Lecture 16: Remarketing and RLSA: Remarketing Lists for Search Ads

Short answer: Remarketing Lists for Search Ads (RLSA) let you layer audience data from your website visitors, past customers, and CRM contacts onto your existing search campaigns. Instead of building brand-new campaigns, you use RLSA to bid more aggressively when a known, high-intent visitor searches again, bid down or exclude people who already converted, and unlock broader keywords for segments you already trust. Combined with display and video remarketing, dynamic product ads, and Customer Match, RLSA turns your search account from a keyword-only engine into a full-funnel, audience-aware system.

What You'll Learn in This Lecture

• What RLSA actually is and how it differs mechanically from standard display/video remarketing
• How to build remarketing lists from website visitors using tags, GA4, and app data
• How to segment lists by behavior: cart abandoners, past purchasers, high-intent page visitors, content readers
• How to use RLSA to bid up and expand keywords for warm audiences searching broader terms
• How to bid down or exclude converters so you stop paying full price to re-win existing customers
• How list duration and membership windows affect reach, relevance, and decay
• How to combine RLSA with Similar Audiences and Customer Match for a layered targeting stack
• The practical differences between display/video remarketing and search remarketing
• How dynamic remarketing with product-level ads personalizes creative automatically
• How to frequency cap remarketing so you build trust instead of annoyance
• How to measure remarketing's true incremental value instead of trusting last-click credit
• Common RLSA setup mistakes that quietly shrink your audience to nothing
• A practical rollout plan for adding RLSA to an existing search account this week

What Is RLSA and How It Differs From Standard Remarketing

Standard remarketing — the kind most advertisers learn first — shows display banners or video ads to people who previously visited your site, using the Google Display Network or YouTube. It is a targeting method layered on top of placements you don't control directly; the ad appears while someone reads a news article or watches a video, essentially interrupting their attention with a reminder of your brand.

RLSA is different in one crucial way: it doesn't add new inventory, it adds an audience layer on top of search inventory you already bid on. When someone who is on one of your remarketing lists types a query into Google, your existing search campaign can react to that fact. You are not showing an ad in a new place; you are changing how you compete for an ad you would have shown anyway. That distinction matters because search remarketing captures active intent (someone typing a query right now) rather than passive attention (someone scrolling a webpage), which is why RLSA audiences frequently convert at two to four times the rate of the general search population.

Mechanically, RLSA works through two settings: Observation and Targeting. In Observation mode, your ads keep serving to everyone who matches your keywords, but you can add bid adjustments and get reporting broken out by list membership. In Targeting mode, your ads only serve to people who are on the list AND match your keywords — this is how you build audience-only campaigns with much broader, even generic keywords, because the audience layer is doing the qualifying work that the keyword alone can't do.

Example: A software company sells project management software under branded keyword campaigns and a separate generic campaign for terms like "team collaboration tool." On its own, "team collaboration tool" is too broad and expensive to bid on profitably — most searchers are early-stage browsers. But when the company builds an RLSA-targeting campaign restricted to people who visited its pricing page in the last 30 days, that same broad term becomes highly profitable, because everyone eligible to see the ad has already shown strong purchase intent once.

Building Remarketing Lists From Website Visitors

Every remarketing list starts with a tag that fires on your site and records visitor activity against a Google Ads or Google Analytics identifier. There are three main technical paths, and most mature accounts use a mix of all three.

1. The Google Ads tag (global site tag / Google tag). This is the most direct method: install the base tag site-wide, then define audience lists directly inside Google Ads based on URL rules, page visits, or custom events. It is fast to set up and doesn't require Analytics, but the segmentation logic is comparatively basic — mostly URL-contains rules and simple event parameters.

2. GA4 audiences shared into Google Ads. Because GA4 tracks a full event model (page_view, add_to_cart, purchase, scroll depth, video engagement, custom events), you can build far richer audiences in GA4 — for example, "users who viewed at least 3 product pages in a session but did not purchase within 14 days" — and then export that audience into Google Ads as a remarketing list. This is the recommended approach for anything beyond simple page-visit segmentation, since GA4's audience builder supports sequences, exclusions, and multi-condition logic that the raw Ads tag cannot express.

