Competitor ad analysis: the framework we run on 50,000+ real ads

Most articles on competitor ad analysis are written by people who have never analysed a competitor's ads. This one shows you the exact qualification rubric we run on every creative we track — field by field — then gives you the baselines it produces across 50,982 qualified ads from 158 UK brands, including a sector-by-sector table so you can judge your own market against the right comparison instead of a global average.

Why most competitor ad analysis advice is useless

Search this topic and you'll find the same article a dozen times: "identify your competitors, look at their ads, note their messaging, find gaps". It's a framework with no data behind it, usually published by a tool vendor, and it leaves you exactly where you started — staring at a wall of screenshots with no idea what any of it means.

The problem isn't collecting the ads. The platforms' own transparency libraries, such as Google's Ads Transparency Center and Meta's Ad Library, hand you the creatives for free — we've written step-by-step guides to all four (Google, Meta, TikTok, LinkedIn), and a combined walkthrough if you're starting from zero. The problem is what you do next. An ad library gives you two hundred thumbnails. Analysis means turning each thumbnail into structured, comparable fields, and then knowing what normal looks like, so you can spot when a rival deviates from it.

We do this at scale for AdRival's reports, which means we can publish two things a generic article can't: the actual rubric, and the actual distributions.

The qualification rubric: what to extract from every ad

Every creative we capture goes through an AI qualification pass that scores it against 25 dimensions. You can apply the same rubric by hand with a spreadsheet — it's slow but nothing about it is secret. These are the fields that earn their keep:

The remaining fields — featured_product, product_category, seasonal_theme, cta_type, price_shown, factual claims, a differentiation_note and two composite scores — matter for deeper reads, but the seventeen above do most of the work. If you're doing this manually, start with offer_type, hook_type and primary_objective: three columns in a spreadsheet already tell you more than any amount of unstructured scrolling.

What 50,000 qualified ads actually show

Here's the part no generic framework can give you: baselines. The numbers below come from the 50,982 ads we had qualified as at 4 November 2026, across Google (30,466), Meta (15,365) and TikTok (5,151) for 158 UK brands. Your category will deviate from these — that deviation is the analysis.

Two-thirds of ads carry no offer at all

66% of qualified ads have offer_type: none — 33,768 of them. No discount, no free trial, no free shipping — just product, message and brand. If your mental model of competitor advertising is "everyone's always on sale", the data disagrees. Most advertising is evergreen demand capture and brand-building, which is exactly why the minority of ads that do carry offers are worth isolating and watching.

Among the 17,044 ads that do carry an offer:

• Percentage-off dominates: 35% of all offers. Nothing else comes close.
• Finance deals (buy now pay later, 0% APR) take 16% — bigger than most people expect, driven by furniture, mattress and home-improvement advertisers.
• Free trials (14%), free shipping (13%) and fixed amount-off (12%) form the middle tier.
• BOGO is nearly extinct: 72 ads out of 50,982, barely one in seven hundred. If a rival is running buy-one-get-one, that alone is unusual enough to note.
• Only 4.9% of offer-carrying ads show a visible promo code — most discounts are applied on the landing page, which is one reason tracking competitor promotions from ads alone understates what's really running.

Discount depth clusters hard at 20–25%

Across the 6,439 ads with a percentage discount recorded against them (as at 4 November 2026), the median is 25% and the single most common figure is 20% (1,318 ads), followed by 25% (810) and then 50% (783). The average is dragged up to 32.5% by deep clearance ads. Practical read: if your competitor advertises 20% off, they're at the market's default setting. If they advertise 50%+, they're either clearing stock or escalating — check their launch pacing before assuming either.

Treat any discount-depth figure, ours or a tool's, as directional. About 16% of the percentages in our data are our model's read of the brand's prevailing offer rather than a number printed on the creative — see the caveat in our promotions guide. Share-of-offer is the robust metric; depth is worth spot-checking against the ad itself.

Hooks: claims and discounts open most ads

Those shares are of all 50,982 qualified ads; the missing 3.7% are creatives where our pass couldn't identify a single dominant hook, which we'd rather leave out of the table than redistribute.

Social proof, stats and testimonials — the hooks every creative-strategy thread on LinkedIn tells you to use — together open about one ad in ten. Either the advice is wrong or the opportunity is real; both are worth testing, but now you know the baseline you're testing against.

Urgency and tone: calmer than you'd think

58.5% of ads score 20 or below on our 0–100 urgency scale; only 13.1% score 50 or higher. "Hurry, ends tonight" is a niche tactic, not the norm. Tone follows the same pattern: 44.5% of ads read as practical, with aspirational (22.8%) and trust (15.4%) behind. Humour is just 2.1% — 1,058 ads out of 50,982. And despite every best-practice deck saying faces lift performance, only 20.3% of qualified ads contain a human face.

Sector baselines: the comparison that actually matters

A market-wide average is the wrong yardstick for almost everybody. "33% of ads carry an offer" is true across our whole set and misleading in any specific category — a mattress brand and a fintech are playing different games, and measuring either against the blended figure tells you nothing. So here is the same analysis cut by sector, every category where we hold at least 400 qualified ads from three or more brands, as at 4 November 2026.

