Kova Commerce · Clementine & Mint

Google Ads Audit

Account 8007771142 · 28 September 2026 · windows: last 30 / 90 / 365 days, plus Q4 2025 for seasonality

ROAS · 365d
9.89
£9.89 back per £1
Revenue · 365d
£79,550
1,537 orders
Spend · 365d
£8,040
avg £22/day
Non-brand ROAS · 365d
7.07
excluding searches for your name
The short version This account is well built — clean structure, correct tracking, strong returns. The single biggest opportunity is not a fix, it's a release: PMax has been capped at its daily budget for six weeks, heading into the two months that earn the most all year.

1Scorecard

What's already right, and what we'd change.

Working well — don't change

  • ROAS 9.89 over 12 months on £8,040 spend
  • Conversion tracking is clean — no double counting, data-driven attribution live, values agree to within 2.5%
  • Brand excluded from PMax correctly, so Shopping isn't buying your own name
  • Brand search is exceptional — ROAS 119, Quality Score 10/10
  • Feed-only PMax is a deliberate, sound choice — keeps spend on Shopping
  • Headbands earn their share — best conversion rate in the account at 6.87%
  • Clean naming and structure — easy to manage and extend

Where the money is

  • Raise the PMax budget — it has spent its full budget for six weeks
  • Show up on 90%+ of searches for your name — raise brand max CPC 10p → 20p, budget to £4/day
  • Rewrite product titles — customers search "bridal bag", not "pouch"
  • Restock the Medium tartan pyjamas — £368 last Q4, sold out so not showing
  • Fund the starved winners — laptop sleeves, robes, bows
  • Standardise product images — same size, same framing, 2048 × 2048
  • Ramp budget for Christmas on a dated calendar

2Brand vs non-brand

Two very different jobs, so we judge them separately.

Brand is people typing "Clementine & Mint" into Google. They already know you, so it's cheap and returns a lot. Non-brand is Performance Max showing your products to people searching for things like "bridal headband" or "velvet bow". That's where new customers come from, and it's the number to judge the budget on.

PeriodTypeSpendRevenueROASShare of revenue
Last 12 monthsBrand£203£24,107118.930%
Non-brand£7,837£55,4437.0770%
Last Christmas
Oct–Dec 2025
Brand£71£9,574135.142%
Non-brand£1,771£12,9977.3458%
Last 90 daysBrand£49£4,40090.521%
Non-brand£2,395£16,5056.8979%
Last 30 daysBrand£17£98658.420%
Non-brand£746£3,9035.2380%
What this means The 9.89 headline ROAS includes brand search. Brand is 2.5% of spend but 30% of revenue, so it lifts the average. The honest number for growth is non-brand at 7.07, and that's the one we use for every budget recommendation below. At Christmas brand jumps to 42% of revenue, because far more people search for you by name. That's why the brand changes in section 8 matter most in November and December.

Brand is cleanly separated: your name is excluded from Performance Max, and we checked every one of its search categories to confirm your name isn't leaking in.

3The budget is the bottleneck

PMax has spent its whole budget for six weeks. It averaged £27.63 a day on a £28 budget. When the budget was cut to £22 on 11 September, spend dropped straight to £22.20 a day. The budget is setting the spend, not demand.

