Why this matters
At the intermediate level you split a catalogue by margin for one channel. At scale the harder question is what each group is for: a thin-margin best seller may earn its place by winning new customers, short-dated stock is judged on the cash it recovers, and a launch needs exposure before it can be judged. Those roles must hold across channels, or one platform undoes another's logic. The Performance Max structure guide says to split by target, not taxonomy, and the Meta catalogue ads guide shows product sets carrying business logic into prospecting and retargeting.
The store and its single campaign
Peakform Nutrition is an invented online store selling third-party and own-brand sports nutrition. It runs one Performance Max campaign over every product with no brand exclusion, and one Meta catalogue ad set showing the full catalogue to past visitors and new people alike.
Peakform Nutrition: one month by product group
Monthly spend: Performance Max £20,000, Meta catalogue ads £10,000. Margin is contribution before marketing; for short-dated stock it is price less delivery and fees, since unsold stock is written off.
| Group | Margin | Stock | PMax spend | PMax ROAS | PMax orders | Meta spend | Meta ROAS |
|---|---|---|---|---|---|---|---|
| Branded whey (third party) | 18% | Deep | 46% | 5.2 | 1,000 | 40% | 3.9 |
| Own-brand protein | 48% | Deep | 12% | 4.1 | 250 | 15% | 3.2 |
| Own-brand vitamins | 55% | Deep | 6% | 3.5 | 170 | 8% | 2.6 |
| Pre-workout (third party) | 26% | Medium | 18% | 4.6 | 470 | 17% | 3.4 |
| Short-dated stock, eight weeks to clear | 60% | Fixed | 8% | 6.1 | 490 | 12% | 4.8 |
| Plant protein, launched six weeks ago | 45% | Deep | 2% | 1.8 | 18 | 3% | 1.5 |
| Shakers and accessories | 35% | Deep | 8% | 2.4 | 320 | 5% | 2.0 |
- Search term insights: Peakform brand terms carry about 20% of Performance Max conversion value.
- Of customers whose first order was branded whey, 30% bought own-brand products within six months.
- About 70% of Meta purchases come from people who visited the site in the previous 30 days.
- The owner wants 8 points of revenue kept as profit after ad cost.
| Filter view | Search demand | Products shown |
|---|---|---|
| Whey protein, chocolate | High | 34 |
| Whey protein, vanilla | High | 29 |
| Whey protein, 2.5kg | Moderate | 18 size variants of products on the parent page |
| Pre-workout, caffeine-free | Moderate | 11 |
| Plant protein, strawberry | Very low | 2 |
| Any flavour plus size pair | None measured | 1 to 6 |
Classifying products by margin and role
Give each group one role. A hero wins first orders and may run on a thinner return if the data shows what those customers do next. A profit line earns the margin and deserves more reach. Clearance recovers cash before a deadline. New lines need protected exposure. An add-on sells alongside something else. Stock depth decides whether a line can take more demand. The role sets the target, the budget rule and the group's place in each channel, carried by the custom labels the product feeds guide describes.
Task
- Calculate break-even and target return for each group, and judge the reported figures.
- Assign a role to each group, with the evidence for it.
- Map the groups to Performance Max campaigns, asset groups and listing groups, with targets and budget rules.
- Map the groups to Meta product sets for prospecting, retargeting and cross-sell.
- Decide which filter views become indexable pages and how the rest are handled.
- Write the monthly rule for moving products between groups.
Worked answer
Peakform Nutrition: break-even, targets and roles
Break-even is 1 ÷ margin; target is 1 ÷ (margin − 0.08).
| Group | Break-even | Target | Role |
|---|---|---|---|
| Branded whey | 5.6 | 5.6, held at break-even | Hero, new customers only |
| Own-brand protein | 2.1 | 2.5 | Profit |
| Own-brand vitamins | 1.8 | 2.1 | Profit |
| Pre-workout | 3.8 | 5.6 | Profit, thin |
| Short-dated stock | 1.7 | 1.9 | Clearance |
| Plant protein | 2.2 | None until it has history | New |
| Shakers and accessories | 2.9 | Not advertised alone | Add-on |
Branded whey on Performance Max: £9,200 of spend × 5.2 = £47,840 of revenue × 18% ≈ £8,611 of contribution, about £590 short of its ad cost before any brand traffic is removed.
Performance Max. Three campaigns, one per target. Own brand holds protein and vitamins at the protein target, with an asset group per message and one for plant protein, whose handful of orders is too few to learn alone. Third party holds whey and pre-workout at the shared figure (break-even for whey, target for pre-workout), with the customer acquisition goal bidding higher for new customers. Clearance has a fixed budget and an end date. Shakers are excluded from every listing group, and the brand goes to brand search, so no campaign reports brand demand as its own ROASRevenue generated per unit of ad spend..
Meta. Prospecting uses a product set of deep-stock own-brand lines, excluding customers and recent visitors. A RemarketingAdvertising to previous visitors or engagers. ladder keeps the full catalogue minus accessories and low stock, with basket abandoners on a short window and viewers on a longer one. A cross-sell set shows shakers and vitamins to recent protein buyers. Clearance appears only in retargeting and customer sets. Branded whey leaves prospecting: its Meta return sits below break-even even with past visitors flattering it.
Filter pages. Following the faceted navigation guide, chocolate and vanilla whey and caffeine-free pre-workout become indexable landing pages with their own copy, linked from the category, featuring own-brand options where they genuinely fit. The size view shows variants of the same products, so its Canonical TagTells search engines the preferred URL. points to the whey category. Strawberry plant protein and every flavour-plus-size pair stay out of the index and out of crawlable links. Margin never decides indexing; demand, product count and distinctness do.
The moving rule. Read both platforms' product reports monthly. A product with spend and no sales goes to a feed check before any move; a launch joins the profit group after a quarter of sales; a profit line low on stock leaves prospecting until restocked.
Common mistakes
- Calling every best seller a hero. The role needs evidence of what new customers do next.
- Judging clearance on full margin. Written-off stock has no cost of goods to recover.
- Different roles per platform. They pull budget in opposite directions.
- Indexing by margin. Demand decides which filter pages exist.
Next, take the catalogue scale checkpoint or revisit single-channel segmentation at the intermediate e-commerce level.
