Remarketing Lists for Search Ads: What Actually Changed in 2026, and Why Your Bid Adjustments Might Be Doing Nothing

For most of the last decade, Remarketing Lists for Search Ads sat just out of reach for a huge share of advertisers. The feature itself was never complicated — show a different ad, or bid differently, to someone who’s already been to your site and is now searching again. The problem was the gate in front of it: Google required a list to hold at least 1,000 active users before it could be used on Search at all. Local service businesses, niche B2B software, low-traffic e-commerce stores — most of them simply never got there. RLSA existed on paper and was invisible in practice for a huge share of the accounts that could have benefited from it most.

That gate came down in December 2025, when Google finalized a change lowering the minimum audience size to 100 users across Search, Display, and YouTube, for both remarketing lists and customer lists. It’s one of those changes that didn’t get much attention outside PPC circles, but it quietly rewired who gets to use one of the higher-leverage tools in Google Ads. If your account never had enough traffic to build a usable remarketing list before, there’s a decent chance it does now.

At the same time, a lot of accounts that already had RLSA running are sitting on a different, less visible problem: bid adjustments on audience segments that look configured correctly in the interface and do absolutely nothing, because the campaign switched to Target CPA or Target ROAS at some point and nobody went back to check whether the old settings still mattered. Google doesn’t warn you when this happens. The adjustment stays visible, the percentage stays set, and it just stops functioning as a lever.

This piece covers both sides of that: what RLSA actually is and how the mechanics work in 2026, and where most of the effort people put into “audience strategy” is quietly wasted because it’s aimed at a control Smart Bidding no longer listens to.

What RLSA Actually Does

Remarketing Lists for Search Ads lets you layer audience data on top of your existing Search campaigns. It doesn’t create new traffic and it doesn’t change which keywords trigger your ads on its own — someone still has to search a term you’re bidding on. What it changes is what happens once they do, based on whether that specific person has interacted with your site before and, depending on which list, what they did while they were there.

The reason this is more useful than it sounds is that it combines two signals that are individually weaker than they are together. A search query tells you what someone wants right now. A remarketing list tells you they already know who you are. Someone searching your category term for the first time and someone searching that same term after already comparing three product pages on your site last week are not the same lead, even though the keyword report treats them identically. RLSA is the mechanism that lets you treat them differently — different ad copy, a different bid, or in some setups, a completely separate campaign structure built only for people who’ve already been exposed to your brand.

Under the hood, RLSA runs on the same audience infrastructure as everything else in Google Ads’ Audience Manager: remarketing lists built from site visitor tags, Customer Match lists uploaded from your own customer data, and Google’s own affinity, in-market, and life-event segments. What makes it “RLSA” specifically is where you apply that audience data — inside a Search campaign or ad group, rather than Display, YouTube, or Performance Max, where those same lists can also be used.

How RLSA Differs From Ordinary Display Remarketing

It’s worth being precise about this, because the two get conflated constantly, including by people running both. Ordinary Display remarketing shows banner or responsive ads to people from your list as they browse other sites and apps — it’s a proactive, interruption-based format. You decide when someone sees your ad again based on where they are and what they’re doing, not because they’ve told you anything about current intent.

RLSA is reactive. Nothing happens until the person on your list actively searches something you’re bidding on. That single difference changes the entire economics of the audience. A Display remarketing impression reaches someone who may or may not be thinking about your category at that moment. An RLSA-triggered impression, by definition, only happens when someone from your list is actively searching — meaning intent and familiarity are both present at the same time, which is a genuinely rare combination and the reason RLSA campaigns often post stronger conversion rates than either cold Search or Display remarketing on their own.

The practical implication is that the two shouldn’t be treated as competing tactics or a simple either/or budget decision. They cover different moments in the same customer’s journey — Display remarketing keeps you visible while someone isn’t actively looking, and RLSA sharpens what happens on the occasions they come back and search directly. Most accounts running both get more out of them together than either alone, and the lists themselves are frequently interchangeable between the two, since a well-built remarketing list works equally well whether it’s deployed on Display or layered onto a Search campaign.

Observation vs. Targeting: The Setting That Decides Everything

When you add an audience segment to a Search campaign, Google asks you to choose between two modes, and the choice changes what the audience actually does.

Observation mode is the default and the safer starting point. Your campaign keeps bidding on its keywords exactly as it would without the audience attached, reaching everyone who searches those terms regardless of whether they’re on your list. The audience segment just gets layered on top for reporting and, if you’re on Manual CPC, optional bid adjustments. You can see how people from that list perform compared to everyone else without giving up any reach to find out.