£0£10£20£30£402026-08-14: £33.71 of a £28 budget2026-08-15: £24.26 of a £28 budget2026-08-16: £32.38 of a £28 budget2026-08-17: £24.06 of a £28 budget2026-08-18: £28.42 of a £28 budget2026-08-19: £26.21 of a £28 budget2026-08-20: £23.86 of a £28 budget2026-08-21: £30.47 of a £28 budget2026-08-22: £21.85 of a £28 budget2026-08-23: £33.94 of a £28 budget2026-08-24: £29.77 of a £28 budget2026-08-25: £25.44 of a £28 budget2026-08-26: £24.55 of a £28 budget2026-08-27: £26.13 of a £28 budget2026-08-28: £24.98 of a £28 budget2026-08-29: £27.22 of a £28 budget2026-08-30: £21.25 of a £28 budget2026-08-31: £29.48 of a £28 budget2026-09-01: £22.33 of a £28 budget2026-09-02: £37.61 of a £28 budget2026-09-03: £25.06 of a £28 budget2026-09-04: £21.06 of a £28 budget2026-09-05: £38.01 of a £28 budget2026-09-06: £33.52 of a £28 budget2026-09-07: £28.46 of a £28 budget2026-09-08: £33.62 of a £28 budget2026-09-09: £27.52 of a £28 budget2026-09-10: £24.04 of a £28 budget2026-09-11: £22.01 of a £28 budget2026-09-12: £22.45 of a £22 budget2026-09-13: £22.4 of a £22 budget2026-09-14: £19.62 of a £22 budget2026-09-15: £18.47 of a £22 budget2026-09-16: £28.23 of a £22 budget2026-09-17: £23.38 of a £22 budget2026-09-18: £17.46 of a £22 budget2026-09-19: £28.86 of a £22 budget2026-09-20: £21.94 of a £22 budget2026-09-21: £20.82 of a £22 budget2026-09-22: £17.8 of a £22 budget2026-09-23: £23.26 of a £22 budget2026-09-24: £26.54 of a £22 budget2026-09-25: £22.5 of a £22 budget2026-09-26: £20.93 of a £22 budget2026-09-27: £20.47 of a £22 budgetavg £27.63 a day on a £28 budgetavg £22.20 on £22 14/0825/0805/0916/0927/09
Daily PMax spendDaily budget: £28, cut to £22 on 11 Sep
Why this matters right now The same thing happened last Christmas. PMax hit its cap on 18 of 61 days in Nov–Dec 2025, 14 of them in December — spending £42 on a day budgeted at £22, while returning a ROAS of 11–15. The best weeks of the year were rationed.

Note: the campaign's own "budget lost impression share" reads 1.3%, which looks healthy. That metric is unreliable on Performance Max — spend ÷ budget is the real test.

Recommendation

CampaignTodayRecommend nowReason
PMax — All Products£22/day£32/daySpends its full budget; was £28 until 11 Sep
Brand — UK Search£1.50/day£4/dayWill bind once CPC rises + Q4 demand

Also worth adding a target ROAS of 5.5–6.0 to PMax. It currently runs on Maximise Conversion Value with no floor, so there's nothing protecting efficiency as budget goes up.

4Christmas is the whole game

November and December carry the year. The pattern has repeated two years running.

010202025-08: ROAS 7.69, £4,771 value2025-09: ROAS 7.12, £4,369 value2025-10: ROAS 6.7, £3,729 value2025-11: ROAS 16.85, £8,805 value16.92025-12: ROAS 13.17, £10,038 value13.22026-01: ROAS 8.66, £4,052 value2026-02: ROAS 5.61, £2,614 value2026-03: ROAS 9.04, £4,705 value2026-04: ROAS 11.17, £6,748 value2026-05: ROAS 11.66, £9,385 value2026-06: ROAS 9.5, £8,221 value2026-07: ROAS 10.86, £9,399 value2026-08: ROAS 7.44, £6,470 value2026-09: ROAS 6.75, £3,400 valueA25SONDJ26FMAMJJAS
November & December

Where the efficiency actually sits, week by week

01020w/c 2025-09-29: ROAS 7.44, £18.96/day7w/c 2025-10-06: ROAS 5.34, £20.43/day5w/c 2025-10-13: ROAS 5.06, £18.49/day5w/c 2025-10-20: ROAS 10.36, £15.45/day10w/c 2025-10-27: ROAS 5.98, £15.15/day6w/c 2025-11-03: ROAS 14.5, £16.7/day15w/c 2025-11-10: ROAS 15.66, £15.97/day16w/c 2025-11-17: ROAS 16.19, £16.31/day16w/c 2025-11-24: ROAS 23.48, £21.07/day23w/c 2025-12-01: ROAS 15.6, £26/day16w/c 2025-12-08: ROAS 11.05, £33.29/day11w/c 2025-12-15: ROAS 14.89, £26.97/day15w/c 2025-12-22: ROAS 10.98, £16.63/day11w/c 2025-12-29: ROAS 11.1, £13.91/day1129Sep6Oct13Oct20Oct27Oct3Nov10Nov17Nov24Nov1Dec8Dec15Dec22Dec29Dec
BFCM week — ROAS 23.5High-efficiency window (3 Nov – 21 Dec)

Early and mid October is the weakest stretch of Q4 (ROAS 5–6). Don't ramp then — ramp from late October.

