What Is a Good ROAS? Industry Benchmarks for 2026

ROAS in one sentence

ROAS — Return On Ad Spend — is revenue divided by ad spend. A 5× ROAS means £5 of attributable revenue for every £1 of paid media. It's a top-line metric: it doesn't account for fulfilment, gross margin, management fees, or platform costs. That's where most ROAS conversations go wrong.

ROI is different. ROI measures profit per pound of total cost (ad spend + fees + cost of goods + overheads). MER — Marketing Efficiency Ratio — is total revenue divided by total marketing spend across all channels, used as a dataset-level health check. Three different metrics, three different decisions.

For the full ROAS definition with formulas, examples, and platform-specific reporting nuances, see the companion piece on Google Ads CPC benchmarks.

Use the ROAS calculator

Enter your monthly ad spend, attributable revenue, gross margin, and industry. The calculator returns your ROAS, your POAS (Profit on Ad Spend), your break-even ROAS, and where you sit against the industry benchmark range. POAS is the number that actually predicts whether the campaign makes you money.

Directional estimate. Industry benchmarks based on WordStream 2026 + Visionary survey & tracking dataset Q1 2026 (47 accounts).

What "good"actually means — break-even ROAS

The most important number in any ROAS conversation is your break-even ROAS — the ROAS at which gross profit from incremental revenue exactly matches the ad spend that produced it. The formula is simple: 1 ÷ gross margin.

A 30% margin business needs a ROAS of 3.33× just to break even on ad spend. A 70% margin business breaks even at 1.43×. This single number explains why the same headline ROAS produces wildly different outcomes across verticals.

Most healthy accounts target 2× break-even — the "comfortable ROAS"column above. That gives enough headroom to absorb attribution noise, seasonal swings, and CPC inflation without dropping into loss-making territory.

ROAS benchmarks by industry — full breakdown

The table below combines public 2026 benchmarks (WordStream) with our own survey & tracking data from 47 EU accounts in Q1 2026. Sortable — click any column to re-rank. Use the median as your initial target; aim for the top of the range as the account matures.

Sources: WordStream 2026; Visionary survey & tracking dataset Q1 2026 (47 accounts). Last reviewed April 2026.

Real client ROAS proof

Why ROAS varies so much by channel and bid strategy

Channel ROAS norms differ materially. Brand-defence Search Ads against your own brand terms commonly return 8–20× because the buyer is already converting; non-brand search returns 3–6× because you're acquiring net-new demand. Shopping and Performance Max sit in between. Display and YouTube are upper-funnel and rarely produce direct-response ROAS comparable to Search.

Bid strategy compounds the variance. Manual CPC bidding gives full control but rarely beats the algorithm above 1,000 weekly clicks. Target ROAS constrains spend to a return target. Maximise Conversion Value scales aggressively but can over-spend on low-margin SKUs unless margin-tier custom labels are configured.

Target ROAS vs actual ROAS — what bid strategy you set vs what you get

Setting a Target ROAS of 5× doesn't guarantee a 5× return. Account variables affect actual delivered ROAS heavily — feed quality, conversion-tracking accuracy, audience signals, asset group structure, landing-page conversion rate. Over-restrictive Target ROAS chokes volume and the algorithm under-spends. Under-restrictive Target ROAS bleeds budget chasing impressions.

The practical rule: set Target ROAS at 80–90% of your trailing 30-day actual ROAS, then nudge upward as performance proves out. Setting it at 130% of current actual is the most common cause of suddenly stalled spend.

How to lift your ROAS

Eight practical levers, in rough order of impact for most accounts:

1. Margin-tier custom labels for Shopping

Populate custom_label_0With margin tier (Tier 1 = highest margin, Tier 3 = lowest). Run Tier 1 with aggressive Target ROAS, Tier 3 with Maximum CPC caps. This single change typically lifts blended ROAS 20–40% without changing spend.

2. Conversion-tracking accuracy + value-based tracking

Smart Bidding optimises against the data you give it. Pass actual order value (not flat conversion = 1), include refund signals, and use enhanced conversions for cookieless attribution.

3. Feed quality (titles, attributes, custom labels)

Title structure alone — moving from "SKU-7741-K"to "Bunk Bed Triple Sleeper Solid Pine Single + Single Over Double White" — typically lifts CTR 30–80% and ROAS proportionally. Full deep-dive in our 47-point feed optimisation checklist.

4. Audience signals on PMax asset groups

Performance Max blends signals across Search, Shopping, Display and YouTube. Without audience signals, asset groups optimise blind. With customer-match lists, in-market segments, and detailed demographics, asset-group ROAS typically climbs 15–30%.

5. Negative keyword discipline

Weekly negative-keyword reviews on Search campaigns, monthly on Shopping and PMax (via search-terms reports + brand-exclusion lists for non-brand campaigns).

