The remarketing setup that converts (and the audiences that don't)

Remarketing is one of the most misused and simultaneously most underutilised tools in digital advertising. Misused, because many advertisers simply retarget everyone who ever visited their website without any meaningful selection. Underutilised, because a well-structured remarketing setup is one of the most powerful conversion instruments available to any advertiser. The difference lies entirely in which audiences you choose, how you segment them, which ones you exclude, and how you align your bidding strategy accordingly. This article explains exactly which audiences convert, which ones burn your budget without return, and how AdBrains uses its own AI technology to achieve structurally better results in this area.

What is remarketing and why is it so powerful?

Remarketing, also referred to as retargeting, means reaching people again who have previously interacted with your website, app, or advertisements. They have already seen your brand, they are already familiar with your offer, and that puts them in a significantly better position to convert than a completely cold visitor. This explains why remarketing audiences show higher conversion rates than prospecting campaigns in virtually every industry.

In Google Ads, remarketing audiences are used via RLSA (Remarketing Lists for Search Ads), Display remarketing, YouTube remarketing, and in Performance Max (PMax) campaigns as an audience signal. In Meta Ads, Custom Audiences work based on website visits (via the Meta Pixel or server-side events), video engagement, and Instagram or Facebook profile activity. Both platforms offer rich possibilities, but the quality of execution determines entirely whether remarketing becomes a profit machine or a budget drain.

The audiences that convert: segmentation is everything

Conversion rate by remarketing audience segment

Average industry benchmarks e-commerce & lead gen (2026)

Cart abandoners (1-7 days)

4.8%

Product page visitors (1-14 days)

3.1%

Previous buyers (upsell)

3.9%

All website visitors (30 days)

1.4%

All website visitors (90+ days)

0.6%

Lookalike without exclusions

0.4%

The biggest mistake in remarketing is using one broad audience: all website visitors from the past 30, 60, or 90 days. This sounds logical but is in practice a recipe for disappointment. Someone who visited your homepage for five seconds nine weeks ago has a completely different profile from someone who added three products to their cart yesterday and abandoned it at the last moment. Yet in a standard remarketing setup, both people are approached identically, with the same bid and the same ad copy.

The audiences that consistently show the highest conversion rates are the following:

• Cart abandoners (1-7 days): people who added a product to their cart but did not complete the purchase. This is the warmest segment that exists. They have demonstrated purchase intent and only dropped off at the final step. A specific message, potentially with urgency or a reminder, works powerfully here.
• Product page visitors (1-14 days): visitors who viewed one or more specific product pages but did not proceed to purchase. They are orienting but serious. Dynamic ads featuring exactly the products they viewed perform excellently in this segment.
• Previous buyers (upsell and cross-sell): customers who have already purchased from you know your brand, trust you, and are more likely to make a repeat purchase. For ToetsJeKennis.nl, this segment works particularly well: someone who has completed an exam preparation course is an ideal candidate for a follow-up or supplementary course.
• Form starters (lead gen): for lead generation campaigns, such as those run by Clima-Active.nl for air conditioning and heat pump quotations, visitors who started the contact form but did not complete it are an extremely valuable segment. They have actively shown interest and only need a small extra nudge.
• Video viewers (75%+ on Meta): on Meta Ads, people who have watched more than 75% of an ad video are a strong remarketing audience. They have consciously chosen to keep watching and are therefore well advanced in their orientation.
• Email list uploads: a customer list of existing clients or leads, uploaded as a Custom Audience in Meta or as Customer Match in Google Ads, typically delivers excellent results in upsell and reactivation campaigns.

The key across all these segments is the time window. A cart abandoner from yesterday is fundamentally different from a cart abandoner from twelve days ago. Always split your audiences into multiple time windows (0-7 days, 8-14 days, 15-30 days) and adjust your bidding strategy and ad copy per window. The shorter the time window, the higher the intent, and the more aggressively you can bid.

The audiences that burn your budget

Just as important as knowing which audiences perform is knowing which audiences you should actively exclude or not target at all. Here is an overview of the most common pitfalls:

• All website visitors without a time-window filter: a pool of everyone who has ever been on your site is too broad and too cold to remarket profitably. Intent is too diffuse, competition for this audience is high, and conversion rates are low.
• Recent converters who have not been excluded: someone who just made a purchase or submitted a quote request yesterday should not immediately be bombarded with an ad for the same product or service. This causes irritation and wastes budget. Automatically and immediately excluding converters is a basic requirement that is surprisingly often overlooked in practice.
• Lookalike audiences without further qualification: Lookalike and Similar audiences are useful for prospecting but do not belong in a remarketing structure. They are by definition cold and have had no brand interaction yet. Including them in a remarketing campaign dilutes the quality of your audience pool.
• Bounced visitors from irrelevant pages: someone who arrived on your homepage via an accidental search query, stayed for three seconds, and left has shown almost no intent. Pages with an average session duration below ten seconds are rarely valuable remarketing sources.
• Geographically irrelevant visitors: if you only operate in one country but have international visitors in your remarketing pool (which happens unnoticed on many websites), you are paying for impressions to people you can never serve as a customer.

