Why this matters
Most bidding problems start with a target nobody can explain, copied from last quarter's average or accepted from the platform's suggestion. The Smart Bidding strategies guide explains what each strategy optimises and how to change targets safely; the performance economics guide explains why the target has to come from margin. Here you put the two together on one account, where the economic target and current performance rarely agree and the job is to decide how to get from one to the other.
The margins and conversion data
Oakhaven Furniture is an invented online retailer of sofas, beds and home accessories. Its finance lead has supplied each line's contribution margin after product cost, delivery, payment fees and returns.
Oakhaven Furniture: margins by product line
| Product line | Average order value | Contribution margin |
|---|---|---|
| Sofas | £860 | 40% |
| Beds and mattresses | £370 | 30% |
| Home accessories | £65 | 20% |
The owner wants each line to keep at least 10 percentage points of revenue as profit after ad cost. On a £860 sofa order, that means £86 of profit once the advertising that won the order has been paid for.
The account's four campaigns, over a closed period, with purchase values verified against the order system:
Oakhaven Furniture: campaign results over eight weeks
| Campaign | Current strategy | Spend | Orders | Conversion value | ROAS | Search lost impression share (budget) |
|---|---|---|---|---|---|---|
| Brand Search | Manual CPC | £1,600 | 96 | £52,800 | 3,300% | 0% |
| Sofas Search | Maximise clicks | £14,000 | 70 | £60,200 | 430% | 35% |
| Beds Search | Manual CPC | £7,400 | 74 | £27,380 | 370% | 5% |
| Accessories Shopping | Manual CPC | £3,200 | 128 | £8,320 | 260% | 0% |
From margin to target
Two numbers come from each margin. The break-even return is one divided by the contribution margin; below it, every reported sale loses money. The target return is one divided by the margin left after the owner's profit is set aside. The break-even calculator does the first step, and the ROAS calculator shows what each campaign is actually returning.
The target is a destination, not necessarily a starting point. A campaign beating its target can be given a looser one to find more volume; a campaign below it needs walking towards it, because a ROASRevenue generated per unit of ad spend. target far above what the campaign achieves makes the model bid on few auctions and spend collapses. As the budgets and pacing guide explains, a budget raise only matters to a campaign that is limited by budget.
Task
Answer before reading on.
- Calculate the break-even ROAS and the target ROAS for each product line.
- Calculate the contribution profit after ad cost for each campaign over the period, using the margin of the product line it sells. Treat Brand Search as a blend and judge it on purpose rather than margin.
- Choose a bid strategy for each campaign and say in one sentence why that strategy fits its goal and data.
- Set a starting target for each campaign that uses one, and say how and when you would move it.
- Decide whether any campaign should have its budget raised, held or removed, and what you would check first.
Worked answer
The arithmetic first, then the decisions.
Oakhaven's targets and campaign profit
| Product line | Break-even ROAS | Target ROAS |
|---|---|---|
| Sofas (40% margin) | 1 ÷ 0.40 = 250% | 1 ÷ (0.40 − 0.10) = 333% |
| Beds (30% margin) | 1 ÷ 0.30 = 333% | 1 ÷ (0.30 − 0.10) = 500% |
| Accessories (20% margin) | 1 ÷ 0.20 = 500% | 1 ÷ (0.20 − 0.10) = 1,000% |
Profit after ad cost over the eight weeks:
- Sofas Search: £60,200 × 0.40 = £24,080 margin, less £14,000 spend = £10,080 profit.
- Beds Search: £27,380 × 0.30 = £8,214 margin, less £7,400 spend = £814 profit.
- Accessories Shopping: £8,320 × 0.20 = £1,664 margin, less £3,200 spend = a loss of £1,536.
Decisions:
- Brand Search: target impression share, absolute top, with a maximum CPC cap and its own budget.
- Sofas Search: maximise conversion value with a target ROAS of 400%, then a stepped budget increase.
- Beds Search: maximise conversion value with a target ROAS of 400%, raised to 450% and then 500% as the campaign settles.
- Accessories Shopping: paused as a standalone campaign.
Brand Search. The goal is presence on the retailer's own name; its return is high because those people were looking for Oakhaven already. Target impression share with a cost cap keeps the ad on top without letting a competitor drive the price up unchecked, and its own budget stops generic campaigns starving it.
Sofas Search. It beats its target, but on maximise clicks it ignores which clicks convert, and it has enough orders to move to maximise conversion value. A target slightly below its trailing return and well above the economic one lets the model buy more volume while protecting profit. It is the one campaign losing impressions to budget, so it deserves more money, but only after the strategy change has settled, in steps, checking that the marginal return rather than the average stays above target.
Beds Search. Above break-even but well below target, it makes almost nothing. The full economic target at once would choke spend, so start near the trailing return and raise it in steps. Expect lower revenue and higher profit.
Accessories Shopping. It returns about half its break-even, so every first order loses money. An economic target would be so far from reality that the model would barely bid: a pause with a wasted learning period. Pause it, and revisit only if order data shows accessories reliably joining sofa or bed baskets.
Common mistakes
- Setting targets from the trailing average. Beds would keep its current return and its near-zero profit.
- Jumping straight to the economic target. Spend collapses on Beds while the reported return looks fine on the little it buys.
- Treating break-even as the target. It is the floor, with no room for attribution error or overheads.
- Changing strategy and budget together on Sofas. Two learning triggers, one unreadable result.
- One target for the whole account. It subsidises accessories with sofa margin.
- Judging Brand Search on return. It measures existing demand.
Next, test yourself with the bidding and budgets checkpoint or return to the intermediate Google Ads level. For a lead generation account judged on cost per acquisition (CPACost of each conversion.), the same logic holds: the maximum comes from a customer's margin, and the working target sits below it.
