Ad Strength and RSA Asset Reporting in 2026: What to Actually Change
“Ad Strength: Average” — with a helpful nudge to add three more headlines. You add them, the bar turns green, and your CPA does not move. That is the standard experience, and the reason is simple: Ad Strength and the asset performance report answer different questions, and most advertisers only ever look at the louder one.
This article covers what Ad Strength actually measures (and what it does not influence), how to read Low / Good / Best asset ratings, what to do about Learning and Pending, when pinning is the right call, the replacement order that produces results, and how to verify a change without fooling yourself with numbers.
What Ad Strength is, and what it is not
Ad Strength rates a responsive search ad on a scale from Incomplete to Poor, Average, Good and Excellent. It looks at four things:
- the number of unique headlines and descriptions;
- how different those headlines are from one another;
- whether relevant keywords appear in the copy;
- how many assets are pinned — heavy pinning lowers the rating.
What it is not: a ranking factor. Ad Strength does not enter Ad Rank directly and does not replace Quality Score — the mechanics of what actually moves your cost per click are covered in our breakdown of impression share and auction insights. Ad Strength is a completeness diagnostic: it tells you whether the system has enough raw material to assemble combinations from.
“Excellent” does not mean profitable. It means the system has options. Profit is decided by which assets actually get served, not by the colour of the bar.
Hence the practical rule: Good is enough. Chasing Excellent by padding the set with near-synonyms is a reliable way to dilute relevance in exchange for a green label.
The asset performance report is where the data lives
Find it via Campaign → Ads and assets → select the ad → View assets (or the Assets tab at ad group level). Each headline and description receives one of these labels:
| Rating | Meaning | Action |
|---|---|---|
| Learning | In rotation, not enough data yet | Leave it alone and let impressions accumulate |
| Pending | Has not gathered enough impressions to be rated | Check whether stronger assets are crowding it out |
| Low | Underperforms relative to the rest of this set | Replace it, do not mass-delete |
| Good | Performing acceptably | Keep as a baseline reference |
| Best | Among the strongest in the set | Do not touch; extend the idea into new assets |
The nuance that breaks most people’s interpretation: these ratings are relative. Low does not mean “bad by industry standards”, it means “worse than its neighbours in this ad”. Delete every Low and within a month some of your Good assets become the new Low. That is not decay, it is the middle being redefined.
Why some assets sit in Pending forever
Google does not rotate every combination evenly. It finds a few that work and gives them most of the impressions. An asset that missed the early winning combinations may never accumulate a sample. That is normal at modest volumes: at 30–50 impressions a day, waiting for ratings is pointless — you need traffic first.
The combinations report: the most ignored tab in the account
Next to assets sits a Combinations report showing which headline and description pairings actually served, and how often. What to look for:
- Repeated meaning inside one ad. The classic: “Next-day delivery” and “Fast shipping” served together. The reader gets one idea twice and half the ad is wasted.
- Contradictory promises. “Free consultation” beside “Pay only for results” reads as confusion, not generosity.
- Impression concentration. If 80% of impressions sit in three combinations, the other twelve headlines are decoration for the Ad Strength bar.
- Mobile truncation. Long headlines paired together produce an ellipsis. Check the mobile preview, not the desktop one.
Pinning is a precision tool, not a sin
Pinning locks an asset into a position. It lowers Ad Strength — and there are cases where that trade is clearly correct.
| Pin when | Example |
|---|---|
| Legal or regulatory wording is required | Mandatory licence or disclaimer text |
| Brand belongs in position 1 | Recognition and brand traffic defence |
| A price or condition must always show | “From $39/month” cannot silently disappear |
| A call to action belongs in position 3 | Consistent message architecture |
The workable compromise is pinning two or three variants to a position rather than a single asset. Structure survives, and the system still has room to test. Fifteen headlines with twelve unpinned is a healthy baseline.
How to change assets in an order that produces results
- Build a base first. Change nothing until the ad group has roughly 2,000–3,000 impressions — a working benchmark for mid-sized budgets.
- Map the angles. Sort headlines by type: benefit, price or terms, trust, urgency, objection handling, product feature. Usually you discover that fifteen headlines are three ideas rewritten five times.
- Change one or two at a time. Swapping eight headlines at once makes conclusions impossible and restarts learning.
- Replace a Low asset with a new idea, not a synonym. If “Fast shipping” rated Low, “Rapid delivery” will rate the same.
- Extend your Best assets. A Best rating is a confirmed hypothesis about what the audience cares about. Write three variations of that specific idea.
- Give a new asset 14 days. Anything sooner and you are reading a Learning label as a verdict.
On idea sourcing: the best headline generator is your search terms report and the language customers use, not a brainstorm. If queries show people searching for “with money-back guarantee”, that is a finished headline. For the testing discipline around this, see our guide to A/B testing responsive search ads.