6. CSS partner activation

Routing Shopping spend through a CSS partner reduces effective CPC by ~20%. Lower CPC at constant revenue = higher ROAS. Read the CSS partner explainer.

7. Brand vs non-brand campaign separation

Brand campaigns typically return 8–20× ROAS but represent capture of existing demand, not net-new acquisition. Reporting them blended with non-brand inflates headline ROAS and disguises true acquisition cost.

8. Customer-LTV-based bid logic

For high-LTV verticals (B2B SaaS, subscription, repeat e-com), bidding to first-purchase value under-weights long-term contribution. Pass predicted LTV as conversion value where attribution allows.

When low ROAS is actually fine

Counter-intuitively, sometimes a low ROAS is the right strategy. Five common scenarios:

• New customer acquisition where LTV > first purchase.A subscription business acquiring a customer at break-even on first purchase is actually highly profitable over the contract.
• Brand-building campaigns.YouTube TOFU and Demand Gen rarely return direct-response ROAS but lift branded search and direct traffic for months afterwards.
• Product launches.Acquiring early reviewers and demand-validation traffic at low ROAS is normal investment, not waste.
• Market expansion.Entering a new geo or vertical typically runs at half normal ROAS for the first 60–90 days as the algorithm learns.
• Competitive defence.Bidding on competitor brand terms returns lower ROAS but denies them airtime.

The mistake is optimising for ROAS at the cost of strategic objectives. The fix is to set ROAS targets per campaign type, not blanket account-level.

The seven most common ROAS mistakes we see in audits

Across the 47-account Q1 2026 audit cohort, seven ROAS mistakes appeared in more than half of accounts. Each one silently caps performance for months before anyone spots it. Fixing them accounts for the bulk of the 20–40% ROAS lift we typically deliver in the first 90 days.

Mistake one: reporting blended ROAS at account level. A 6× account ROAS composed of 15× brand and 3× non-brand looks healthy but hides an acquisition problem. Segmented reporting — brand, non-brand, Shopping, PMax, retargeting — surfaces where the money is actually being made and where it is being wasted. The fix is a five-minute Looker Studio filter, not a strategy overhaul.

Mistake two: setting Target ROAS above trailing 30-day actual. If your account has run at 4× for a month and you set Target ROAS to 6×, Smart Bidding pulls back into the highest-intent auctions only. Spend drops 40–70% overnight. Revenue drops with it. The correct move is Target ROAS at 80–90% of trailing 30-day actual, nudged up by 5% per fortnight while volume holds.

Mistake three: flat conversion value. Passing conversion value = 1 on every purchase tells Smart Bidding that a £5 sock and a £500 sofa are worth the same. The algorithm cannot possibly optimise for margin pounds. Passing actual order value is a two-line change in your tag manager and typically lifts ROAS 15–25% within a fortnight.

Mistake four: no refund signals. Fashion accounts with 30%+ return rates that don't feed refunds back to Google Ads are optimising against phantom revenue. Enable enhanced conversions for leads/refunds and pipe returns into the conversion adjustment API. Blended ROAS falls on paper but the underlying number is finally honest, and the algorithm stops chasing high-return SKUs.

Mistake five: comparing this year's ROAS to last year's without adjusting for CPC inflation. UK CPCs rose ~11% year on year in 2025. A ROAS that dropped from 6× to 5.4× at flat AOV is exactly the CPC inflation figure — the account hasn't got worse. Adjust benchmarks annually or you'll fire agencies for delivering flat real-terms performance.

Mistake six: ignoring POAS. ROAS optimisation without margin data leads to over-investment in low-margin SKUs. Populate custom_label_0 with margin tier, run three campaign tiers, and bid Target ROAS at each tier's break-even × 2. Almost every account we run this on adds contribution while spend stays flat.

Mistake seven: no attribution reconciliation. GA4, Google Ads, and Shopify report different conversion totals — commonly a 20–40% gap. Picking one platform as source of truth and reporting it consistently prevents the fortnightly "our ROAS is 4× or 6× depending who you ask" argument.

Attribution windows and their effect on reported ROAS

Two identical accounts can report ROAS 30% apart purely because of attribution settings. The default Google Ads window is 30-day click, 1-day view; GA4 defaults to data-driven attribution with a 30/1 window; Shopify last-click attributes to the final referring source at checkout. None of these is wrong, but they are not comparable.

For a considered-purchase category (furniture, B2B, luxury) with a 14–45 day decision cycle, shortening the click window to 7 days will typically drop reported ROAS by 15–30%. For impulse purchases (beauty, apparel under £50) a 7-day window changes reported ROAS by less than 5%. Match the window to your buying cycle, then hold it constant.

