More Sales, Less Profit: Reading Your Weekly Campaign Numbers

A campaign can report more orders and still leave less money. Refunds, payment fees, fulfilment costs, discounts, and a change in audience quality can grow at the same time as sales. A visitor counter or a revenue total cannot show that trade-off on its own.

A weekly contribution report gives you a clearer decision. It combines spend, confirmed revenue, variable costs, refunds, and qualified outcomes by source and week. You can then decide whether to scale, hold, or pause based on money retained after advertising. The figures below are hypothetical examples, so they show a method rather than a client result.

Start with the money that remains after each sale

Suppose a campaign spends $1,200 in one week and receives $4,000 in reported sales. At first glance, $4,000 may support a larger budget. Now subtract $1,100 for product or service delivery, $280 in payment and fulfilment costs, and $300 in refunds. The campaign contribution before advertising is $2,320, and the contribution after advertising is $1,120.

Contribution means revenue left after costs that change with the sale. In this report, the practical formula is:

Net revenue = confirmed revenue - refunds - credits
Variable costs = product or delivery cost + payment fees + fulfilment + commissions
Contribution after advertising = net revenue - variable costs - campaign spend
Post-ad contribution rate = contribution after advertising / net revenue

This measure is different from revenue and from return on ad spend. ROAS divides reported revenue by ad spend, while contribution after advertising also recognises the costs needed to make and keep the sale. This is not net profit: fixed overhead, tax and other excluded costs still need to be covered. A positive number does not prove that a campaign is scalable, but it gives you a better starting point than a larger sales count.

Set up a weekly contribution report you can act on

A report is useful only when each row has a clear period, source, and financial meaning. If your campaign data is spread across separate exports, choose one currency and one timezone, then record the extraction date. You can use the pricing and campaign planning page when estimating a traffic budget, but keep the traffic cost as its own line in the report rather than hiding it inside revenue.

A practical step is to use one row for each source and week. Your minimum fields are week start, source, campaign or placement, spend, reported orders, confirmed orders, gross revenue, refunds, credits, variable delivery cost, payment and fulfilment costs, qualified outcomes, and decision. A qualified outcome is the event that your business values, such as a paid order that passed a review or a lead accepted by sales. You may want to define it before reviewing performance so the definition does not change to rescue a weak result.

The downloadable weekly contribution report CSV worksheet provides a copyable starting structure. You can replace its hypothetical values with your own data, add rows for each source, and keep the same definitions from week to week.

Example calculation

Assume the following hypothetical row for one source:

The arithmetic is $4,000 - $300 = $3,700 net revenue. Then $3,700 - $1,100 - $280 - $1,200 = $1,120 contribution after advertising. The post-ad contribution rate is $1,120 / $3,700 = 30.27%, rounded to 30.3%.

Keep recent weeks open until conversions and refunds settle

A week ending yesterday may look weak because some customers have not completed checkout, a lead has not been accepted, or a refund has not been recorded. Treating incomplete data as final can cause a pause that removes a useful source, while treating early orders as final can cause an expensive increase.

It helps to add a data-status field such as open, provisional, or closed. A practical step is to record the normal conversion lag and refund review period for your business. For example, if a typical paid order is confirmed within five days and refunds are checked after fourteen days, a report for the last seven days may remain provisional. That does not mean you ignore it; it means you separate an early signal from a settled decision.

A practical review uses two dates: the event week and the confirmation or refund date. It is better not to silently move a late conversion into a new week just to make the newest week look better. You may want to use a consistent rule, such as assigning an order to the week of the qualifying campaign interaction, and document exceptions where your analytics setup cannot support that rule.

Compare sources by contribution and qualified outcomes

A source with the lowest cost per order may still bring low-value or refunded orders. Another source may bring fewer orders but more confirmed revenue and accepted leads. Put spend, contribution, and quality beside each other instead of ranking sources on one metric.

Useful weekly checks include:

• Contribution after advertising: the amount available after variable costs and spend.
• Contribution per qualified outcome: contribution after advertising divided by qualified outcomes. Leave it blank when the outcome count is zero.
• Refund rate: refunds divided by reported revenue or orders, using one documented basis.
• Confirmation rate: confirmed orders divided by reported orders.
• Spend share: source spend divided by total campaign spend.