Budget calendar

PeriodPMax /dayBrand /dayWhy
Now → 12 Oct£32£4Back above the old £28
13 – 26 Oct£30£4Weakest ROAS stretch — hold
27 Oct → 16 Nov£40£5ROAS climbs to 14–16
17 – 30 Nov BFCM£55–65£8Peak efficiency: ROAS 23.5, CVR 8.9%
1 – 18 Dec£55£8Highest volume; 14 capped days last year
19 – 24 Dec£35£5Last posting dates
25 Dec → 5 Jan£22£3Volume collapses

Step up by about 30% at a time, and make the BFCM increase by 17 November so the campaign is settled before peak. Judge it on marginal return, not blended ROAS.

5Product mix

Headbands dominate — and they've earned it. The problem is elsewhere.

Headbands
53.7%
7.81
Pouchesfell to ROAS 4.31 last 30d
20.4%
5.65
Children's headbandROAS 8.61 last Q4
6.2%
2.21
Bows30% of Q4 spend
4.1%
6.57
Laptop sleeves
3.9%
10.81
Robeshighest AOV ~£214
2.9%
10.02
Pyjamas
3.0%
2.33
Scrunchies
2.5%
9.42
Children's bow
1.3%
2.19
Spend shareROAS (account average 6.89)
Good news Headbands take 53.7% of spend and return 61.6% of conversions at the best conversion rate in the account. They're not crowding out better performers — they are the best performer.
The actual problem Pouches are the second-biggest spend line and fell from ROAS 5.65 to 4.31 in the last 30 days. The single biggest-spending product in the account is a pouch returning 4.21 — well below average. Section 6 explains why, and it's fixable.

Underfunded winners

LineShare of spendROASNote
Laptop sleeves3.9%10.81Brown Gingham sleeve returns ROAS 19.90, the second-best product in the account
Robes2.9%10.02Highest AOV in the catalogue, ~£214/order
Scrunchies2.5%9.42ROAS 12.05 last Q4
Bows4.1%6.57Quiet now, but 30% of spend last Q4
Why they can't be funded today All three live product groups overlap — the "All Products" group already contains 90 of the 92 headband products and 39 of the 42 pouch products. Product groups don't have their own budgets, so the catch-all wins the spend. Funding these lines needs a separate campaign, not another product group.

6Product titles

The highest-value fix on this list, and the data tells us exactly what to write.

Nobody searches "pouch" They search bridal bag satin (4.5 sales), bridal bag (4.3) and bridal clutch (2.9). The best-selling pouch is titled "The White Ruffled Pouch" — it only ranks because the boilerplate at the end happens to contain "Bridal Clutch bag".

What customers actually type

SearchSalesValue
wedding guest headband8.9£377
velvet alice band8.2£458
butter yellow headband7.6£346
linen headband women7.1£310
pale blue headband6.3£384
padded headband6.1£298
bridal bag satin4.5£259
bridal bag4.3£203
bridal clutch2.9£140

The winning pattern is colour + material + product + occasion. Nearly every title instead starts with "The" and ends with the same brand boilerplate.

Suggested rewrites

NowThe White Ruffled Pouch | Luxury Lifestyle Accessories by Clementine & Mint, Personalised gift, Bridal Clutch bag.
BetterWhite Linen Bridal Clutch Bag | Ruffled Personalised Makeup Pouch
NowTHE LIGHT BLUE SATIN HEADBAND' | Luxury Lifestyle Accessories, Linen Headbands by Clementine & Mint'
BetterLight Blue Satin Headband | Pale Blue Wedding Guest & Bridal Headband
NowTHE BROWN SATIN HEADBAND' | Luxury Lifestyle Accessories, Linen Headbands by Clementine & Mint'
BetterBrown Satin Headband | Slim Alice Band for Women, Wedding Guest

