RLSA (Remarketing Lists for Search Ads)

RLSA is a Google Ads audience targeting feature that enables advertisers to reach people who have previously visited their website when those users later search for relevant keywords on Google — both conditions (list membership and an active search query) must be satisfied simultaneously for an ad to serve. This makes RLSA fundamentally different from display remarketing: it activates at the intersection of prior engagement and current intent, not just on prior engagement alone. (Store Growers, storegrowers.com, 2026-01-08)

---

Naming note

Google now labels the feature "Remarketing for Search Ads" (RDSA) in some documentation, though both "RLSA" and "RDSA" appear in the interface as of mid-2026. The underlying mechanics — applying an audience segment to keyword-based search campaigns — are unchanged. (SolidHQ, solidhq.com, 2026-05-17)

In the current Google Ads interface, the language has also shifted toward "audience segments" and "your data" as the generic labels, but the strategic concept — combining previous engagement with current search intent — is the same. (groas.com, 2026)

---

How it works mechanically

RLSA audiences are built from website or app visitors tracked via the Google Tag (GT-prefix) or Google Analytics 4 (G-prefix). As of 2024, Universal Analytics audiences stopped feeding Google Ads, meaning all remarketing setup now flows through GA4 or the Google Tag. (SolidHQ, solidhq.com, 2026-05-17)

Minimum thresholds (as-of 2026-07-19)

Network	Minimum active users in list
Search and Shopping	1,000 cookies
Display, YouTube, Discover	100

(Google Analytics Help, support.google.com/analytics/answer/9759219, undated — treat as volatile)

List parameters

• Maximum membership duration: 540 days (as-of 2026-07-19, Google Analytics Help — undated, volatile)
• Bid adjustment range (Manual CPC): −90% to +900% (as-of 2026-02-06, Bigeye Agency, bigeyeagency.com)
• Maximum lifespan of 90-day window: groas.com (2026) reports that visitors from more than 90 days ago show sharply declining effectiveness for most ecommerce verticals, even though the technical maximum is 540 days.

Audience eligibility restriction

"Google Analytics audiences that contain demographic dimensions such as age, gender, or in-market are not eligible to be used on Search, YouTube, Display, or Gmail campaigns when exported to Google Ads." (Google Analytics Help, support.google.com/analytics/answer/9759219, undated)

GA4 consent-mode enforcement (June 2026)

From 15 June 2026, Google enforces stricter GA4–Google Ads data controls: only users who granted ad_storage consent are eligible for remarketing. Lists that depend on consent-denied users will shrink or stop refreshing. (SolidHQ, solidhq.com, 2026-05-17)

---

Two campaign modes

Observation mode (Google's recommended default)

In Observation mode, an RLSA audience is layered onto an existing campaign without narrowing delivery. Ads serve to all qualifying searchers as usual; bid adjustments and performance segmentation are applied per audience segment.

Google's official documentation explicitly recommends Observation as the default setting for all Search campaigns: "The 'Observation' setting is recommended for all Search campaigns as well as the Display campaigns of more advanced advertisers." (Google Ads Help, support.google.com/google-ads/answer/7365594, undated)

Under Smart Bidding, first-party audience segments added via Observation serve as signals for Smart Bidding strategies (Google Ads Help, ibid.) — but manual bid adjustment values are ignored (see Smart Bidding interaction below).

Targeting mode (restrict to list members)

In Targeting mode, the campaign or ad group serves exclusively to users on the remarketing list. This enables a powerful advanced tactic: bidding on broad, high-volume, otherwise unprofitable keywords — such as "beds" or "project management" — knowing that the prior visit de-risks the low intent. (Store Growers, storegrowers.com, 2026-01-08; groas.com, 2026)

Store Growers recommends dedicated Targeting-mode RLSA campaigns only for stores with more than 100,000 monthly visitors, because smaller sites typically cannot generate sufficient list size to drive meaningful traffic. Example given: a 5,400–31,000-user list generated only 267 clicks in 30 days for one managed account. (Store Growers, storegrowers.com, 2026-01-08)

Critical mistake: switching an existing well-performing campaign from Observation to Targeting mode causes traffic to dry up immediately, as the campaign stops receiving cold-traffic impressions. The recommended practice is to create a new, separate campaign for Targeting mode rather than converting an existing one. (Key Commerce / Sam, youtube.com/watch?v=py8hvd5WZFY, approx. 2020)

---

RLSA and Smart Bidding

This is the most practically important friction point in 2026. Under Smart Bidding strategies (Target CPA, Target ROAS, Maximize Conversions, Maximize Conversion Value), manual audience bid adjustments — including RLSA audience adjustments — are not supported and are ignored by the algorithm, which factors audience signals internally at auction time.

"The algorithm already factors audience signals — including remarketing list membership, customer match status, in-market segment membership — into every auction-time bid decision." (Bigeye Agency, bigeyeagency.com, 2026-02-06)

The only remaining actionable levers under Smart Bidding are:

• −100% device exclusion (removes a device from auction entirely)
• Conversion value rules (available on tROAS / Maximize Conversion Value only; inflate reported revenue, not universally recommended)

Google now evaluates approximately 3,847 auction-time signals in 100 milliseconds (cited in Bigeye Agency, 2026-02-06 — citing Google documentation; treat as volatile).