Targeting mode is the one that actually restricts who sees your ad. With targeting mode selected, your keywords will only trigger for searchers who are also members of the audience segment you’ve attached. This is what most people picture when they hear “RLSA” — a dedicated campaign or ad group that exists purely to catch people who’ve already visited your site, often bidding more aggressively on broader or more expensive terms than you’d ever justify for cold traffic, because you know something about this specific searcher that changes the math.

The practical pattern most accounts land on is starting every audience in observation mode to see what the data actually shows before restricting anything. If a past-visitor segment converts meaningfully better than your account average on a given keyword theme, that’s the signal to consider carving out a targeting-mode campaign specifically for that combination, rather than assuming it’ll work before you’ve checked. Reversing that order — building targeting-mode campaigns from a hunch and finding out months later they never had the volume or the performance to justify existing — is a common way accounts end up with structural bloat nobody wants to clean up. If you’re trying to figure out whether a given audience or keyword theme is actually earning its keep, that’s the same kind of digging reading a search term report properly is built for — the two reports tell you related but different halves of the same story.

The Segment Types Actually Worth Building

Not every audience segment available in Audience Manager is equally useful for Search. A handful do most of the real work:

• Website visitor remarketing lists. The foundation — built from your own site tag, segmented by the pages people actually visited. A generic “all visitors” list is the weakest version of this. Cart abandoners, pricing page visitors, people who viewed a specific product category, and people who reached a thank-you page but didn’t convert on a second, higher-value action are all meaningfully different audiences with different intent, and treating them as one blended list throws away most of the signal.
• Customer Match lists. Built from your own first-party data — email addresses, phone numbers, or mailing addresses you already have, hashed and uploaded to Google. These let you do things remarketing tags can’t, like excluding existing customers from acquisition campaigns or bidding up on searches from people already in your CRM as high-value accounts. If you haven’t built one recently, what actually changed with Customer Match this year is worth reading alongside this, since the two features are frequently used together — a Customer Match list segmenting who someone is, layered onto an RLSA structure controlling what happens when they search.
• In-market and affinity segments. Google-built audiences based on browsing and purchase-intent signals across the web, not tied to your own site data. These are weaker as a standalone RLSA lever on Search specifically, since they don’t confirm the person has ever interacted with your brand, but they’re useful layered as observation data or fed into Performance Max audience signals.
• Combined segments. Built using AND/OR/NOT logic across your other lists — for example, people on your remarketing list AND not on your customer-match “already purchased” list, which is a cleaner way to isolate genuine prospects than trying to build that exclusion logic through campaign structure alone.

One category that’s mostly gone: Similar Audiences, which Google deprecated for both targeting and reporting back in 2023. Anyone whose account still has strategy documentation referencing “similar audiences” as a lever is working from an outdated playbook — Google replaced the concept with optimized targeting and audience expansion, which work by algorithmically widening delivery around a seed audience rather than giving you a discrete list to target directly. Those features exist mainly in Display, Video, and Performance Max rather than as a Search-specific RLSA tool, which is one more reason a clean, well-segmented set of your own remarketing and Customer Match lists matters more for Search than chasing Google’s automated audience-expansion features.

The 100-User Threshold, and Who It Actually Changes Things For

Before December 2025, the practical minimum for any list you wanted to use on Search was 1,000 active users in the trailing 30 days. That threshold quietly excluded a large share of advertisers from RLSA entirely — not because the feature wasn’t worth using, but because their traffic never grew a list fast enough to clear the bar. A local service business getting a few hundred visits a month, a B2B tool with a long, considered sales cycle and low search volume, a new e-commerce store still building traffic: all of them were locked out by list size long before they got anywhere near strategy questions like observation versus targeting mode.

Google’s finalized change drops that floor to 100 active users across Search, Display, and YouTube, for both remarketing lists and Customer Match lists. That’s a genuinely different situation for smaller accounts. A list that would have taken the better part of a year to reach 1,000 visitors can clear 100 in a matter of weeks for a lot of businesses, which means RLSA becomes a realistic tool for exactly the kind of accounts that historically wrote it off as “something bigger advertisers do.”

The catch is that a 100-user list is a genuinely small sample, and it’s worth treating it that way rather than expecting it to behave like a mature audience. Performance data on a segment that thin is noisy — a handful of conversions either way can swing your read on whether the audience is working. The sequencing that makes sense for smaller accounts newly eligible for this is the same one larger accounts should already be following: start in observation mode, let the list grow past the bare minimum before drawing conclusions from its performance, and resist the urge to build a dedicated targeting-mode campaign around a segment that’s barely cleared the threshold. If your business runs on a smaller local footprint, this is also the kind of change worth weighing against the rest of your account setup — what a Google Ads setup built for a local service business actually needs tends to differ from what a national e-commerce account needs, and audience thresholds are one more input into that. The same logic applies to anyone managing their own account without an agency — a lower list-size floor is one less reason a DIY advertiser needs to skip audience strategy entirely.