How these were built. Start with colour + material + product, so one title catches both the broad search ("brown headband", 6 sales) and the specific one ("brown satin headband", 2 sales from 8 clicks). Then add the words people actually bought with: "wedding guest" (9 sales), "bridal" (5), "pale blue" (6), "alice band" as two words (11), "for women", "makeup bag". Leave out words nobody searches, like "handmade", "luxury" and "hair accessory". They belong in the description. Only use words that are true of the product: the brown headband is slim, so "padded" would pull in people who then don't buy, and the white pouch is linen, so the satin searches ("bridal bag satin", "white satin bag", about 10 sales) belong on the satin pouches' titles instead.

Length. Google allows 150 characters, but shoppers see far less: roughly the first 70 on desktop and around 30 on a phone, where most of your sales happen. So the first 30 characters must say what the product is (all three above do), and the whole title should aim for 70 or under. That's why "Clementine & Mint" is gone from the end: your shop name already shows on every Shopping ad, and your name is excluded from Performance Max anyway. Your current titles run to about 100 characters, and most of the visible part goes on "THE", capitals and the brand boilerplate rather than words people search.

What's wrong, by the numbers

IssueProductsImpact
Written in ALL CAPS187 of 234Reads as spam; 68% of spend
Start with the word "The"233 of 234Wastes the highest-weighted position
End in the same brand boilerplate228 of 234Spends title space on words nobody searches
Stray apostrophe in the title (e.g. HEADBAND' |)mostLooks unfinished in the Shopping tile
Pouches & pyjamas tagged "Linen Headbands"14Actively tells Google the wrong category
Duplicate titles competing with each other6 listingsSplits sales data; one twin gets 0 sales
Product groups built from hand-typed item IDs282 IDsNew products stay off until added by hand

Start with the top 20 products by spend — they carry 54% of spend, so the work is bounded.

Build product groups from product type, not item IDs

Right now your campaigns pick products from 282 item IDs typed in by hand. Any new product stays switched off until someone adds its ID. Every live product is included right now, but only because the list is being kept up to date by hand.

You don't need to change anything in the feed to fix this. Every product already has a product type in Shopify ("headband", "bow", "pouch" and so on), and Google Ads can build product groups straight from product type. A new headband then joins the headband group on its own.

Product groupShopify product types in itProducts
Headbandsheadband69
Bowsbow39
Scrunchiesscrunchie, scrunchie with bow33
Bagspouch, laptop sleeve25
Kidschildren's headband, children's bow, the little one43
Nightwearpyjamas, robe, women's stripe set13
Decorationsstocking decoration10
What to do 1. Keep product type filled in on every product in Shopify, including new ones. Only two are blank today: the Personalised Linen Gift Bag and the Red Linen Scrunchie.
2. The Blue Gingham Bunny Basket is typed "stocking decoration". Give it its own type (e.g. "easter basket") so it doesn't run with the Christmas stockings.
3. When the product groups are rebuilt, build them on product type instead of item IDs. It's safe to tidy the product types any time, because nothing changes in the ads until the groups are switched over.

This is also what makes the January campaign for laptop sleeves, robes and bows (section 5) a few clicks rather than a rebuild. Custom labels are only worth using for things product type can't say, such as tagging your velvet and tartan pieces as "christmas" across every category. That's optional.

7Don't prune before Christmas

Products that look dead in September are the ones that carry December.

The trap The Burgundy Velvet Maxi Bow returned £1,649 on £268 last Q4 — the biggest product of the year. Today it shows £10.92 and zero sales, which puts it on any "underperformer" list.
ProductNow (90d)Last Q4Verdict
Burgundy Velvet Maxi Bow0 sales£1,649 · ROAS 6.15Leave alone
Red Velvet Scrunchie0 sales£215 · ROAS 62.68Leave alone
Navy Velvet Headband0 sales£260 · ROAS 22.71Leave alone
Black Watch Tartan HeadbandROAS 1.36£589 · ROAS 13.56Leave alone
Children's headbands (line)ROAS 2.21ROAS 8.61Hold, review January
Three of last Christmas's products aren't showing — we checked why They earned £864 last Q4 and get no impressions today. All three are still set up in the campaign, so nothing is broken in Google Ads. Two are no longer on your website, and one is sold out in one size.
ProductLast Q4TodayWhat to do
Black Watch Tartan Pyjamas — Medium£368 · ROAS 9.06Still on sale. Medium is sold out; Small and Large are runningRestock Medium before November if you can. It switches back on by itself.
Cream Tweed Headband£265 · ROAS 19.24No longer on your website (last ad Feb 2026)Nothing, unless it's coming back
Navy Velvet & Tartan Ruffled Pouch£232 · ROAS 7.51No longer on your website (last ad Nov 2025)Nothing, unless it's coming back