You can compare the current week with a settled prior period, not just the previous seven calendar days. A short table can expose a quality shift:

The table is hypothetical. Its purpose is to show why Source A and Source B should not be judged by order count alone, and why Source C needs a clear correction or pause rule.

Record changes before you explain a number

Weekly totals can move because targeting, placement mix, offer, landing page, pricing, tracking, or fulfilment changed. Without a change log, you may attribute a revenue drop to traffic when the real cause was a stock problem or a new refund policy.

It helps to add a short note beside every material movement. A practical step is to record what changed, when it changed, which sources were affected, and what result you expected. Then test that explanation against the data. If a source gained 30% more visits after a targeting change but its confirmation rate fell from 62% to 41%, the extra traffic is not automatically progress.

If a placement looks weak, use the seven checks before pausing an ad placement before removing it. If order totals appear unusually high, review duplicate purchase events and their effect on reporting. Both checks protect the contribution calculation from decisions based on broken or incomplete measurement.

Choose scale, hold, or pause with a decision log

A budget decision should state its evidence and its next review date. It should not rely on a universal increase such as “add 20% every week,” because the affordable increase depends on cash limits, conversion lag, delivery capacity, and how much uncertainty remains.

You may want to use this decision sequence:

1. Check data status. Keep the row provisional if conversions or refunds are still within the normal settling period.
2. Check contribution. A negative settled contribution needs a pause, correction, or a smaller controlled test unless a documented strategic reason applies.
3. Check quality. Compare confirmation, refund, and qualified-outcome rates with the settled baseline.
4. Check capacity. Confirm that fulfilment, sales follow-up, and cash flow can support additional demand.
5. Set a bounded test. Choose an amount you can afford to lose if the next period repeats the current contribution rate.
6. Set the rule. Write the scale, hold, or pause threshold before the test starts.

For a hypothetical example, suppose a source has $1,120 settled contribution after $1,200 spend. You may test an additional $200 rather than applying a percentage increase. A scale rule could require settled contribution to remain above $0, confirmation rate to stay at or above 55%, and no material capacity issue.

A pause rule could apply after two settled weeks below $0 or after a confirmed tracking error. These thresholds are examples, not universal benchmarks.

Decision-log template

Review week:
Source and campaign:
Data status: open / provisional / closed
Spend:
Net revenue:
Variable costs:
Contribution after advertising:
Qualified outcomes:
Change made this period:
Decision: scale / hold / pause
Next test budget:
Scale rule:
Pause rule:
Next review date:
Owner and evidence location:

The next test should answer one question, such as whether a new audience improves qualified-outcome rate or whether a placement change reduces refunds. Changing several inputs at once makes the result harder to interpret.

Use experiments when attribution overstates the contribution

A conversion credited to a campaign is not always an additional conversion caused by it. Some people would have purchased without the campaign, and multiple sources may claim the same order. Attribution is useful for operational reporting, but budget allocation benefits from a causal test when the spend is large enough to justify one.

An incrementality test compares a treatment group exposed to the campaign with a comparable control group that is not exposed. The difference in outcomes estimates additional revenue or contribution, subject to the test design and its limits. A geographic or audience split may be appropriate, but privacy, sample size, interference between groups, and the length of the test need review before launch.

You can add an incrementality_status field to the worksheet. Use values such as not tested, planned, and tested with notes. It is better not to present an attributed contribution as causal proof. If a test is not practical, label the report as attribution-based and use conservative scale limits.

Close the week with one accountable decision

A good report ends with a decision, not a dashboard screenshot. It says what happened, whether the numbers are settled, what changed, what money remained after advertising, and what the next bounded test will prove.

You can review the CSV with finance, fulfilment, and campaign owners when their inputs affect refunds, costs, or qualified outcomes. Keep definitions stable, preserve the prior version, and explain any correction. More sales can be useful, but the next budget should follow retained contribution and reliable evidence rather than a larger visitor counter.