3. App and CRM data. For advertisers with a mobile app, Firebase/GA4 for Firebase generates audiences from in-app events (app_open, level_complete, in_app_purchase). For advertisers with offline sales or a customer database, first-party data uploaded via Customer Match (covered later in this lecture) supplements web-based lists with people who never converted online at all.

Whichever path you choose, a few technical requirements are non-negotiable: your tag must fire consistently across desktop and mobile, respect consent-mode signals in regions with cookie consent laws, and reach a minimum list size before Google will serve ads against it — currently 1,000 members for search remarketing lists (vs 100 for display). Small sites with low traffic often can't build a cart-abandoner list that reaches 1,000 people in a reasonable window; in that case, broaden the membership rule (e.g., "all site visitors" instead of "cart page visitors") until volume clears the threshold, then narrow again as traffic grows.

Segmenting Lists by Behavior (Cart Abandoners, Past Purchasers, High-Intent Page Visitors)

A single "all visitors" list is a starting point, not a strategy. The real value of remarketing comes from splitting visitors into behavioral tiers so your bids and messaging match true intent. A practical tiering structure looks like this:

• Tier 1 — Cart/checkout abandoners: Visited a checkout or cart page but did not complete a purchase within a session. This is your highest-intent, highest-urgency segment; these people were one step from converting and often just got distracted, hit a shipping-cost objection, or needed to compare one more option.
• Tier 2 — Product/service page viewers: Viewed specific product, pricing, or service pages but never reached cart or checkout. Strong intent, but earlier in the decision process — good candidates for RLSA bid-ups on category and competitor terms.
• Tier 3 — Deep content engagers: Read multiple blog posts, watched a demo video, spent an extended session on the site, or downloaded a resource. Interested but not yet transactional; a good segment for RLSA-informed bid adjustments on brand and educational queries rather than aggressive commercial terms.
• Tier 4 — Past purchasers/converters: Already bought or already converted on the primary goal. This list exists mainly for exclusion or cross-sell/upsell targeting, not for standard acquisition bidding.
• Tier 5 — All visitors (catch-all): Anyone who touched the site, used mainly to hit minimum list-size thresholds or for a light, low bid-adjustment blanket layer.

Each tier deserves its own membership duration and its own bid adjustment, because a cart abandoner from yesterday behaves very differently from a cart abandoner from 25 days ago. Where volume allows, split cart abandoners further into "abandoned in the last 3 days" (aggressive bid-up, urgency messaging) versus "abandoned 4-14 days ago" (moderate bid-up) versus "abandoned 15-30 days ago" (light bid-up, since intent has likely cooled or the buyer purchased elsewhere).

Example: An online furniture retailer builds four lists: cart abandoners (3-day window), sofa-category viewers (14-day window), all-site visitors (30-day window), and past purchasers (540-day exclusion window for cross-sell). On the generic search campaign for "modern sofa," it applies a +60% bid adjustment for cart abandoners, +25% for category viewers, and a -100% exclusion for past purchasers of sofas specifically — while leaving them eligible for a separate rug and decor campaign as a cross-sell audience.

Using RLSA to Bid Up on Past Visitors Searching Broader Terms

The single highest-leverage RLSA tactic is using warm audiences to justify bidding on keywords that would otherwise be too broad, too expensive, or too low-intent to run profitably. The logic: a generic keyword's average conversion rate blends everyone who searches it, from casual researchers to ready buyers. When you restrict — or heavily bid up — that keyword for people already on a remarketing list, you are effectively filtering out the low-intent share of the query and keeping only searchers you already know are warm.

Two structural approaches accomplish this:

Bid adjustments in Observation mode. Keep your existing campaign structure and keyword set exactly as is, then layer remarketing lists as an audience criterion set to Observation, with positive bid adjustments (commonly +20% to +100%, sometimes higher for high-value cart abandoners). This is the lower-risk approach: your existing traffic and volume are untouched, and you simply pay more to win auctions when a warm visitor searches.

