Marketing ROI tracking across campaign cohorts

Marketing ROI tracking compares campaign return snapshots using consistent costs, attribution rules, and maturity windows. Keep the old snapshots so you can explain revisions, then compare different campaigns at the same age. A percentage without its cohort, observation window, and reporting cutoff cannot support a dependable trend.

Define what stays consistent across snapshots

Start with a campaign register containing cohort ID, start date, eligible outcomes, attribution rule, cost boundary, and responsible owner. Define observation age as days elapsed since the cohort start. Record the actual as-of date and source-extract dates beside every snapshot.

Keep the outcome basis stable: bookings, invoiced sales, and collected cash are different measures. For the example below, the measure is attributed net sales after known discounts and refunds, less related variable costs. Marketing costs sit outside those variable costs and are deducted once.

Keep the return convention explicit too. Google's ROI reference describes ROI in relation to net profit and costs, with the calculation depending on the advertiser's goals. The marketing ROI versus ROAS guide explains the specific contribution and marketing-investment convention used here.

Choose the ages you will compare before reading the outcome. A campaign at an early age has had less time to generate sales and recognize refunds than an older one. Comparing current calendar-month totals alone can confuse that difference with a change in performance.

Retain each cohort's history

Each source extract should have a snapshot ID. Preserve the rows as they were known at that cutoff, either in separate extracts or in a history table with the snapshot ID attached. A workbook refresh should not quietly replace the evidence behind a prior review.

For a US distributor, the following illustrative extracts show the January and February 2026 acquisition cohorts at two ages. The rows are cumulative views, not additional transactions. Amounts are USD; percentages are rounded to one decimal place.

With this summary in columns A through I, enter contribution in F2:

=D2-E2

Enter contribution after marketing in H2:

=F2-G2

Enter attributed ROI in I2:

=IF(G2=0,"",H2/G2)

Fill down and format column I as a percentage. A missing cost should remain unresolved; a known zero cost makes the return ratio undefined, so the formula returns a blank.

To assemble net sales, suppose Outcomes has snapshot ID in A, cohort ID in B, and the already matched, window-eligible net amount in C. In summary D2, use:

=SUMIFS(Outcomes!$C$2:$C$1000,Outcomes!$A$2:$A$1000,$A2,Outcomes!$B$2:$B$1000,$B2)

Apply the same snapshot and cohort filters to the related variable costs and the approved marketing-cost register. Microsoft's SUMIFS documentation confirms that the function sums values meeting multiple criteria. Extend all ranges to the complete extracts.

Covirage is not affiliated with Microsoft or Google.

Reconcile the movement before calling it improvement

January's contribution after marketing moves from negative $400 at day 30 to positive $3,500 at day 60. Explain that change with a bridge drawn from the same rows:

The check is negative $400 plus $3,900 equals $3,500. This proves the arithmetic; supporting order and invoice rows explain whether the additions represent later activity, corrected records, or late-arriving costs. It does not show that a budget change made the campaign more productive.

Compare January day 60 with February day 60 for the campaign comparison. February's 44.0% exceeds January's 33.3%, but the next review still needs audience, offer, and product-mix context. Improving marketing ROI covers the decision that follows this measurement work.

Produce a portfolio view without repeated outcomes

Select only the day-60 rows. Their net sales total $80,000, variable costs $48,000, contribution $32,000, and marketing costs $23,000. Contribution after marketing is $9,000, so combined attributed ROI is 39.1%.

For the displayed worksheet, row 3 and row 5 are the selected snapshots:

=(H3+H5)/(G3+G5)

Tie the $23,000 cost total to the selected campaign-cost register. Reconcile eligible sales as attributed plus unallocated sales against the finance export. For acquisition cohorts, keep customer acquisition cost alongside the return; a changing customer mix can affect both measures.

Where tracking goes wrong

• Cumulative snapshots are summed. An early snapshot is already contained in its later version. Select one age per cohort for portfolio totals.
• Attribution rules change without a version. Record the rule used for each snapshot and, when restating history, preserve both versions. Google's attribution settings guidance distinguishes credit allocation and lookback settings; neither should be an invisible assumption in your own report.
• Maturity is mistaken for operating improvement. Late sales and costs can move an old cohort's return while current execution remains unchanged.
• One campaign has broader costs. Adding labor to one cohort while excluding it elsewhere creates a false ranking.
• Unmatched rows vanish. Show their dollar value and owner. Missing allocation is a reconciliation issue, rather than evidence of no return.

Track comparable snapshots on your own file in Covirage

Bring your saved campaign snapshots, matched sales, and cost exports to Covirage as files or a Google Sheet. Covirage maps the columns, and its tools compute and check the figures; the external AI model explains the result with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.

See customer profitability for the contribution view behind campaign return.