Assets beyond headlines: extensions get rated too
Campaign-level assets — callouts, structured snippets, sitelinks, images — have their own performance reporting, and their contribution to CTR is often larger than a fifteenth headline’s. How to build a complete set without losing SERP real estate is covered in our article on ad extensions and assets in Google Ads.
A practical minimum per campaign: 6+ sitelinks, 8+ callouts, 2 structured snippets, images where the vertical allows, and a call asset if phone sales matter. Empty slots are surrendered screen space your competitor occupies for free.
Automatically created assets in 2026
Google can generate headlines from your site content. The upside is gap-filling; the downside is that you do not control the wording and may end up advertising a promise you never made.
- Review the automatically created assets section at least fortnightly.
- Remove anything containing numbers, deadlines, guarantees or prices that do not exist on your site.
- In regulated verticals — finance, healthcare, legal — turning the feature off entirely is usually the safer call.
Ad Strength in Performance Max and Demand Gen
In PMax the rating applies to the asset group and includes images, video and logos. The logic is the same — completeness — but the stakes are higher: a thin asset group limits which inventory the system can reach. No vertical video means no Shorts placements; no square image means part of Discover is unavailable.
A quick asset group checklist: 3–5 headlines plus 1–2 long headlines, 2–4 descriptions, images in three ratios (1:1, 1.91:1, 4:5), logos in two, and at least one video. Skip the video and the system will assemble one for you, usually worse than you would. For measuring what that video actually does when it produces almost no clicks, see our piece on Brand Lift and Search Lift on YouTube.
Verifying that an asset change worked
Week-over-week before-and-after comparisons almost always lie: seasonality, competition and query mix all shift underneath you. The honest route:
- Run a 50/50 split experiment: control campaign with the old set, treatment with the new one.
- Decide on CPA or ROAS, never CTR. Rising CTR with falling conversion rate means you got better at attracting the curious.
- Allow at least one full conversion window plus a buffer for lag.
- Do not touch bids or budgets mid-test.
The full procedure lives in our guide to experiments in Google Ads. If you lack the volume for a formal experiment, at minimum compare like-for-like periods with an identical campaign set, and keep the confidence interval in mind before declaring victory.
Six mistakes that repeat everywhere
- Chasing Excellent at any cost. Padding with synonyms trades relevance for a rating.
- Deleting every Low at once. You erase accumulated data and buy another learning cycle with no new knowledge.
- Grading assets by manual CTR math. Google withholds per-asset click data by design; assets contribute to combinations, not to isolated metrics.
- One RSA serving an ad group of 40 keywords. No headline set can be relevant to everything. Split ad groups by query intent.
- Pinning everything. At that point you have rebuilt expanded text ads with worse combination coverage.
- Forgetting the landing page. The headline promises “a quote in two minutes” and the page opens an eleven-field form. What drops is not CTR, it is conversion rate.
A 20-minute checklist
- Ad Strength on all active RSAs: target Good or better; Excellent is optional.
- Asset report: is anything rated Best? If everything is Good, the set is monotonous.
- Combinations report: check the top five servings for duplicated meaning.
- Pinning: no more than two or three positions, with 2+ variants each.
- Campaign assets: every available type populated.
- Automated assets: no promises your site does not make.
- Headline-to-landing-page match: the promise is fulfilled above the fold.
In a large account these edits go faster in bulk than one ad at a time in the web interface. And if the problem is broader than assets — you are assembling the whole paid process — PPC Rebels runs a Google PPC agency service for high-volume advertisers, and the English blog archive covers the surrounding fundamentals.
How to write ten genuinely different headlines
The usual reason behind “everything is Good and nothing is Best” is that the set was written from one head in twenty minutes. What fixes it is not inspiration but an angle matrix: decide in advance which message types must be represented, then write to each slot.
| Angle | What it tells the reader | Example |
|---|---|---|
| Direct benefit | What they get | “Invoice reports in three minutes” |
| Price / terms | The entry point | “Plans from $39 a month” |
| Objection handling | Removes the main fear | “Migrate data with zero downtime” |
| Proof | Why believe you | “430 companies already switched” |
| Speed / timeline | When results arrive | “Live in one business day” |
| Segment call-out | This is for me | “Built for Shopify stores” |
| Product feature | Functional specifics | “CRM and inventory integration” |
| Guarantee / risk | What if it does not fit | “14-day money-back guarantee” |
| Alternative comparison | Why not carry on as before | “Replaces spreadsheets and manual checks” |
| Exact query match | Keyword relevance | “Inventory management software” |
Ten angles, ten headlines, each carrying a distinct idea. The system then tells you which angle your audience responds to — and instead of an abstract “Best” label you get a usable insight: they care about speed, not price. That insight travels to your landing page, your emails and your sales scripts, which is where it pays for itself several times over.