Membership Duration: The Setting Almost Nobody Revisits

Every remarketing list has a membership duration — how long someone stays on the list after the qualifying action, before they age off automatically. This setting gets configured once, usually left at whatever the default was when the list was created, and then essentially never looked at again. That’s a mistake, because the right duration depends entirely on your buying cycle, and buying cycles don’t stay constant.

Google caps membership duration for lists used specifically on Search campaigns at 180 days, meaningfully shorter than the up-to-540-day ceiling available for lists used on Display and for Customer Match lists generally. As of April 2025, Google also began actively enforcing that 540-day maximum on Customer Match specifically — any contact that hasn’t been added or updated within that window gets automatically dropped from the list, whether or not you’d noticed.

The default when you create a new list is typically 30 days, which is far too short for a lot of businesses. If your average customer takes six weeks to decide between quotes on a home renovation, or a B2B buyer needs two months to get budget signed off, a 30-day remarketing window ages people out of your RLSA audience before they’ve even finished evaluating you. On the other end, keeping someone on a “recently converted” exclusion list for the full 180 or 540 days when your purchase cycle is genuinely a one-time event for most customers just wastes list capacity without adding anything useful.

The fix isn’t complicated — it’s auditing existing lists against how long your sales cycle actually runs, rather than assuming whatever was set at creation still makes sense. It’s exactly the kind of setting that drifts out of alignment with the business silently, the same way a lot of account configuration drifts without anyone deliberately changing it — nobody decided the membership duration was wrong, it was just set once by whoever built the list originally and never revisited as the business or the sales cycle changed.

Why Your Bid Adjustments Might Be Doing Nothing

Here’s the part that catches the most accounts, and it’s almost invisible unless you specifically go looking for it. If your Search campaigns run on Manual CPC, an audience bid adjustment functions exactly the way it looks like it should — a straightforward multiplier on your bid for anyone in that segment, positive or negative, that Google applies directly.

Move that same campaign to Target CPA or Target ROAS, and audience bid adjustments get ignored. Not removed from the interface, not flagged as inactive — ignored. The setting stays visible in the campaign, the percentage you configured months or years ago is still sitting there looking exactly like it’s doing something, and it simply isn’t a factor in how Smart Bidding sets bids anymore. The same is true of location, demographic, and ad-schedule bid adjustments under most Smart Bidding strategies — Target CPA and Target ROAS treat all of these as inputs the algorithm considers as signals among a much larger set (device, browser, time of day, conversion path history, and audience membership among many others), rather than as fixed multipliers you control directly. Device bid adjustments are the one meaningful exception still functioning as an actual on/off lever under Target CPA, including the ability to exclude a device entirely with a -100% adjustment.

This is a genuinely common gap, and it tends to show up at a specific moment: an account that built out a careful RLSA bid-adjustment structure while running Manual CPC, then migrated to Smart Bidding at some point for the (correct) reason that Smart Bidding generally outperforms manual bidding once there’s enough conversion data. Nobody goes back and audits which of the old manual-bidding settings still do anything after that migration, because there’s no prompt telling you to. The campaign keeps running, the bid adjustments keep displaying their configured percentages, and everyone reasonably assumes the audience strategy they built is still in effect. If you’re mid-decision on which Smart Bidding strategy fits your account, or auditing one that’s already running, how to actually choose between Target CPA, Target ROAS, and Maximize Conversions is worth reading alongside this, since the choice of strategy is exactly what determines whether your audience-layer bid adjustments do anything at all.

What Actually Still Works Under Smart Bidding

None of this means RLSA becomes useless once an account moves to automated bidding — it means the lever you’re pulling has to change. A few things still function, and function well, regardless of bidding strategy:

• Targeting mode itself. Restricting a campaign or ad group to only serve audience-segment members is a delivery control, not a bid adjustment, and it works identically under Manual CPC and Smart Bidding. A dedicated targeting-mode campaign for past visitors, running under Target CPA, still only shows ads to people on that list — Smart Bidding just decides how much to bid within that already-restricted pool.
• Audiences as a Smart Bidding signal. Even though the manual percentage adjustment stops functioning, audience membership doesn’t disappear from what the algorithm considers — it becomes one signal among many that Smart Bidding factors into its real-time bid calculation, rather than a lever you set once and control directly. The difference is you’re no longer telling the algorithm exactly how much more to bid; you’re trusting it to weigh that signal against everything else it knows.
• Ad copy and creative tailored to targeting-mode campaigns. If you’ve built a dedicated campaign for return visitors, there’s no reason the ad copy needs to look identical to your cold-traffic ads. Referencing that someone’s already looked at a specific product, or leading with a different offer for people who abandoned a cart rather than people searching cold, is a lever that has nothing to do with bidding strategy and works exactly the same regardless of how bids get set.
• Exclusions. Removing recent converters or existing customers from acquisition-focused campaigns via a Customer Match or remarketing exclusion list works the same way under any bidding strategy — it’s a delivery filter, not a bid signal, so Smart Bidding has no mechanism to override it.
• Feeding audience signals into Performance Max. The same well-built remarketing and Customer Match lists doing RLSA work on Search can be reused as audience signals in Performance Max asset groups, giving Google’s automated system a stronger seed to learn from during the ramp-up phase. If you’re already maintaining these lists for Search, there’s very little additional cost to also pointing them at a Performance Max campaign — though it’s worth pairing with a real look at how to audit a Performance Max campaign when you can’t see inside it, since audience signals are an input Google’s AI considers, not a guarantee of where your budget actually goes.