Data-driven attribution — the current Google Ads default — redistributes credit across all touchpoints in the path. Compared with last-click, DDA typically lifts Display and YouTube ROAS by 30–60% (because they get partial credit for assists) and drops brand-search ROAS by 10–20% (because brand no longer takes 100% of the credit). Neither model is "true" — both are estimates. The value of DDA is that it stops you from cutting upper-funnel channels that are actually driving downstream revenue.

For the cleanest single-number reporting, we recommend picking one attribution model, documenting the window in every board deck, and running MER (total revenue ÷ total marketing spend) as an unattributed sanity check. When MER and blended ROAS move in opposite directions, attribution is masking a real problem.

ROAS through the year — seasonality patterns

ROAS is not stable across the calendar. UK e-commerce accounts in the Q1 2026 dataset showed peak ROAS in Q4 (Nov–Dec, median 7.8×) driven by cyber-week AOV lift and higher intent. Q1 (Jan–Mar) troughs at median 4.9× as CPC inflation outpaces post-Christmas demand. Q2 recovers to 5.6× and Q3 sits at 6.1×.

Bidding to a flat annual Target ROAS misses the seasonal opportunity. The correct approach is a seasonality-adjusted Target ROAS schedule: lift the target 15–20% for Q4, drop it 10–15% for Q1, and use Google Ads' seasonality adjustments feature for known one-off promotional windows (Black Friday, Boxing Day, brand-specific launches). Under-adjusting Q4 leaves growth on the table; over-restricting Q1 hands share to competitors who ride the CPC dip.

B2B seasonality inverts: Q1 and Q3 are strongest as budget cycles trigger buying; December is the weakest month across every B2B account in the dataset. Setting Q4 Target ROAS to Q1 levels wastes budget chasing an audience that isn't buying.

Portfolio bidding — when to pool budgets across campaigns

Standard bidding runs one strategy per campaign. Portfolio bidding pools campaigns under a single Target ROAS or Maximise Conversion Value strategy, letting Smart Bidding shift budget between campaigns based on real-time opportunity. Done well, it lifts blended ROAS 10–20% without any content change. Done badly, it starves niche high-margin campaigns of budget.

Pool campaigns that share margin structure, target audience, and product category. Do not pool brand with non-brand — they have different underlying economics. Do not pool prospecting with retargeting for the same reason. A three-portfolio structure (brand pool, non-brand acquisition pool, retargeting pool) covers most accounts.

For accounts under £5,000/month spend, portfolio bidding usually hurts more than it helps — there isn't enough conversion volume for the algorithm to reallocate meaningfully. Above £15,000/month spend, portfolios are almost always a net positive. Between those, test with 10% of budget for four weeks and compare portfolio ROAS to matched-cell control campaigns.

Methodology

Industry benchmarks above combine WordStream 2026 cross-vertical data with the Visionary survey & tracking dataset Q1 2026 — 47 EU accounts spanning e-commerce, B2B SaaS, B2B services, lead gen, local services, travel, and education. Median figures are weighted by spend; range bounds reflect 10th and 90th percentile observed performance.

Client case-study figures are reported as published in their respective case study pages and are validated against direct platform exports (Google Ads, Microsoft Ads, GA4) before publication.

Last reviewed: April 2026. Next review: July 2026.

Why LTV changes the ROAS conversation completely

First-purchase ROAS is what every reporting dashboard shows. Lifetime-value ROAS is what actually predicts business health. For any brand with genuine repeat purchase economics — subscription, consumables, high-consideration replacement categories — reporting only first-purchase ROAS materially understates the value of paid media and drives systematically wrong spend decisions.

A concrete example: a UK skincare subscription client at £45 AOV showed first-purchase ROAS of 2.4× on Meta prospecting — below the 3× target the founder had set. Six-month LTV analysis revealed the same cohort produced £142 average revenue by month six, taking effective LTV ROAS to 7.6×. Cutting Meta spend because first-purchase ROAS looked weak would have destroyed the account's growth engine.

The practical fix is a two-metric reporting standard: report both first-purchase ROAS (for optimisation velocity) and cohort LTV ROAS at 90 and 180 days (for strategic health). Set your break-even threshold on LTV, not first-purchase. For Smart Bidding, pass a blended conversion value that includes predicted LTV multiplier for known repeat categories — most tag managers can compute this in a single custom variable.

For one-off purchase categories (furniture, home appliances, wedding) LTV multipliers are usually below 1.3× and don't warrant the reporting complexity. For subscription and consumable categories where LTV multipliers routinely hit 3–8×, ignoring LTV is the single most expensive reporting mistake a founder can make.

Reviewing ROAS quarterly rather than monthly filters out attribution noise and seasonality. Monthly ROAS reviews create panic decisions; quarterly reviews create real optimisation. Set the cadence at the board level and stick to it.

Frequently asked questions