A strong remarketing setup is just as much a matter of exclusions as it is of inclusions. Those who do not select will waste budget.

RLSA: remarketing in Google Search

One of the most powerful but least used remarketing methods in Google Ads is RLSA: Remarketing Lists for Search Ads. With RLSA, you add remarketing audiences to your search campaigns, not to show separate ads, but to adjust your bids for people who have already visited your website when they perform a relevant search again.

For E-4motion.com, the webshop for new electric folding bikes, this means concretely: someone who previously visited the product page of a specific e-bike model and then searches again for "buy electric folding bike" receives a higher bid via RLSA. The probability that this person now wants to convert is significantly higher than for a completely new searcher. By linking the RLSA audience to Smart Bidding and Target ROAS, the algorithm receives an additional signal that the conversion probability is higher, leading to more efficient budget allocation.

RLSA works best in combination with bid adjustments per audience layer. An "observation" setting lets you collect data without restricting your targeting, while "targeting" limits your bidding to people on the list. For most campaigns, observation is the right starting position, with targeted bid increases for the warmest segments.

Remarketing on Meta Ads: Custom Audiences and the right funnel

Remarketing funnel: from visitor to converting customer

Throughput per stage with a well-segmented RLSA approach (percent mode)

Website visitors

All visitors in the audience pool

Qualified segments

After time-window and behaviour split

Ad impressions

Reach after budget allocation and targeting

Clicks

Average CTR remarketing vs. prospecting

Conversions

Final result after landing page match

Average 4.2% conversion rate with a segmented remarketing approach

On Meta Ads, remarketing revolves around Custom Audiences. You can build audiences based on website visits (via the Meta Pixel or server-side tracking), app activity, video engagement, Instagram profile interactions, and uploaded customer lists. The power of Meta remarketing lies in the combination of visual format and precise segmentation.

An effective Meta remarketing structure for a lead generator like Clima-Active.nl looks like this: visitors to the quotation page who have not completed the form receive a carousel ad featuring social proof (customer reviews, installation guarantees) and a clear call-to-action. This is combined with an exclusion of people who have already successfully submitted the form. At the same time, existing customers (uploaded customer list) receive a different ad aimed at expansion, such as a heat pump as a complement to an existing air conditioning installation.

On Meta, the quality of conversion signals determines how well the system can reach the right people. Server-side tracking via a proprietary sGTM infrastructure ensures that more conversions are measured and returned to Meta, which structurally improves both audience building and ad optimisation. Advertisers who use server-side signals see an average of 23% more tracked conversions than advertisers who rely solely on the standard pixel.

How AdBrains AI automates remarketing

Remarketing sounds manageable when you think about it once, but in practice it is a dynamic system that constantly demands attention. Audiences grow and shrink, time windows expire, converters need to be excluded, bids need to be adjusted, and ad copy needs to match the visitor's stage in the funnel. Keeping up with all of this manually is labour-intensive and error-prone. This is precisely where the value of AdBrains' own AI technology lies.

The audience management automation system at AdBrains creates and refreshes PROD audiences, Incubator audiences, and RLSA audiences on a weekly basis automatically. This means the right time windows are always current, converters are excluded immediately without manual action, and new segments are created as soon as visit volume is large enough to draw statistically valid conclusions.

The multi-agent verification system plays a critical role here as well. Every change to an audience setting, whether an exclusion, a bid adjustment, or an audience swap, is reviewed by four independent AI agents before the change is applied. This prevents errors that occur regularly in manual management, such as accidentally excluding a valuable segment or activating an overlap that doubles budget without additional reach.

The automatic tCPA/tROAS optimisation then ensures that bids per audience segment are adjusted daily based on current conversion data. A cart abandoner from the past 48 hours automatically receives a higher bid than a product page visitor from two weeks ago. The account manager does not need to adjust this manually; the AI handles it based on the most recent conversion volumes and the margin targets per client.

The RSA improvement system finally ensures that ad copy is aligned with the audience stage. Someone in the cart abandoner audience sees an ad that plays on urgency and trust, while someone in an earlier orientation stage receives an informational message. The AI analyses Ad Strength scores per ad group and automatically rewrites ad copy with a POOR status, so that relevance for each audience layer remains optimal at all times.