If the tweed headband or the tartan pouch is coming back this winter, let us know. Both were among last Christmas's most efficient products. It also explains why the search "tweed headband" got 19 clicks and no sales: there's no tweed headband to buy any more. If tweed isn't returning, add "tweed" as a negative keyword.

What to actually pause

ProductSpendWhy
Navy Blue Linen Aliceband£17.97ROAS 0.20. The "pre order June delivery" wording has now come off the title, so give it three weeks and pause it if sales don't follow
Duplicate Black Watch Tartan Pyjamas listing£27.6373 clicks, 0 sales, while its identical twin sells

8Brand search

The goal: your ad shows on 90% or more of Google searches for your name. Today it's about 85%.

Genuinely excellent ROAS 119 over 12 months. Quality Score 10/10. Ad strength Excellent. Last Christmas it returned £4,462 on £26 of spend in November.
KeywordMatchImpression shareLost to rankROAS
clementine and mintExact93.2%6.7%106
clementine and mintPhrase57.8%40.2%38
clementine & mintPhrase74.2%25.8%49
Why 90% Someone typing "Clementine & Mint" has already chosen you. These are the cheapest sales in the account, at about 4p a click. At 85%, roughly 1 in 7 of those searches doesn't show your ad, which leaves the top of the page open to a competitor or reseller. Impression share has sat at 82–86% every month since March.

Your exact-match keyword is already at 93%. The ~85% campaign average is dragged down by the two phrase keywords, and all of that loss is bid, not budget (budget lost: 0.5%). So the bid is the only lever.

Two changes 1. Raise max CPC from 10p to 20p to get impression share to 90%+. Quality Score is already perfect and you're only paying 3.8p a click, so the 10p ceiling is the only thing holding you back. Check again after two weeks. If it's still under 90%, go to 25–30p.
2. Raise the budget to £4/day at the same time, or a higher CPC plus Christmas demand will simply hit a different ceiling.

One correction to an earlier assumption

Search partners were worth turning off — but they delivered 4 impressions and £0.00 in 90 days. It's tidy housekeeping, not a performance fix, and it won't move impression share.

9Product imagery

The same kind of product looks bigger in one photo and smaller in the next.

Six of your current headband main images, shown side by side at the same width. The Cornflower Blue fills well under half its frame while the Brown Houndstooth and Forest Green fill most of theirs. Some stand upright and face the camera, others are shot from the side. Next to each other in a Google Shopping grid, that makes some headbands look bigger or more premium than others for no reason.

MeasureFinding
Main images in a 4:5 shape140 of 227
Main images in a 5:6 shape76 of 227
Other shapes11 of 227
Different image sizes across all 959 product photos88
Square photos1 of 959
The standard to use for every main image Main image: 2048 × 2048 pixels, square. This is the one to get right first. It's the most versatile shape: Google shows it cleanly in Shopping results and when it reuses your product images in remarketing across Display, YouTube and Discover. It's also Shopify's recommended size, and Google recommends 1500 × 1500 or larger.
Vertical version (optional): 1676 × 2048 pixels (a 9:11 shape). Google's guide for clothing and accessories names 9:11 as the ideal shape for vertical images. Google decides when to show a taller layout, not you, so keep the square as the main image and add the vertical as an extra image. Your current 4:5 photos are close to this shape already, so recropping them is a good start.
Framing: the product fills 75–90% of the frame, which is Google's own guidance, and the same amount on every product in a category.
Angle and position: the same angle for every product of a type (for example, every headband upright and front-on), centred, and on the same background.