Dedicated RLSA campaigns in Targeting mode. Build a new campaign using the same or broader keyword list, but restrict the audience to Targeting mode so ads only show to list members. This lets you bid on head terms and category terms that you would never run as open, unrestricted keywords because the cost per click or conversion rate wouldn't clear your threshold for a cold audience. Because the audience is pre-qualified, you can also write more assumptive ad copy ("Still deciding? Here's 10% off" or "Continue your order") that would feel presumptuous to a first-time searcher.

A common rollout sequence is to start in Observation mode to gather data on how each list performs against your existing keywords, confirm the lift is real (not just correlation with people who were already going to convert), and only then graduate the best-performing segments into a dedicated Targeting-mode campaign with expanded keyword coverage.

Using RLSA to Bid Down or Exclude Converters

The mirror image of bidding up warm prospects is making sure you stop paying acquisition-level prices to "reconvert" people who already bought. This is one of the most overlooked levers in mature accounts — teams obsess over acquiring new remarketing bids but forget to protect against wasted spend on their own existing customers.

Build a "converters" or "customers" list from your purchase/lead confirmation page, then apply it to your search campaigns in one of two ways depending on your business model:

• Hard exclusion: For one-time purchase products (a mattress, a one-off service, a single online course) where repeat purchase is rare, exclude converters entirely from the campaign. There is no reason to keep bidding on "mattress store near me" for someone who bought a mattress from you last month.
• Bid-down, not exclusion: For subscription businesses, repeat-purchase retail, or multi-product catalogs, a full exclusion can backfire — the same customer might be a great candidate for renewal, upsell, or a different product line. Instead, apply a negative bid adjustment (-30% to -90%) so you remain present but pay much less to reach them, or route them into a separate, cheaper cross-sell campaign instead of your acquisition campaign.

This tactic protects both your cost-per-acquisition metrics and your overall account health: without it, branded search terms in particular become artificially inflated in cost as you compete against your own bidding for customers who were always going to navigate straight back to you regardless of the ad.

List Duration and Membership Windows

Membership duration is the number of days someone stays on a list after triggering the qualifying action, and it is one of the most consequential — and most neglected — settings in remarketing. Google Ads allows durations up to 540 days, but the right number depends entirely on your sales cycle and the behavior being tracked.

Guidelines by scenario:

• Cart/checkout abandoners: 3 to 14 days. Intent decays fast; someone who abandoned a cart three weeks ago has likely already bought elsewhere or moved on entirely, so keeping them on an aggressively bid-up list past that window mostly wastes budget.
• Product/category page viewers: 14 to 45 days, matching a typical consideration window for the product category. A $30 impulse item warrants a short window; a $30,000 B2B software purchase warrants a much longer one.
• Content/blog readers: 30 to 90 days, since these visitors are earlier in the funnel and take longer to become sales-ready.
• Past purchasers (for exclusion or cross-sell): 180 to 540 days, or permanent if you maintain the list via Customer Match instead of a decaying pixel-based list.

Remember that remarketing lists are rolling windows, not fixed cohorts — a person re-enters the clock every time they repeat the qualifying action, and they silently age out once the membership duration passes without a repeat visit. This means list composition is always in flux: check list size trends periodically, since a sudden shrinkage often signals a tracking break (a tag removed during a site redesign, a consent banner blocking cookies) rather than an actual drop in visitors.

Combining RLSA With Similar Audiences and Customer Match

RLSA becomes significantly more powerful once you stop treating it as a single lever and start layering it with other first-party and modeled audience types.

Customer Match lets you upload hashed customer data — emails, phone numbers, mailing addresses — directly into Google Ads, independent of any website pixel. This closes two major gaps that pixel-based remarketing can't reach: people who bought offline or through a channel without tracking, and people who bought so long ago that their pixel-based membership expired. A Customer Match list of "customers in the last 3 years" can serve as a permanent, durable exclusion list for search campaigns, immune to cookie consent issues or ad blockers that can degrade pixel-based lists.

