Paid Social · Traffic Routing · 90-day window
Where should Meta
ad traffic land?
Every spending Meta ad over the last 90 days (21 Apr - 20 Jul 2026), classified by the page it actually sends people to, then scored on conversion and return. Because Meta is a discovery channel for Pott, performance is measured on first-click attribution - crediting the landing page that introduced the eventual buyer, not the page that happened to close them. Subscription and retention campaigns are excluded, since they target existing customers and inflate the numbers.
Meta spend analysed
£77,857
414 ads, subscription excl.
Blended Meta ROAS
1.95x
CPA £37.4
Site sessions
144k
92.9% new visitors
Site conversion
2.96%
6.48% add-to-cart
Destinations classified
98.8%
of spend
The verdict
Collection page
PRIMARY
2.17x
first-click ROAS
£38,008spend
£33.4CPA
£72AOV
Home page
TOP-OF-FUNNEL
2.29x
first-click ROAS
£27,877spend
£23.5CPA
£54AOV
Single product page
AVOID (cold)
1.32x
first-click ROAS
£11,002spend
£49.8CPA
£66AOV
Send Meta traffic to the relevant scent/product collection by default. It is the only destination that stays profitable under every attribution model, at the largest scale, with the highest order value (£72).
The home page is a legitimate top-of-funnel introducer once you credit first-click - it starts a lot of journeys that close later - but it has the lowest AOV and rarely closes in-session, so pair it with retargeting and email.
Dropping cold prospecting straight onto a single product page is the weakest play on every lens.
A dedicated cold-acquisition landing page is still untested - the only prior landing-page spend was subscription retention (now excluded), so there is no clean signal yet. It remains the clearest experiment to run.
Destination performance — first-click
First Meta click: the landing page credited with introducing each eventual buyer. 2,579 Meta-attributed orders. Breakeven marked at 1.00x.
| Destination | First-click ROAS | Spend | Orders | Revenue | CPA | AOV |
| Collection |
|
£38,008 | 1,139 | £82,562 |
£33.4 | £72 |
| Home |
|
£27,877 | 1,185 | £63,952 |
£23.5 | £54 |
| Single product |
|
£11,002 | 221 | £14,541 |
£49.8 | £66 |
Does the ranking hold?
The same spend re-scored under three attribution models. Collection leads or ties on every one. Home strengthens under first-click (its discovery role) but weakens under multi-touch; product trails throughout. This is why the conclusion is not an attribution artifact.
| Destination |
First Meta click
primary |
First-click strict
Meta = absolute first touch |
Linear all
multi-touch |
Meta-native
7d-click / 1d-view |
|---|
| Collection | 2.17x | 1.50x | 1.33x | 2.26x |
| Home | 2.29x | 1.20x | 0.92x | 1.55x |
| Single product | 1.32x | 0.61x | 0.66x | 1.37x |
On-page conversion & intent
Meta-reported behaviour of clickers who actually landed. This isolates what the page itself does: add-to-cart rate and landing-to-purchase conversion. Collections convert clickers best; a lone product page converts worst and adds to cart least.
| Destination | Landing views | Add-to-cart rate | Conversion (of landings) | AOV |
| Collection | 23,616 |
18.1% | 4.81% |
£76 |
| Home | 14,217 |
19.8% | 4.28% |
£71 |
| Single product | 5,924 |
15.6% | 3.54% |
£72 |
Top landing pages by spend
The specific URLs Meta budget flows to today, scored on first-click ROAS. Scent and product collections dominate the profitable end; single product pages and the homepage sit at the lower-return end.
| Landing page | Type | Spend | First-click ROAS | CPA |
| / | Home |
£27,877 | 2.29x | £23.5 |
| /collections/cook-candles | Collection |
£12,728 | 1.29x | £57.3 |
| /collections/tomato-vine-refillable-candle | Collection |
£9,567 | 3.03x | £23.5 |
| /collections/the-outdoor-collection | Collection |
£9,282 | 2.13x | £39.8 |
| /products/ocean-seasalt-starter-pack | Single product |
£6,896 | 1.10x | £57.5 |
| /collections/roses | Collection |
£6,270 | 2.09x | £32.3 |
| /collections/seasalt-collection/products/ocean-seasalt-starter-pack | Single product |
£1,313 | 1.01x | £69.1 |
| /products/seasalt-refillable-candle-starter-pack-standard-size | Single product |
£1,231 | 1.05x | £82.0 |
| /products/refillable-candle-starter-pack-old-copy | Single product |
£495 | 3.26x | £17.7 |
Method & the one caveat
How this was measured
- Attribution. Since Meta introduces buyers who often close via email, direct or branded search, last-click understates it. Primary lens is first-click (Triple Whale journey data), crediting the Meta ad that started the journey. Three further models are shown so no page wins on an artifact.
- Destination truth. Most Pott ads are page-post ads whose link is not exposed by Meta's API, so the true landing page for every ad was reconstructed from 8,631 Triple Whale customer journeys - correlating each ad click to the next on-site page load. Actual arrival behaviour, not the creative's stated link (the two disagreed on 19% of ads). 98.8% of spend classified.
- Exclusions. Subscription and retention campaigns (39 ads, £1,641, including the /pages/subscribe-save opt-in page) are removed. They target existing customers on a warm audience, which inflates conversion and return and would skew an acquisition-routing decision.
- Return figures use Triple Whale attributed revenue against Meta spend; on-page conversion uses Meta-reported landing views.
The honest caveat. Destination is confounded with campaign quality - seasonal launch campaigns (strong creative, real demand) tend to point at collections, while broad prospecting points at home. The mechanism is sound and holds across every de-duplicated lens, but to prove it is the page and not the campaign, run one clean test: same creative, same audience, split destination collection vs single product vs a new landing page.
What users do on the page
Microsoft Clarity behaviour by destination type. A 3-day directional snapshot of live traffic (the Sea Salt test pages were not running in-window, so these reflect the page types in general). It supports the ranking: collections get explored, product pages get ignored, and the home page gets bounced back from fastest.
| Destination | Avg scroll depth | Active engagement | Quick-back rate | Dead-click rate |
| Collection | 47% | 13s | 15.5% | 0.8% |
| Home | 25% | 17s | 19.1% | 0.0% |
| Single product | 34% | 4s | 10.0% | 3.7% |
Read: collections show the deepest scroll (people browse the range); single product pages show the shallowest engagement (about 4 seconds) and the most dead clicks (disengagement on a lone PDP); the home page has the highest quick-back rate, so cold clickers bounce back from it fastest. To learn the why, read the click, scroll and rage maps for these pages in the Clarity dashboard (the API does not export the visual maps).
Recommended moves
1Default to the scent/product collection. Make the relevant collection the standard destination for prospecting and launch campaigns. It is the proven best blend of scale, return and order value.
2Keep home for pure top-of-funnel, not as a closer. It earns its first-click return by introducing buyers; support it with retargeting + email flows to finish the sale rather than expecting same-session conversion.
3Stop sending cold traffic to a single product page. Worst return on every lens and lowest add-to-cart; if a PDP must be used, wrap it in a collection or landing experience.
4Build and test a dedicated cold-acquisition landing page. Collection-style: refill-ritual story, multiple scents, proof. There is no clean read on this yet (prior landing-page spend was subscription retention, now excluded), so treat it as the priority experiment rather than a proven winner - run it against the top collection.
5Run the de-confounding test (same creative + audience, split destination) to lock the collection-vs-PDP-vs-LP decision with clean causal evidence.