Comparison: manual vs. AI-driven remarketing management

Manual vs. AI-driven remarketing management

Structural differences in approach and results

Manual management

• Audiences updated monthly or quarterly
• No automatic exclusion of converters
• Broad segments without time-window splits
• Bid adjustments set by gut feeling
• Campaign overlap goes unnoticed
• RSA copy not tailored to audience stage
• No RLSA integration with Smart Bidding

AdBrains AI managementRecommended

• Audiences automatically created and refreshed weekly
• Converters immediately and automatically excluded
• Segments split by time window and behaviour
• Bid adjustments driven by tCPA/tROAS data
• Campaign overlap detected and resolved
• Ad copy optimised per audience stage
• RLSA audiences integrated into Smart Bidding signals

The comparison above makes clear that the difference between manual and AI-driven remarketing management is not about one or two isolated improvements. It is a fundamentally different approach, in which every part of the funnel is structurally managed better. Manual management is by definition reactive; AI-driven management is proactive and continuous.

For clients like LeroyBrouwer.nl and Clima-Active.nl, who are strongly dependent on high-quality leads and therefore on precise audience selection, this distinction is directly visible in the cost per lead and the quality of incoming requests. The more precise the audience definition, the more relevant the advertisement, and the greater the probability that the visitor actually converts.

Best practices for a conversion-focused remarketing setup

Summarised below are the most important principles for a remarketing structure that converts consistently:

• Always segment by behaviour and time window, never by single broad categories.
• Exclude converters immediately and automatically from all active remarketing campaigns.
• Link RLSA audiences to Smart Bidding for maximum efficiency in search campaigns.
• Use server-side tracking to maximise the quality of your conversion signals.
• Align ad copy and creatives to the user's stage in the funnel.
• Actively monitor campaign overlap and resolve it before budget is doubled without additional reach.
• Review your audience structure at least monthly based on current conversion data.

Remarketing is not a set-and-forget tactic. It is a living system that requires constant attention, whether you manage it manually or through an AI-driven platform like AdBrains. The difference is how much of that attention you need to invest yourself, and how consistently that attention is applied.

Frequently asked questions about remarketing audiences

What is the difference between remarketing and retargeting?

The terms are often used interchangeably, but in practice retargeting more often refers to reaching visitors again via paid ads on external platforms (such as Meta or Display), while remarketing can also include email campaigns. In Google Ads terminology, the official term used is "remarketing." The principle is the same in both cases: reaching people again who have previously interacted with your brand.

How large does a remarketing audience need to be to work?

For Display remarketing and Meta Custom Audiences, you need at least 100 active users in an audience before ads are served. For RLSA in Google Search, a minimum of 1,000 users applies. For smaller advertisers, this volume requirement can make RLSA challenging to set up. In that case, you can extend time windows (from 7 to 30 days) or combine broad behavioural segments to reach the minimum threshold, but be careful not to lose too much intent signal in the process.

Should I always exclude converters from remarketing?

In most cases the answer is yes, but with nuance. Recent converters, people who just made a purchase or submitted a quote request, are excluded from campaigns targeting the same conversion action. But those same people can then be added to a separate upsell or cross-sell audience for a follow-up offer. At ToetsJeKennis.nl, someone who has just completed an exam course is excluded from the acquisition campaign but simultaneously added to a retention list for supplementary course modules. This prevents irritation and maximises the lifetime value of the customer.

Does remarketing work for B2B and lead generation too?

Absolutely. For lead generation, remarketing can be particularly powerful because the decision cycle is longer than in e-commerce. Someone who visited a quote page but did not convert is an extremely warm segment that typically needs several more touchpoints before the final enquiry. For Clima-Active.nl, active in air conditioning and heat pump installation, RLSA audiences on search queries such as "heat pump quote" or "air conditioning installer" are among the most profitable segments in the entire campaign structure. The cost per lead in remarketing segments is structurally lower than in prospecting, while lead quality is equal or higher.

How do I combine remarketing with Performance Max?

In Performance Max campaigns, you can add remarketing audiences as audience signals. This gives the PMax algorithm a strong signal about who your most valuable visitors are, after which it autonomously finds similar profiles across all Google channels simultaneously. It is important to monitor whether PMax is not leaning too heavily on your remarketing pool at the expense of prospecting. Use audience insights to analyse the mix and adjust via campaign objectives and budget allocation. A separate remarketing campaign alongside PMax can in certain cases be more efficient, because you retain full control over bids and ad copy per audience segment.