Similar Audiences (where still available in your account/market — Google has been phasing this feature in and out of different campaign types) let Google find new users who share behavioral patterns with an existing list, without requiring that new user to have visited your site. Historically used mainly on Display and YouTube, similar-audience-style modeled expansion increasingly happens automatically inside Performance Max and broad-match-plus-Smart-Bidding combinations, meaning the strategic move today is less about manually selecting a "Similar to Cart Abandoners" list and more about feeding Smart Bidding rich, well-labeled conversion signals (via Customer Match and refined remarketing lists) so its automated audience expansion has good source data to learn from.

Example: A subscription meal-kit company combines three layers on one campaign: an RLSA list of trial-signup page visitors (bid +40%), a Customer Match list of lapsed subscribers who canceled 60-180 days ago (separate win-back campaign with tailored copy), and a hard Customer Match exclusion of active subscribers so it never pays to "acquire" someone already paying them monthly.

Display and Video Remarketing vs Search Remarketing

It's worth being explicit about how these channels differ, since teams often conflate them or assume one subsumes the other.

• Inventory and intent: Search remarketing (RLSA) only fires when the user actively searches — it rides on top of expressed intent. Display and video remarketing fire based on the user's browsing or viewing context, regardless of whether they're currently looking for your product — it's interruption-based, not intent-based.
• Creative format: Search remarketing uses your existing text/responsive search ads; there's no separate creative to build. Display and video remarketing require dedicated banner sizes, responsive display assets, or video creative, and dynamic remarketing (below) requires a product feed.
• Cost and volume: Display and video impressions are typically far cheaper per impression but convert at a lower rate; search remarketing clicks cost more but convert at a materially higher rate because of active intent.
• Best use in the funnel: Use display/video remarketing to stay top-of-mind during the consideration window with minimal cost, and use RLSA to capture the moment that consideration turns back into an active search — the two are complementary stages of the same customer journey, not competing budgets.

A well-built account runs both simultaneously: a low-cost display/video remarketing campaign keeps the brand visible, while RLSA ensures that the moment a warm visitor searches again, the search campaign is bidding aggressively enough to win that click before a competitor does.

Dynamic Remarketing With Product-Level Ads

Dynamic remarketing extends the basic remarketing concept by automatically showing the exact products or services a visitor viewed, rather than a generic brand banner. This requires two components: a product feed (the same type of feed used for Shopping campaigns, covered in the next lecture) uploaded to Google Merchant Center or directly to Google Ads, and dynamic remarketing tag parameters that pass product IDs, page type, and other identifiers back to Google as visitors browse.

Once configured, dynamic remarketing can assemble ads on the fly: a visitor who viewed a specific pair of running shoes and a specific jacket sees a carousel ad featuring those exact items, often with the price and stock status pulled live from the feed. This dramatically outperforms static remarketing banners for catalog-heavy businesses (retail, travel, real estate, job boards) because the ad content is directly relevant to what the individual actually looked at, rather than a generic message hoping to be relevant to everyone on the list.

While dynamic remarketing itself runs on Display and YouTube inventory (not classic text-based search ads), it pairs naturally with RLSA: the same product-level feed and viewing data that powers a dynamic display carousel can also define the remarketing list used to bid up related search terms, keeping messaging and targeting consistent across both channels for the same visitor.

Frequency Capping and Avoiding Ad Fatigue

Because remarketing repeatedly targets the same people, it is uniquely prone to over-exposure. Seeing the same ad ten times a day reads as desperate or invasive rather than helpful, and it measurably depresses click-through and conversion rates over time — a phenomenon usually called ad fatigue or banner blindness.

Frequency capping controls this directly on Display and video remarketing campaigns, letting you set a maximum number of impressions per user per day, week, or month. There is no universal right number, but many advertisers find 3-7 impressions per user per day across a campaign, or roughly 15-20 per week, keeps visibility high without tipping into annoyance — though the right cap depends heavily on product price point and consideration length (a $15 impulse buy warrants a much lower cap than a $15,000 B2B service with a long sales cycle).

Search remarketing (RLSA) is inherently self-limiting on frequency, since it only serves when the user actively searches — you can't over-serve someone who isn't searching. But you should still manage fatigue at the creative level: rotate ad copy and, on Display/video, rotate creative assets on a defined schedule (every 2-4 weeks is common) so the same exact banner or video isn't shown for months on end. Pair frequency capping with sensible list durations (covered above) — a list that never expires membership is a common hidden cause of fatigue, since people keep seeing ads long after their interest realistically would have faded.