Where to source the wording
- Search terms report. Your audience’s own language: “with guarantee”, “no upfront payment”, “for small teams”.
- Sales objections. The first three questions on every discovery call are finished rebuttal headlines.
- Reviews and support tickets. A problem stated in customer words almost always outperforms the marketing version.
- Competitor ads. Not to copy but to find the unoccupied angle: if everyone promises a low price, the one talking about timelines wins the click.
Assets and the landing page: the promise has to be kept
Asset ratings reflect more than CTR — post-click behaviour feeds into them too. A headline that sends people to a page which does not answer its promise drifts toward Low over time, and that is a correct signal rather than a glitch.
The consistency check takes five minutes: open the landing page and try to find a direct answer to each headline above the fold.
| Headline | What must be above the fold | Common failure |
|---|---|---|
| “Quote in two minutes” | Short form or calculator immediately | An eleven-field form three screens down |
| “From $39 a month” | The same figure without scrolling | “Contact us for pricing” |
| “Built for Shopify stores” | Explicit segment mention | A generic “for business” page |
| “Live in one day” | The timeline in the first block | Timeline buried in the FAQ |
When the promise is not honoured you have two options: change the headline or change the page. The second is usually the better investment, since it lifts conversion for all traffic rather than for one ad. The practical side is covered in our articles on landing page conversion for paid traffic and landing page speed and Core Web Vitals — a slow page erases every gain you make in the ad.
A worked iteration
An illustrative example. One ad group, one RSA, 15 headlines, 30 days, 42,000 impressions.
Starting point: Ad Strength Excellent, CTR 4.1%, conversion rate 2.3%, CPA $64. Asset report: 2 Best, 6 Good, 4 Low, 3 pending.
What the review revealed:
- Both Best assets were headlines with a concrete timeline (“live in one business day”, “setup within 24 hours”).
- All four Low assets were variations on “great prices”.
- The three pending assets were the longest ones, so they rarely made it into mobile combinations.
The iteration: three price headlines replaced with new angles (guarantee, segment, objection handling), long headlines trimmed to 25–28 characters, and the brand headline pinned to position 1. Ad Strength dropped to Good because of the pin.
Thirty days later: CTR 3.8% (lower), conversion rate 3.1% (higher), CPA $51. The classic pattern — the ad became less enticing and more precise. Which is exactly why the decision is made on CPA, not on CTR and certainly not on the Ad Strength bar.
The transferable conclusion: this audience buys on speed, not price. Worth verifying at the targeting layer too — how to choose audiences around a confirmed motive is covered in our guide to optimized targeting and audience signals.
FAQ
Does Ad Strength affect cost per click?
Not directly. The auction is driven by your bid, Quality Score and expected impact from ad assets. Ad Strength is a completeness indicator, not a bid multiplier.
Do I have to fill all 15 headlines?
No. Eight to ten genuinely different headlines beat fifteen where half are synonyms. Variety of meaning matters more than filled slots.
All my assets are Good and none are Best — what does that mean?
It usually signals monotony: the system cannot find a standout. Add assets with a fundamentally different angle — price instead of benefit, objection handling instead of a feature list.
Why does a new ad show no ratings for weeks?
It needs impressions. On a small budget the sample may never arrive, in which case judge the ad as a whole rather than individual assets.
Should I delete Low-rated assets?
Replace, yes. Mass-delete, no. Swap one or two at a time for a new idea and allow two weeks for a rating.
Does pinning hurt performance?
It reduces the number of possible combinations, but when a pin protects a required message — brand, price, disclaimer — the gain in precision usually outweighs the loss in variety.
How do Ad Strength and Quality Score relate?
They are independent. Quality Score evaluates keyword relevance and landing page experience; Ad Strength evaluates the completeness and diversity of your asset set.
Can I copy winning assets into other campaigns?
Yes, with a caveat: ratings are relative to their set. A Best asset in one ad group may rate Good in another simply because its neighbours are stronger there.
How does asset reporting work in Performance Max?
Ratings are applied at asset group level across text, image and video. A thin asset group restricts available placements, so completeness matters even more than in Search.
Should I turn off automatically created assets?
In regulated verticals, generally yes. Elsewhere you can keep them, provided you audit the wording regularly — any promise not present on your site is a future complaint.
How many RSAs should an ad group have?
Two is usually right: one primary, one under test. Three or more spread impressions thin and slow down data collection.
My assets look strong but conversions are flat. Now what?
The bottleneck is further down the funnel — landing page or offer. Check load speed, whether the promise is honoured above the fold, and form length. Ad copy cannot fix any of those.
Bottom line: Ad Strength answers “do I have enough material?” and the asset report answers “which of it works?”. The first takes an hour to fix. The second takes discipline: change one asset at a time, wait two weeks, and grow your winners instead of endlessly recycling losers.