The spread is the point: offer share runs from 60% to 11% depending purely on what you sell. Four reads worth lifting out of that table:

• Discounting intensity is structural, not strategic. Mattresses and home improvement lead-gen sit at 60% offer-carrying, nearly double the 33% blended baseline, because both sell infrequent, high-ticket, heavily comparison-shopped purchases where a deal is the reason to act today. At the other end, fintech (11%) and charity (12%) barely discount at all, for opposite reasons: you can't put 20% off a current account, and a charity asking for money isn't running a sale. If you're in a low-offer category, a rival suddenly running offers is a much louder signal than the same move in beauty.
• Mattresses are the urgency capital of UK advertising. 31% of mattress and sleep ads score 50+ on urgency, against a 13.1% market baseline — well over double. This is the permanent-sale category, where "ends Sunday" runs every week. Our promotions tracking guide covers how to tell a real deadline from a rolling one, which in this sector is most of the job.
• In travel, the hook outruns the offer. Of all 22 sectors here, travel is the only one where the discount-led hook share (36.5%) exceeds the share of ads that actually carry an offer (33.4%). Discount language is doing more work than discount substance. Travel also prices its deals in pounds about as often as in percentages — 652 amount-off ads against 596 percentage-off — because "£200 off" reads bigger than the 8% it might represent. It isn't alone in that: the car market (303 amount-off against 25 percentage-off) and telecoms (241 against 33) lean on pounds-off far harder still. Which unit a sector quotes in is itself a competitive tell, and it's the reason a pure percentage-depth comparison flatters whoever happens to quote in percent.
• Fashion's offer isn't a discount — it's credit. Fashion shows 33% of ads carrying an offer but only 7% leading on discount, and the reason is in the offer mix: 424 of its offer-carrying ads are finance deals against just 19 percentage-off. Buy-now-pay-later has quietly become the category's standard promotional mechanic. B2B SMB tools show the same offer-without-discount shape (38% vs 9%) for a different reason: the offer is a free trial, which is the single most common offer type in the sector.

Reading a single competitor: two real examples

Baselines only matter when you put a brand against them. Two from our tracked set:

HelloFresh is among the most offer-saturated brands we track: 78% of its qualified ads carry an offer, and 42% open discount-led — nearly double the market's 23%. That's a deliberate acquisition machine built on introductory pricing, and it tells you exactly what you'd be walking into if you competed on price in meal kits.

Gymshark is the opposite pole: 7% of its ads carry an offer and just 2% are discount-led, while 83% read as aspirational. Nearly everything is product-forward brand work — discounts are saved for events, not drip-fed through always-on ads. Interestingly, Lidl shows the same pattern (6% offer-led) for a different reason: when low prices are the brand, you don't need an offer field to say so.

Contrast either with DFS — 46% of its 843 qualified ads carry an offer, heavily finance and percentage-off, with 39% discount-led hooks — and you can see how three spreadsheet columns separate three completely different commercial strategies at a glance. Note too that DFS's 46% looks aggressive against the blended baseline but sits well above its own furniture sector's 30%, which is the comparison that actually indicts it.

A workflow you can actually sustain

1. Pick 3–5 competitors. More than that by hand and you'll quietly stop doing it.
2. Capture monthly, same week each month. Pull each brand's live ads from the four libraries. Screenshot or export — Meta's library only shows what's active right now, so anything you don't capture is gone.
3. Qualify every ad against the rubric above. For 5 brands × ~50 ads this is an afternoon's work. If you need a reporting structure to pour the results into, our competitor ad report template gives you the nine sections with a benchmark column already filled in for each.
4. Compare to baseline and to last month. Offer share up? Discount depth deepening? New hook types appearing? The deltas are the intelligence; a single snapshot is just description.
5. Sanity-check the dates. Library first-shown dates are less reliable than they look — we've measured how and why in how accurate is ad launch data, including the point at which Google's Transparency Center silently stops paginating for big advertisers.

Honest caveat: the manual version of this is genuinely doable for a handful of competitors, and we'd rather you do it badly than not at all. Where it breaks is scale and continuity — qualifying every creative, every week, across four platforms, with history. That's the part we built AdRival to do: the same rubric, run on every ad we capture from a brand. If that's the gap you've hit, the brand pages show live samples of the output and the pricing page covers reports and trackers.

The short version

• Collecting competitor ads is free; the value is in structured qualification, not screenshots.
• Score every ad on offer, hook, objective, tone, urgency and longevity — three fields minimum, 25 if you're serious.
• Baselines from 50,982 qualified ads: 66% of ads carry no offer; percentage-off is the dominant offer type at a median 25% depth; bold claims and discounts open nearly half of all ads; high urgency and humour are rare.
• Compare against your sector, not the market. Offer share ranges from 60% in mattresses to 11% in fintech — the blended average flatters some categories and damns others.
• Analyse deviation from baseline, and deltas over time — not snapshots.