Measuring Remarketing's True Incremental Value

The hardest and most important remarketing question is not "did this convert" but "would this person have converted anyway, without the ad?" Remarketing audiences are, by definition, people who already showed strong intent — many would return and buy directly, via a bookmark, a branded search, or simply typing the URL, regardless of whether they saw a remarketing ad. Last-click and even most multi-touch attribution models tend to overstate remarketing's contribution because they credit whichever touchpoint happened last, and remarketing ads are structurally positioned to be that last touchpoint.

To estimate true incremental value, use one or more of these methods:

• Conversion lift studies / ghost ads: Google Ads' built-in Conversion Lift measurement runs a holdout: some eligible users see your remarketing ad, an equivalent group sees a public service ad or nothing instead, and the platform compares conversion rates between groups. This isolates the causal effect of the ad rather than relying on attribution modeling.
• Geo holdout tests: Pause remarketing (or a specific segment of it) in a subset of matched regions while running as normal elsewhere, then compare conversion rate or revenue trends between the holdout and control regions over several weeks.
• Frequency/exposure comparison: Compare conversion rates of users who saw the remarketing ad zero times vs. once vs. multiple times, controlling as much as possible for the fact that more-engaged users are naturally more likely to both see more ads and convert more — a useful directional signal, though weaker than a true experiment since it isn't randomized.
• Data-driven attribution in GA4/Google Ads: Rather than last-click, use a data-driven model that distributes credit across the full path, reducing (though not eliminating) the tendency to over-credit the final remarketing touch.

In practice, most advertisers find that remarketing delivers real incremental value, but meaningfully less than raw last-click numbers suggest — often somewhere in the range of 20-60% of reported conversions being truly incremental, with cart abandoners and high-intent segments showing higher genuine lift than broad "all visitors" lists. Use this understanding to set realistic ROAS/CPA targets for remarketing campaigns rather than judging them by the same standard as pure cold-traffic acquisition, while still holding them accountable to a measured lift rather than assuming all credit is deserved.

Bringing It Together: A Rollout Plan

If your account doesn't yet use RLSA, a practical first-week rollout looks like this: (1) confirm your Google tag or GA4 sharing is active and audiences are populating; (2) build four core lists — all visitors, category/product viewers, cart abandoners, and past purchasers/converters; (3) add all four as Observation-mode audiences on your best-performing existing search campaigns with modest bid adjustments; (4) after two to three weeks of data, identify which segments show a real, defensible lift and graduate the strongest into a dedicated Targeting-mode campaign with broader keywords; (5) add a converter exclusion or bid-down layer everywhere it hasn't been applied; (6) layer in Customer Match for durable, pixel-independent audiences; and (7) schedule a conversion lift study or geo holdout once volume supports it, so remarketing's budget is justified by measured incrementality rather than last-click assumption. This same audience infrastructure — lists, exclusions, and behavioral segments — carries directly into Shopping campaigns, the subject of the next lecture, where product feeds and dynamic remarketing data become even more central to how ads are built and targeted.

Related Lessons Across SEO, AEO, GEO, SEM, and PPC

Use these connected lessons to move through organic search, answer engines, generative AI visibility, paid search, and PPC without losing the bigger strategy.

• Lecture 1: What Is SEM? How Search Engine Marketing Works in 2026 (SEM) - return to the course foundation when you need the big picture.
• Lecture 1: What Is PPC? How Pay-Per-Click Advertising Works (PPC) - separate SEM strategy from PPC execution.
• Lecture 38: PPC Strategy Roadmap: Bringing Search, Social, Retail, and AI Together (PPC) - connect SEM with the full PPC channel roadmap.
• Lecture 4: Keyword Research Fundamentals (SEO) - connect organic keyword research with the same demand signals.
• Lecture - 7: Prompt Research: Finding What People Ask AI Tools About Your Topic (GEO) - turn search queries into AI prompt research.