What still works: adding RLSA audiences in Observation mode on Smart Bidding campaigns, for reporting and audience-level performance segmentation. The data is surfaced even though bid adjustments have no auction-level effect. (Bigeye Agency, bigeyeagency.com, 2026-02-06)

Risk: stacking multiple manual bid adjustments multiplicatively — e.g. +30% mobile × +25% RLSA × +20% ad schedule = 1.95× combined modifier — produces undetected cost spikes under Manual CPC campaigns. (Bigeye Agency, bigeyeagency.com, 2026-02-06)

August 2026 Smart Bidding target enforcement change

Starting 17 August 2026, Google is changing its bidding systems so that campaigns deliver more closely to the actual Target CPA or Target ROAS set, even when budgets are adjusted. If a campaign's Target CPA is $10 but recent actual CPA is $5, performance will shift toward $10 from that date unless the target is updated. Affects Search, Shopping, Performance Max, Demand Gen, Display, Hotel, and Travel campaigns using Target CPA or ROAS. App Campaigns, Video reach, and Video view campaigns are exempt. (Google Ads Help, support.google.com/google-ads/answer/17061251, 2026-07; effective 2026-08-17 — volatile)

---

---

RLSA and Performance Max

Performance Max (PMax) campaigns expose only "audience signals" — not list-based bid controls. Audience signals in PMax are explicitly not targeting: they tell the system where to start looking, but Google's algorithm is free to serve ads to users outside the signal. This is categorically different from RLSA's Targeting mode. (YouTube practitioner video, youtube.com/watch?v=utk7OFdvtnM, 2025-01)

PMax now handles remarketing via its own mechanisms across all Google surfaces, using RLSA lists, Customer Match, and GA4 segments as input signals for AI bidding.

Industry data (as-of 2026): "Performance Max now accounts for 67% of Shopping spend and 68% of Shopping-attributed sales among advertisers running both [PMax and standard campaigns]." (AdsGo Blog, adsgo.ai, 2026 — single vendor source, volatile)

Practical recommendation (practitioner consensus, not orchestrator's view): SolidHQ recommends running standard RLSA/RDSA campaigns and Performance Max concurrently for accounts with strong audience signal: standard search remarketing gives clear control over creative and frequency for warm audiences; PMax uses those same audiences as an AI bidding signal across all Google surfaces. (SolidHQ, solidhq.com, 2026-05-17)

---

RLSA vs Customer Match

Dimension	RLSA	Customer Match
Data source	Website/app visitors (cookied)	First-party CRM data (email, phone, address)
Cookie dependency	Yes	No (matched via signed-in Google account)
Typical list size	Larger (most sites have more cookied visitors than emails)	Smaller (only confirmed email subscribers/customers)
Durability	Vulnerable to cookie erosion (Safari/Firefox ITP)	Durable — cookie-independent
Time limit	540-day maximum membership	No inherent time limit on the data
Minimum match size	1,000 active users	~1,000 matched users

Customer Match is described in 2026 as "one of the only durable, cookie-resistant audience types left" and is recommended for win-back campaigns, VIP exclusion lists, lookalike seeding inside Performance Max and Demand Gen, and B2B account-based targeting. (SolidHQ, solidhq.com, 2026-05-17)

RLSA lists tend to be considerably larger than Customer Match lists for most ecommerce retailers; Customer Match lists, while smaller, are more durable as third-party cookie deprecation continues across non-Chrome browsers. (SolidHQ, solidhq.com, 2026-05-17)

Cookie context: Safari and Firefox block third-party cookies entirely; ITP shortens cookie lifespans; ad blockers strip tracking parameters. Google reversed its Chrome cookie deprecation plan (reaffirmed April 2025 — third-party cookies remain on by default in Chrome), but list erosion from Safari/Firefox users persists and grows with their browser market share. (groas.com, 2026)

Similar Audiences sunset: Similar Audiences were removed in 2023. Smart Bidding and Performance Max now use Customer Match and GA4 audiences as seed signals in their place. (SolidHQ, solidhq.com, 2026-05-17)

---

Ecommerce use cases

Store Growers (storegrowers.com, 2026-01-08) identifies five primary ecommerce RLSA tactics:

1. Tiered bid adjustments by buyer-journey stage — different bid modifiers for product viewers, cart abandoners, checkout starters, past purchasers
2. Bidding on expensive generic keywords for warm audiences only — a broad keyword like "beds" or "project management" that is unprofitable against cold traffic becomes viable when restricted to list members
3. Cross-sell and upsell to existing customers — serve ads for complementary products to past purchasers searching related terms
4. Competitor brand term bidding for warm audiences — bid on competitor brand terms only for users already in your consideration funnel
5. Time-decay bid ladders — reduce bids the longer it has been since the user's last visit; beyond 90 days, effectiveness drops sharply for most verticals