How to keep it consistent. When shooting, keep the camera on a tripod at a marked distance so every headband comes out the same size. When editing, make one 2048 × 2048 template (Canva works) with guide lines 5–10% in from each edge, and size every product to touch the same guides. The background doesn't have to be white. What matters is that it's the same on every main image. Keep lifestyle and on-the-head shots as the second and third images. Start with the top 30 products by spend.

10Conversion tracking

We stress-tested this before trusting any number in the report. It's healthy.

CheckResultDetail
Double countingCleanThree purchase tags exist; only one counts. Correct.
AttributionData-drivenGenuinely active, not silently falling back to last-click
Value accuracyWithin 2.5%AOV agrees across all three tags (£51.13–£52.34)
Default-value leakageNoneThe £43.92 fallback never fired
Tag outagesNoneFired on 353 of 365 days; gaps are normal variance
Lookback windowRight-sized73% of sales happen within a day
GA4 agreement−28%GA4 undercounts despite a 3× longer window
Enhanced conversionsVerifyNot visible via the API — check in the interface
One thing to know GA4 reports about 28% fewer Google Ads sales than actually happened. It doesn't affect bidding, but any comparison you make in GA4 will make Google Ads look worse than it is.

Low priority, for when you have time. Google Ads shows three warnings on the purchase tag. Sales are recording properly, so none of this affects your results today. They're tidy-ups that give Google a bit more detail about each order:

The easiest way to fix them is to ask Claude to walk you through it. Paste in something like: "My Google Ads purchase conversion says sold item IDs can't be matched to Merchant Center, many conversions don't have cart data, and cost of goods isn't available. I'm on Shopify with the Google & YouTube app. Walk me through fixing each one, step by step."

11Smaller wins

FindingCostAction
11 live products still have "coming-soon" in their web address—Rename the URLs in Shopify (tick "create redirect")
No negative keywords anywhere~£20/moAdd 4–6 exact-match negatives only (see note)
Desktop converts at just over half the mobile rate3.32% vs 5.86%Worth a look at the desktop product pages
Christmas Collection asset group paused (inside the main PMax, where it ran last Christmas: £302 spent, £2,428 back)No extra budgetTurn it back on. Leave the separate "Performance Max - Christmas" campaign off: it has never run
Tracking template writes campaign numbers into GA4—Swap to campaign names
Duplicate paused brand campaign—Delete to keep things clean
Worth saying — what we did NOT recommend The obvious negative keywords turned out to be profitable: "fascinator" returns £344 across 27 searches, "makeup bag" £280, "tartan" £223. Blocking them would have cost real revenue. The genuine waste is about £20 a month — worth ten minutes, not a project.

12What to do, in order

Sequenced so everything is settled before the peak.

This week
1Raise PMax to £32/day and brand to £4/dayThe budget has been the limit for six weeks. 5 minutes.
2Restock the Medium tartan pyjamas£368 last Christmas. Small and Large are still running.
3Raise brand max CPC 10p → 20pTarget: 90%+ impression share on your name, up from ~85%
4Pause the duplicate pyjama listing73 clicks, 0 sales, while its twin sells
5Add the 4–6 exact-match negatives~£20/month, 10 minutes
Before 26 October
6Rewrite titles for the top 20 productsLead pouches with "bridal clutch / bag"
7Tidy product types in ShopifyTwo blanks and the Easter basket. Product groups can then be built on type. See section 6.
8Turn the Christmas Collection asset group back on (inside the main campaign)It ran there last Christmas. Leave the separate Christmas campaign off
9Confirm enhanced conversions is recordingFree accuracy gain before peak
10Consistent square images for the top 30 products2048 × 2048, product filling 75–90% of the frame
The Christmas ramp
1127 Oct — go to £40/dayROAS climbs to 14–16 from here
12By 17 Nov — go to £55–65/dayPeak week: ROAS 23.5. Set it early so delivery is stable.
1319 Dec — step down to £35, then £22Volume falls after last posting dates
January
14Split non-headbands into their own campaignThe only way to genuinely fund laptop sleeves, robes and bows
15Review the children's lines on a full Q4Judge them on their real season, not September