Account 13859523 · GBP
Paid social · account audit
In October, 83% of spend went to three campaigns that produced three purchases between them. Reported ROAS fell to 0.5x, because the ads that actually sell were off. This audit covers what worked, what the money went on instead, the most likely cause of the drop, and the structure to rebuild on.
The only live campaigns are Awareness, Engagement and a Start Trial lead campaign.
Reported ROAS in October, with purchase ads off
Of October spend — £886 of £1,069 — on three campaigns
Purchases those three produced between them
Target CPA on a £40 order, roughly 3x
The band is the account's normal 2024–25 range. Hover any bar for detail.
The full list is in section 07.
These are Meta's reported figures, not an independent measurement. They are used here to compare periods within the same account on the same basis, not as a claim about incremental revenue.
CPM stayed around £4.50 to £4.80. Purchases per landing-page view fell from roughly 10–15% in late 2025 to May 2026, down to 6–8% between June and September 2026. The problem is who is clicking and what they see, not what the auction costs.
Lifetime Bundles: £131.9k spent, 9,357 purchases, £14.09 CPA, 2.92x. Lifetime Coveralls: £67.5k spent, £11.50 CPA, 2.88x.
Bundles frequency is 16.8 lifetime. Account frequency ran between 6 and 11 through 2024.
| Campaign type | CPA | ROAS |
|---|---|---|
| Boosted organic posts | £4.30 | 7.4x |
| Dark ads to warm audiences | £4.80 | 6.65x |
| Generic dark ads | £7.50 | 4.4x |
These campaigns used a different attribution window, which adds engaged-view. Without Weaning Week, May ran at about 1.7x. The direction is real; the multiple is not directly comparable with the rest of this audit.
| Audience | ROAS | CPA | Note |
|---|---|---|---|
| Website visitors, 30 days | 5.2x | £7.11 | |
| Add to cart or view content, 14 days | 4.8x | £7.63 | |
| Customers / LTV | 4.7x | £7.19 | At frequency 65 |
| Engagers, 365 days | 3.8x – 4.7x | — |
Small samples, but they point the same way as Weaning Week.
Retargeting, 1% lookalikes, and the organic-post style ads. All three were performing when they were switched off.
June to October 2026 unless stated. Lifetime figures for Start Trial.
Poor placements together account for about £2.1k between June and October 2026.
After the May promotion ended, the account returned to its normal 2x to 2.6x.
Putting the best ads into a single Advantage+ campaign produced £15.2k spent at £17.73 CPA and 2.26x, on a CPM of £5.92 — against £4.40 to £5.60 in the Bundles campaign, which had been running 2.6x to 3.05x from April to June.
The data fits the consolidation hurting performance, but it does not prove it caused the drop.
Three purchase-optimised campaigns only: prospecting, retargeting, and one promo or event campaign. Target CPA at or below about £13 on a £40 order, which is roughly 3x.
Advantage+ broad, plus a 1% customer-list lookalike test. Optimise for the
new-customer purchase event Leaf_NewCustomer_Purchase, which already exists
in the pixel. Exclude existing customers. Remove Facebook Reels, Marketplace, Threads and
in-stream.
Rebuild website visitors at 14 days and add-to-cart at 7 days. Cap frequency — 17 to 65 is far too high. Keep the budget small, because these pools are limited.
Weaning Week is the model: education-led, organic-post and creator-style. The one structure in this account that has produced a step change rather than a drift.
Lifetime figures except the June consolidation. Anything past the dashed line is losing money at a £40 order value.
Organic-post and creator-style ads, education-led angles like Weaning Week, and three to five new angles a month. Static "62% off" bundle ads are now the baseline, not the winner.
Lead with Bundles, which carries the higher order value — £41 against £33 for Coveralls. Test Coveralls separately.
One change at a time, and leave a campaign alone for five to seven days.
| Step | Completion | Read |
|---|---|---|
| Landing page → add to cart | — | The weak step. Test the product page for cold traffic and show the offer earlier |
| Add to cart → checkout | about 25% | |
| Checkout → purchase | about 70% | Working |
Tracking is in good shape. Purchase match quality is 9.3 out of 10 and server events are firing today; add-to-cart and checkout are at 7.4. The one gap: Click ID capture on purchases is only 19%, and that needs fixing.
Unless the trial has a business reason beyond the ad account. Together they are the majority of current spend and almost none of the purchases.
Not covered by this audit.
As the new prospecting test. Both were working when they were stopped.
Give each one five to seven days.