PMax: Knobs — Campaign Analysis
Summary
A short-lived PMax campaign targeting furniture knobs that ran for 12 days (Dec 12–23) before being paused. Loss-making at 0.37x ROAS with 31% of budget wasted on Display/YouTube channels that produced zero conversions. Even Search-only performance was 0.54x ROAS. However, the failure was the PMax format, not the product economics — see ./knobs-and-handles-analysis.md for the full cross-campaign picture showing actual order AOV of £46–86 and evidence of PMax cannibalising Shopping conversions.
Campaign Configuration
| Setting | Value |
|---|---|
| Campaign ID | 75 |
| Name | PMax: Knobs Only Shopping (SC) |
| Type | Performance Max |
| Bidding strategy | (not recorded — likely MAXIMIZE_CONVERSIONS) |
| Daily budget | £60 |
| Status | PAUSED (last data Dec 23) |
Channel Breakdown (confirmed via segments.ad_network_type)
| Channel | % Spend | Impressions | Clicks | Cost | CPC | Conv | Revenue | ROAS |
|---|---|---|---|---|---|---|---|---|
| SEARCH | 68.9% | 24K | 374 | £514 | £1.37 | 8.0 | £278 | 0.54x |
| CONTENT (Display) | 25.5% | 257K | 3,046 | £190 | £0.06 | 0 | £0 | 0.00x |
| YOUTUBE | 5.4% | 229 | 13 | £41 | £3.12 | 0 | £0 | 0.00x |
| SEARCH_PARTNERS | 0.1% | 30 | 1 | £1 | £0.54 | 0 | £0 | 0.00x |
| TOTAL | 100% | 281K | 3,434 | £745 | £0.22 | 8.0 | £278 | 0.37x |
The worst of both worlds
- Display: 257K impressions and 3,046 clicks for zero conversions. £190 wasted on £0.06 CPC remnant inventory. The highest proportion of Display waste across all three PMax campaigns (25.5% vs Knife Rack’s 12.2% vs Table Tops’ 0.9%).
- YouTube: £3.12 CPC (the most expensive channel) with zero conversions from 13 clicks. £41 wasted.
- Search: Even without Display/YouTube waste, ROAS is 0.54x — losing £0.46 for every £1 spent. The product doesn’t convert profitably through paid search at any efficiency level.
Device Breakdown
Not available — campaign ran Dec 12–23 only, outside the current 30-day data window. Given the 25.5% Display waste pattern, the device split would likely mirror Knife Rack: cheap mobile/tablet clicks from Display remnant inventory, with only desktop showing real Shopping CPCs.
Analysis
Why this failed
PMax was the wrong format. With only 8 conversions over 12 days, PMax never had enough data to optimise. It dumped 31% of budget into Display/YouTube (the worst allocation of any PMax campaign) because it couldn’t find enough profitable Search auctions in the limited run.
The £35 AOV figure is misleading. The 8 PMax conversions averaged £35, but actual order data (3,033 historical orders containing knobs) shows a median AOV of £46 and mean of £86. Customers typically buy 8 units per order. The PMax sample was too small to reflect real buying behaviour.
PMax actively suppressed Shopping. PMax takes auction priority over Standard Shopping (Google’s documented behaviour). During the 12 days PMax ran, Shopping Catch All could only show knob products at 1–5 impressions/day with zero clicks — PMax won every auction. After PMax was paused, Shopping took 5 weeks to ramp up before its first knob conversion (Jan 26). PMax didn’t just cannibalise Shopping’s conversions — it prevented Shopping from building any performance history on these products, creating a weeks-long learning gap after shutdown. Monthly order volumes (~55/month) stayed flat regardless of PMax status, confirming PMax wasn’t generating incremental demand.
PMax channel allocation was the worst observed. 31% of budget went to non-Search channels (vs 15% for Knife Rack, 2% for Table Tops). The lower the product’s natural search demand, the more PMax dumps budget into Display fill. Knobs had the least natural search demand of the three, so PMax allocated the most to Display.
Verdict
Correctly paused. Do not restart PMax. The PMax format is wrong for this product — limited search demand means PMax dumps budget into Display, and it likely cannibalises Shopping conversions rather than generating incremental sales.
However, the product is not unviable for paid search. Shopping Catch All is already generating knob conversions at 2.11x ROAS (early data, 4 conversions). The correct channel is Shopping, not PMax. Recommendations:
- Keep knobs in Shopping Catch All (where they’re already performing)
- Consider incentivising multi-unit purchases on-site to push AOV higher
- Do not create another dedicated knob campaign — let Shopping’s algorithm allocate budget naturally
- See
./knobs-and-handles-analysis.mdfor the full cross-campaign analysis and order data
Comparison Across All PMax Campaigns
| Metric | Table Tops | Knife Rack | Knobs |
|---|---|---|---|
| PMax AOV (sample) | £150 | £53 | £35 (8 conversions) |
| Actual order AOV | £150 | — | £46–86 (3,033 orders) |
| Search ROAS | 4.20x | 2.38x | 0.54x |
| Display % of spend | 0.9% | 12.2% | 25.5% |
| Display conversions | 0 | 0 | 0 |
| Search cost/conv | £36 | £22 | £64 |
| Verdict | Keep | Paused (marginal) | Paused (wrong format) |
Pattern: PMax’s algorithm allocates more to Display when it can’t find profitable Search auctions. Knobs had the most Display waste because search demand for the category is limited — not because the AOV is low. The actual AOV (£46–86 from order data) is comparable to Knife Rack. The 0.54x Search ROAS reflects PMax’s 12-day sample and likely cannibalization of Shopping, not the product’s true paid search viability.
Related Documents
./pmax-table-tops-analysis.md— Table Tops PMax (enabled, healthy)./pmax-knife-rack-analysis.md— Knife Rack PMax (paused, Display waste)../../../market-intelligence.md— PMax channel reporting, placement exclusion options