Recommended audience segmentation tiers

SolidHQ (solidhq.com, 2026-05-17) recommends segmenting into at least four behavioural tiers:

Tier	GA4 event	Recency	Bid orientation
Product viewers	view_item	Last 14 days, no purchase	Moderate increase
Cart abandoners	add_to_cart	Last 7 days, no purchase	Strong increase
Checkout starters	begin_checkout	Last 3 days, no purchase	Aggressive increase
Past purchasers	purchase	Last 60 days	Cross-sell / win-back

groas.com (2026) recommends building multiple lists at different membership durations (7-day, 14-day, 30-day, 60-day, 90-day) and bidding them differently, with the shortest-duration, highest-intent lists carrying the highest bids.

Scale constraint

The 1,000-user minimum per list segment compounds when applying multiple sub-lists simultaneously. Practitioners recommend keeping segmentation simple for accounts under 100,000 monthly visits, as each sub-list must independently reach 1,000 active users to activate. (Store Growers, storegrowers.com, 2026-01-08)

RLSA for Shopping

RLSA can be applied to Shopping campaigns (Google Shopping remarketing lists), where it follows the same observation/targeting mechanics as Search. Tinuiti (tinuiti.com) described RLSA for Shopping as producing CTR and CVR lifts "up to 400% in some reported cases" but the date of this claim is unknown and the benchmark is likely pre-2024.

---

Benchmarks (as-of 2026)

Note: all performance benchmarks below are vendor- or agency-sourced; no independently audited primary research was found. Treat as directional.

Metric	Value	Source
Average retargeting ROAS	~4.2×	SolidHQ, 2026-05-17 (vendor blog, volatile)
Average retargeting CPC	$0.95–$1.25	SolidHQ, 2026-05-17 (volatile; USD)
Average retargeting CTR	0.9–1.2%	SolidHQ, 2026-05-17 (vs ~0.08% cold display)
RLSA CTR uplift vs non-remarketing Search	+37%	groas.com, 2026 (unattributed — no primary source cited)
Average remarketing CPA	~$26 vs $49 cold prospecting	Amir Gomez blog, 2025 (stale-risk; USD)
Cart abandoner RLSA CPA	"2–3× better than general campaign"	Amir Gomez blog, 2025 (stale-risk)

[!unverified] The +37% RLSA CTR uplift cited by groas.com (2026) carries no traceable primary source citation in retrieved content. Do not treat as authoritative.

---

Measurement and incrementality risks

• Display remarketing campaigns frequently appear as the best performers in an account due to inflated view-through conversion attribution. Recommended: set view-through conversion window to 1 day, not the default 30-day window, and evaluate remarketing primarily on click-through conversions. (groas.com, 2026)
• Geographic holdout incrementality testing is the recommended approach for measuring true remarketing lift: pause remarketing in comparable markets and compare total (not attributed) conversions. "Results often reveal that remarketing is less valuable than attribution reports suggest." (groas.com, 2026)
• Smart Bidding campaigns below 15–20 conversions per month in a remarketing campaign struggle to optimise under automated bidding; Manual CPC is recommended for RLSA at that scale. (groas.com, 2026 — volatile)

See also: Incrementality, Multi-Touch Attribution (MTA), Data-Driven Attribution (DDA).

---

What practitioners report

From independent practitioner content (note: Reddit MCP unavailable this run; findings from practitioner newsletters and agency blogs):

• "If you're running the most popular Smart Bidding strategies — which is most accounts in 2026 — Google just ignores most of the bid adjustments you set. The bid adjustment menus still exist in the UI, which makes it feel like you have those controls… But they don't work in most cases." (PPC Edge Newsletter / Miles McNair, ppcmastery.com, 2026-05, practitioner newsletter with 20,000+ subscribers)
• RLSA for search is characterised in a 2026 video tutorial as "the least used feature inside of Google Ads remarketing," described as "more advanced" and positioned after display and YouTube remarketing in standard practitioner curricula. (YouTube, youtube.com/watch?v=VbRC5bFkkbw, 2026)
• The PPC industry is "increasingly polarised" on Performance Max: some agencies take anti-PMax stances to differentiate; others report it is essential for scale. RLSA in standard Search survives partly because it gives practitioners the transparency PMax denies — keyword control, audience transparency, creative control. (PPC Hero, ppchero.com, 2025 — stale-risk)

---

Key terms

Term	Meaning
RLSA / RDSA	Remarketing Lists for Search Ads / Remarketing for Search Ads — same feature, both labels active in 2026
Observation mode	Layer audience onto campaign without restricting reach; enables segmentation and (under Manual CPC) bid adjustments
Targeting mode	Restrict campaign/ad group to list members only; enables broad keyword bidding for warm audiences
Membership duration	How long a user stays on a remarketing list after their qualifying action; max 540 days
1,000-cookie threshold	Minimum active users required before a list activates on Search Network
Audience signal (PMax)	A suggestion to the PMax algorithm — not targeting; does not restrict delivery