The results of these approaches should be evaluated separately.
The quality of retargeting directly depends on the data received by the advertising system. If events are transmitted with errors, the same action is recorded multiple times, or a purchase is not registered at all, audiences will be assembled incorrectly. In this situation, the advertiser begins to optimize the campaign based on a distorted picture.
Before launch, you should check data sources, key events, whether events correspond to actual user actions, and exclusions for those who have already converted. For projects with a website, you should also check the Meta Pixel, Events Manager, and, if necessary, server-side event delivery via the Conversions API.
Meta Pixel is placed on a website and transmits information about visitor actions to the advertising system. It can record page views, product card openings, form submissions, adding products to carts, and other events needed for audience creation and advertising analysis.
Simply installing the code doesn't mean the data is being collected correctly. You need to check which pages the Pixel is triggered on, what events are being sent, whether there are any duplicates, and whether the event corresponds to a real user action.
For example, the "Lead" event shouldn't trigger when a form is simply opened if the company only considers a submitted request a lead. This error inflates the number of conversions and simultaneously creates the wrong audience for subsequent retargeting.
Events describe specific visitor actions. The set depends on the website type and business model, so there's no point in including dozens of events without a clear purpose. It's much more useful to accurately convey a few actions that truly correspond to the funnel stages.
Commonly used are PageView for viewing a page, ViewContent for meaningful content or product viewing, Lead for a lead, AddToCart for adding to cart, InitiateCheckout for starting a checkout, and Purchase for completing a purchase.
After setup, each event should be tested using a real user scenario. If a Purchase event is sent before the actual order is placed, and a Lead event is recorded multiple times during a single request, retargeting segments and reports will be inaccurate.
For an online store, events typically replicate the main stages of a purchase. A user first opens a product page, then adds it to their cart, initiates checkout, and completes the order. This sequence helps segment visitors based on their level of purchase readiness.
The basic logic might look like this:
If a user reaches AddToCart but no Purchase event follows, they can be treated as a separate win-back segment. If the purchase has already been recorded, the user is excluded from cart abandonment ads and, if necessary, transferred to a repeat sales campaign.
Events should correspond to real site actions, so a universal set should not be used without checking the specific structure of the store.
A service website typically has a different workflow. The user reviews the service page, checks the price or terms, opens the form, and submits a request. In some projects, several steps occur between the request and the actual sale, which are then recorded in the CRM.
The working diagram may look like this:
View service → view cost → open form → submit request → qualified lead.
For advertising, it's especially helpful to understand the difference between a simple page view and a real request. If these actions are mixed, the audience quickly becomes too broad, and campaign metrics lose touch with actual requests.
If you have a CRM, it's worth comparing advertising leads with subsequent movement through the funnel, since the same initial lead cost can conceal completely different quality of inquiries.
The Conversions API transmits events server-side and can be used in conjunction with browser-based Meta Pixel signals. This approach is particularly useful for projects that require more robust data transfer and better control over server-side events related to requests, orders, or other target actions.
The setup requires careful deduplication. If the same action is simultaneously transmitted through the browser and the server as two independent events, the advertising system may receive distorted statistics. Therefore, the technical implementation must take into account event identifiers and the chosen website architecture.
Personal data processing requirements and user consent must also be met. Server data transfer does not replace a business's legal and technical obligations to properly handle user information.
Events Manager is used to check data sources, events, and potential transmission issues. Before launching a campaign, it's helpful to ensure that the required events are actually being received, have the expected parameters, and trigger on the actions they were configured for.
Particular attention is required after website changes. A new form, a different checkout, a relocated thank-you page, or a redesigned catalog can disrupt the old logic of events, even if advertising campaigns continue to run.
It's best to test using real-world scenarios: open the desired page, add a product, fill out a test form, or follow a different target path. This way, you can spot errors before they impact your audience and advertising budget.
Segmentation helps connect a user's previous experience with a specific advertising message. The more traffic and data a project collects, the more options for audience segmentation become available. However, segmentation must make practical sense and provide sufficient volume for ad delivery.
It's best to start with two parameters: contact recency and depth of engagement. Later, you can add service type, product category, source of initial visit, customer status, and other factors that truly influence the decision.
Contact freshness affects how well a user remembers a company and their previous interest. Therefore, audiences are often divided into shorter and longer timeframes.
A rough structure might look like this: 1-7 days for recent visitors, 8-30 days for users with relatively recent contact, and 31-90 days for older contacts. These intervals serve as a starting point for testing, not a universal rule.
If the purchase cycle lasts one day, a long period can result in many irrelevant users. If a customer spends a month choosing a service, too short a window will lose some potential customers.
