Why this matters
An automated campaign that may serve on brand searches will usually find them, because they convert best. Its return then rises while the brand campaign shrinks, and the account looks as if it has grown when it has mostly relabelled demand it already had. In paid search (SEMPaid search advertising (sometimes used to include SEO).), brand capture and prospecting are two jobs with two measures. The comparison of Standard Shopping and Performance Max flags brand handling as a core trade-off. Here you measure the effect with the data an account actually has, which never includes a clean split of cost by query.
The candle maker and its campaigns
Petalforge Candles is an invented maker of scented candles with a loyal following under its own name. Last year it ran a brand search campaign and Standard Shopping. This year Standard Shopping was replaced by Performance Max with full asset groups and no brand exclusion, and the brand campaign kept its phrase and broad match brand keywords.
Petalforge Candles: same quarter, last year and this year
| Campaign | Year | Spend | Conversion value | ROAS | Brand search impression share |
|---|---|---|---|---|---|
| Brand Search | Last year | £4,000 | £60,000 | 1,500% | 94% |
| Standard Shopping | Last year | £22,000 | £88,000 | 400% | n/a |
| Brand Search | This year | £1,500 | £21,000 | 1,400% | 41% |
| Performance Max | This year | £30,000 | £150,000 | 500% | n/a |
- Search term insights, Performance Max, share of conversion value: brand name and variants 27%, scented candles 21%, candle gift sets 17%, soy candles 12%, other categories 23%.
- Brand search volume index, from the brand campaign's impressions divided by its impression share: last year 100, this year 98.
- Order system, Performance Max orders this quarter: 3,750, of which 1,350 from new customers at an average first order of £38. New customers acquired last year spent on average a further £45 within twelve months.
- Contribution margin after product and fulfilment: 45%. The owner wants 15 points of revenue kept as profit after ad cost. A small Microsoft Advertising brand campaign also runs.
Telling brand demand from new demand
No report gives Performance Max cost by query, so the estimate is a triangulation. Search term insights show which categories carried conversion value. The brand campaign's Impression ShareShare of eligible impressions you received. shows how much of the brand query volume it lost, and the brand volume index shows whether brand demand itself changed. If demand held steady while the brand campaign lost share, the missing brand value went somewhere, and insights usually say where. New-customer counts from the order system then show how much of the non-brand remainder is genuinely new.
Judge the result against margin, not last year's return. Brand clicks are cheap, so the estimate of brand cost inside Performance Max borrows the brand campaign's own efficiency.
Task
- Estimate the brand conversion value that Performance Max collected, two ways, and check they agree.
- Estimate its spend on brand and recalculate its non-brand ROASRevenue generated per unit of ad spend..
- Calculate break-even and target return from the margin, and judge the non-brand result.
- Judge prospecting on new customers, using first orders and first-year value.
- Propose the structure: brand campaign, Performance Max settings and anything you would leave alone.
- Design the monthly reporting view and the weekly checks that keep the split honest.
Worked answer
Petalforge's brand estimate and recalculated returns
Way one: insights put 27% of £150,000 on brand terms ≈ £40,500. Way two: last year's brand value £60,000 × index 0.98 ≈ £58,800 of brand demand; the brand campaign caught £21,000, leaving ≈ £37,800 unaccounted for. The two estimates agree closely: call it about £40,000.
Brand cost inside Performance Max, at the brand campaign's 1,400%: £40,500 ÷ 14 ≈ £2,900. Non-brand: £150,000 − £40,500 = £109,500 of value on £30,000 − £2,900 = £27,100 of spend ≈ 404%.
Break-even: 1 ÷ 0.45 ≈ 222%. Target: 1 ÷ (0.45 − 0.15) ≈ 333%.
New customers: 1,350 × £38 = £51,300 of first orders ≈ 189% on the non-brand spend, below break-even. First-year value: (£38 + £45) × 1,350 ≈ £112,000; at 45% margin ≈ £50,400 against £27,100 of spend.
Account view: last year £148,000 of value on £26,000; this year £171,000 on £31,500.
What the numbers say. The true non-brand return is about a fifth below the reported one, yet it still clears the target. Prospecting loses money on first orders and pays back within the year, so it is an acquisition investment and should be reported as one. The account's growth is smaller than Performance Max suggests: value rose more slowly than spend.
Structure. Exclude the brand from Performance Max with a brand list. Move the brand campaign to exact and phrase brand keywords with its own budget and a target impression share capped by a maximum cost per click, so it reclaims the queries and reports them. Set the Performance Max customer acquisition goal to bid higher for new customers, as the Performance Max structure guide describes, which needs a maintained customer list. Leave Standard Shopping retired for now; return to it only if specific products need query-level control. Keep the Microsoft brand campaign apart from any non-brand bidding, the same rule the guide to cross-engine bid strategies applies to portfolios.
Reporting view. Three lines every month: brand (cost, value, impression share), non-brand returning customers, and non-brand new customers with cost per new customer and first-year value. Show the account total against the same period last year beside them, because campaign figures will move when brand traffic shifts.
Weekly checks. Following the paid search account management playbook, review search term insights weekly for brand variants leaking back, log the exclusion and the bidding change as separate dated entries, and judge them at the monthly review once both have settled.
Common mistakes
- Reading Performance Max return as prospecting return. Brand queries lift it for free.
- Blaming the brand campaign for shrinking. It lost queries, not effectiveness.
- Judging prospecting on first-order return alone. New customers pay back later; show both views.
- Comparing campaigns instead of the account. Value moves between campaigns; the total is what grew or did not.
Check how this fits the whole level in the Google Ads growth strategy capstone, or revisit Performance Max basics from the intermediate Google Ads level.