The shift in mindset that actually matters is treating audience data as something that shapes campaign structure, exclusions, and creative — the parts of the account you still directly control — rather than something you set once as a bid multiplier and trust to keep working forever.

A Practical Way to Set This Up

If you’re building RLSA into an account that doesn’t have it yet, or auditing a setup that’s been running unattended for a while, this is a reasonable order to work through it:

• Confirm your remarketing tag is actually firing across the pages you think it is, and check the current size of each list in Audience Manager against the new 100-user minimum rather than assuming any list is too small to use.
• Segment visitor lists by intent rather than keeping one blended “all visitors” list — cart abandoners, pricing-page visitors, and people who reached a conversion-adjacent page but didn’t finish are different audiences with different value.
• Add your best lists to existing Search campaigns in observation mode first, and let them accumulate real conversion data before deciding whether a dedicated targeting-mode campaign is justified.
• Check your account’s current bidding strategy on every campaign carrying audience bid adjustments. If it’s Target CPA or Target ROAS, treat any existing adjustment percentage as decorative rather than functional, and redirect that effort toward campaign structure, exclusions, or creative instead.
• Review membership duration on every list against your actual sales cycle, not the default it was created with — six weeks of buying consideration deserves more than a 30-day window.
• Build at least one Customer Match exclusion list for existing customers or recent converters, and apply it to acquisition-focused campaigns regardless of bidding strategy, since it’s one of the few audience controls Smart Bidding can’t override.
• If you’re running Performance Max alongside Search, point your best remarketing and Customer Match lists at its audience signals too, rather than maintaining them for Search alone.

None of this requires rebuilding an account from scratch. Most of it is auditing settings that were configured once, correctly, at a point in time that no longer matches how the account runs today — which describes a large share of what actually goes stale in a Google Ads account generally, whether it’s an audience list, a bid strategy holdover, or a bid adjustment nobody remembered to remove.

Where Ongoing Monitoring Fits In

The bid-adjustment problem covered above is a good example of a broader pattern: almost nothing about it shows up as an error. The campaign doesn’t pause, no warning appears in the interface, and performance doesn’t necessarily crater in an obvious way — it just quietly stops reflecting the audience strategy someone built, sometimes years earlier, and nobody finds out until they go looking specifically. The same is true of a membership duration set for a sales cycle the business no longer has, or a remarketing list nobody rebuilt after a site redesign changed which pages the tag actually fires on.

Growera’s daily account monitoring is built around exactly this category of drift — settings that were correct when configured and quietly stopped being correct as the account, the bidding strategy, or the business changed around them. It won’t rebuild your audience strategy for you, but it’s designed to flag the kind of change that would otherwise sit unnoticed for months, which is precisely how a lot of these RLSA gaps end up costing more than anyone realizes before someone finally audits the account.

The Short Version

RLSA is more accessible now than it’s been at any point since Google introduced it — the drop to a 100-user minimum in December 2025 opened the feature up to a real share of smaller and newer accounts that were locked out by list size alone. That’s the good news, and it’s worth acting on if your account never had enough traffic to qualify before.

The less good news is that a lot of accounts that already had RLSA running have let part of it go quietly stale, usually because a move to Target CPA or Target ROAS silently switched off the bid adjustments the whole structure was originally built around, without any indication in the interface that anything had changed. Fixing that isn’t about abandoning audience strategy under Smart Bidding — it’s about redirecting the effort toward the levers that still work: targeting-mode campaigns, exclusions, tailored creative, and well-maintained lists feeding both Search and Performance Max. If you’re not sure which category your account falls into, checking is a lot faster than most of the other things worth auditing in a Google Ads account, and it’s the kind of gap that’s easy to close once you know it’s there. If you’d rather have that kind of check running continuously instead of once, Growera’s plans and the ERA AI engine behind them are built specifically to catch this category of silent configuration drift before it’s cost you a quarter of missed performance.