Not all visits reflect the same level of commercial interest. Reading an informational article is usually further from making a purchase than checking the price, opening a form, or adding an item to a cart.
Depth can be represented as a sequence:
or for the store:
Category → Product Card → AddToCart → InitiateCheckout → Purchase.
This structure helps determine which message is appropriate for the next step. An article visitor can be shown expert content or a service, while a user who hasn't completed the checkout process can be reminded about the purchase and order terms.
If a company offers multiple services, the overall website audience is often too diverse. For someone who has studied SEO, there's no point in showing an ad for mobile app development just because both services are hosted on the same domain.
Segmentation can be based on URLs, categories, events, or other characteristics that reflect a user's specific interests. In an online store, this could be product categories, in a clinic, medical specialties, or in an agency, specific services.
With low traffic, too much granularity can sometimes create audiences that are too small. In this situation, similar categories can be combined while maintaining a clear, coherent offering.
The actual structure will vary from site to site, but the general logic might look like this:
| Segment | What the user did | What to show |
|---|---|---|
| Content | Read a topical article or guide | Related expert material, case study, or service page |
| Service | Studied a specific service page | Benefits, work process, case study, or clear offer |
| Price | Looked at the cost or rates | Conditions, proof of results, responses to objections |
| Lead | Opened the form but did not complete the submission | Reminder, consultation or additional information |
| Cart | Added an item to cart without purchasing | Product, delivery terms, availability and return to checkout |
The table illustrates the principle, not a ready-made setup for any business. Before launching, you should review the actual customer journey and select only those segments for which a dedicated advertising message can be developed.
The retargeting budget depends on the audience size, cost per result, window duration, and the overall promotion structure. A large budget alone won't generate new users in an existing segment.
If the audience is small, the system may show ads to the same people more often. Therefore, the budget should be evaluated in conjunction with frequency, reach, number of conversions, CTR, and cost per result.
When there are several segments, it's best to prioritize them in advance. Users with incomplete checkouts typically require different attention than those who visited an informational blog in the last three months.
A retargeting audience is limited by the number of people who meet the specified criteria. If a website receives several hundred qualified visitors in a week, increasing the budget won't convert this segment into tens of thousands of new users.
If your budget is excessive, your ads may return to the same users more frequently. As a result, frequency increases, audiences tire of your creative faster, and additional impressions stop generating comparable conversion gains.
Therefore, the budget should be aligned with the actual audience size and its renewal rate. For sites with low traffic, it's sometimes more beneficial to develop the top of the funnel first, and use retargeting as an additional layer.
Frequency shows the average number of ad impressions per person over a selected period. The number itself can't be assessed without context, as different products, time windows, and creatives support different contact frequencies.
If frequency increases simultaneously with a decline in CTR and an increase in CPA or CPL, the audience may be becoming fatigued with the advertising message. In this situation, consider the creative, budget, segment size, and user inclusion window.
There's no universally safe value for all campaigns. Frequency should be analyzed alongside actual conversions, rather than stopping ads just because the metric reaches an arbitrary number.
It's time to rethink your creative when audiences see the same message for extended periods, frequency increases, CTR gradually declines, and the cost per result rises. This combination of signs often indicates that the current campaign is losing traction.
You can change the visuals sooner if the offer is outdated, the product is out of stock, or the landing page information has changed. A separate creative is also needed for a new segment at a different stage of the funnel.
Don't change your ads every day without sufficient data. Frequent, random edits make it difficult to compare results and determine which version actually performed better.
Dynamic retargeting is especially useful for businesses with a large number of products. Instead of manually creating a separate ad for each item, the system uses catalog data and displays relevant products to the user based on a configured scenario.
For proper operation, the catalog, site events, and product IDs must be consistent. If the event ID doesn't match the product in the catalog, the system won't be able to reliably associate the user's view with the correct item.
This approach reduces manual work when dealing with a large product range, but requires high-quality technical setup. Errors in the product feed or events directly impact what the user sees.
When a user opens a product card, the site transmits a corresponding event and item ID. The advertising system links this action to the catalog and can use the product in subsequent advertising for the appropriate segment.
The logic is as follows:
Catalog → product view → ViewContent event → audience → product ad → return to site.
For abandoned carts, the chain may take AddToCart into account, while customers with Purchase are typically excluded from the unfinished order return scenario.
Before launching, you should check the quality of product data, links, images, the presence and correspondence of identifiers between the site and the catalog.
The main use case is online stores, marketplaces, and other projects with large product catalogs. With hundreds or thousands of items, manually creating individual ads for each view is virtually impossible.
A dynamic format is also useful where product offerings change regularly and user interest is tied to specific cards. With a small number of services, manual segmentation and individual creatives sometimes provide a more understandable structure.
The solution depends on the site's architecture and data. If the catalog is populated with errors, events are not transmitted correctly, and products are updated infrequently, the technical foundation should be addressed first.
Evaluation should be based on business results, not just clicks and impressions. A high CTR may look good in your ad account, but it doesn't reflect how many leads turned into sales or how much revenue the campaign generated.
The set of metrics depends on the project. An online store typically analyzes purchases, order value, and ROAS, while a service website typically analyzes leads, cost per lead (CPL), inquiry quality, and subsequent conversion into customers.
It's also useful to compare segments. A seven-day AddToCart audience and ninety-day blog visitors initially have different levels of purchase readiness.
CTR measures the share of clicks relative to the number of impressions and helps understand how well the audience responds to an ad. This metric is useful for comparing creatives within similar contexts, but it doesn't replace the assessment of final results.
A high CTR can be accompanied by expensive or low-quality leads. For example, an attractive ad creative receives many clicks, but the landing page doesn't meet user expectations.
Therefore, CTR should be considered alongside conversions, cost per result, and post-click behavior. For retargeting, tracking changes in this metric as frequency increases is especially useful.
CPA reflects the cost of a target action, while CPL is typically used to estimate the cost per lead. These metrics help compare advertising segments based on the actual number of results received.
However, a cheap lead isn't always more profitable than an expensive one. If one segment generates many untargeted inquiries, while another generates fewer inquiries with a higher probability of purchase, the decision should be made based on further sales stages.
For services, it's advisable to transfer lead quality data from your CRM or at least regularly compare advertising campaigns with sales department results.
Conversion Rate shows what proportion of users completed the desired action. In retargeting, it's useful when comparing audiences with different levels of previous engagement.
For example, AddToCart users may logically show different conversion rates than visitors to informational content. This comparison helps us understand which segments truly require a separate budget.
The metric should be calculated consistently and use the same logic for comparable campaigns. Changing the definition of conversion mid-test makes the conclusions less reliable.
ROAS measures the ratio of revenue attributed by the advertising system to the promotional costs. This metric is especially useful for e-commerce, where the purchase price is transmitted along with the Purchase event.
When analyzing, it's important to consider the quality of data transfer and the chosen attribution model. If some orders aren't transferred or the order value is incorrect, ROAS will also be distorted.
In addition to advertising revenue, businesses need to consider margins, returns, order processing costs, and other expenses. A high advertising ROAS doesn't guarantee the profitability of a particular product.
The advertising account displays results based on its attribution logic, but the actual customer journey can involve multiple channels. A user might see an ad, then return through organic search, open an email, and only then make a purchase.
Therefore, for a serious analysis, it's useful to compare Meta Ads data with CRM, web analytics, and actual sales. This is especially important for B2B and services with long transaction cycles.
If your advertising system reports a large number of leads, but your sales team isn't seeing any corresponding requests, you should first check your events, integrations, and conversion definitions.
Retargeting errors often arise even before an ad is launched. An incorrect event, an overly broad audience, or a lack of exclusions can create a situation where the campaign technically works but shows ads to the wrong people.
The most common problems are as follows:
Once a problem is identified, it's best to change one significant element at a time and test the results. Simultaneously changing the audience, creative, offer, and landing page complicates analysis and doesn't reveal which change actually impacted the campaign.
A stable structure starts with data. First, key user actions are identified, then their delivery is configured, and finally, audiences are created with clear conditions and exclusions. Only on this basis does it make sense to plan ads and budgets.
The workflow looks like this:
Data sources → events → audiences → time windows → exclusions → creatives → advertising campaigns → analytics → adjustments.
An illustrative funnel diagram might look like this:
| Funnel stage | Share of users |
|---|---|
| First visit | 100% |
| Interest in the service | 75% |
| View price | 55% |
| Form / basket | 35% |
| Request / purchase | 20% |
This diagram only illustrates the logic behind audience narrowing and does not represent statistics for a specific campaign. Actual proportions should be calculated based on website data, CRM, and advertising account information.
Once launched, the system needs to be regularly monitored. Audience composition changes, creatives fade, new pages appear on the site, and events can be disrupted by technical updates. Regular monitoring helps spot such issues before they significantly increase advertising costs.
There's no single window for all products. The appropriate window depends on the length of the transaction, the frequency of purchases, the traffic volume, and the specific user action.
An abandoned cart campaign often requires a shorter scenario than a complex B2B service that a customer spends weeks comparing. Therefore, audiences can be segmented by contact recency and analyzed separately.
The final decision should be made based on campaign data and the actual length of the customer journey.