Blog · AI and self-service analytics
Compare campaign spend, recorded conversions and customer revenue from your own exports. Follow a repeatable review with spreadsheet formulas, reconciliation checks and clear limits on attribution.
AI for marketing teams is useful when it turns campaign exports into a comparison the marketing manager can check and act on. Start with spend, recorded conversions and customer revenue on the same agreed basis, then ask the external AI model to explain the computed differences. The first job is to make those numbers comparable.
Write down the decision: which campaigns need investigation before the next spending review? Then specify the conversion event, campaign cohort, revenue cutoff, currency and attribution rule. The broader AI analytics overview explains the division between business calculations and explanations; this workflow applies that division to a recurring marketing review.
For this example, a US business-services company reviews campaigns acquired in March 2026. Conversions are recorded inquiry submissions during March. Revenue is net recognized customer revenue through April 30, 2026, assigned to those campaigns under the company's agreed rule. Everything is in USD.
Attribution assigns credit across marketing interactions, as Google's attribution documentation explains. Save the rule with the exports. Changing it between reviews can move campaign revenue even when customer activity has not changed. Use the existing marketing attribution methods guide to choose the rule rather than designing a new one during the review.
Keep three source sheets with these columns:
| Sheet | A | B | C | D |
|---|---|---|---|---|
| Costs | Campaign ID | Cost date | Spend (USD) | Source reference |
| Conversions | Conversion ID | Campaign ID | Conversion date | Customer ID |
| Revenue | Invoice or adjustment ID | Campaign ID | Recognition date | Net revenue (USD) |
The Revenue sheet is restricted to customers assigned to the March campaign cohort. Include credits as negative amounts and retain an unassigned group where the campaign cannot be established. Each conversion ID must appear once in the agreed event population. Each revenue line must carry only the credit allocated to that campaign.
Aggregate each sheet by campaign before comparing them. Joining every cost row directly to every conversion or invoice can repeat the same cost. Preserve source totals and export timestamps alongside the working comparison, so a reviewer can distinguish a corrected source from a changed formula.
Put campaign names in A2:A4 on a Summary sheet. Columns B, C and D contain spend, conversions and attributed revenue. These Excel formulas assume source rows run through row 1000; extend the ranges to cover the complete export.
In B2, collect March spend:
=SUMIFS(Costs!$C$2:$C$1000,Costs!$A$2:$A$1000,A2,Costs!$B$2:$B$1000,">="&DATE(2026,3,1),Costs!$B$2:$B$1000,"<"&DATE(2026,4,1))
In C2, count March conversion records:
=COUNTIFS(Conversions!$B$2:$B$1000,A2,Conversions!$C$2:$C$1000,">="&DATE(2026,3,1),Conversions!$C$2:$C$1000,"<"&DATE(2026,4,1))
In D2, sum cohort revenue recognized before May:
=SUMIFS(Revenue!$D$2:$D$1000,Revenue!$B$2:$B$1000,A2,Revenue!$C$2:$C$1000,"<"&DATE(2026,5,1))
Microsoft documents SUMIFS for adding values that meet multiple criteria and COUNTIFS for counting records that meet them. The date boundaries include timestamps throughout the last reporting day. Covirage is not affiliated with Microsoft or Google.
In E2 and F2, calculate the two comparisons, then fill all formulas down:
=IF(C2=0,"No conversions",B2/C2)
=IF(B2=0,"No spend",D2/B2)
| Campaign | Spend (USD) | Conversions | Attributed revenue (USD) | Cost per conversion (USD) | Revenue / spend |
|---|---|---|---|---|---|
| Search | 18,000 | 120 | 54,000 | 150.00 | 3.00 |
| Webinar | 12,000 | 80 | 40,000 | 150.00 | 3.33 |
| Partner | 6,000 | 30 | 30,000 | 200.00 | 5.00 |
| Total | 36,000 | 230 | 124,000 | 156.52 | 3.44 |
Dollar ratios are rounded to cents and revenue multiples to two decimals. The total cost per conversion is 36,000 / 230 = $156.52. Total revenue per dollar of spend is 124,000 / 36,000 = 3.44. Calculate both from totals, preserving unrounded values underneath.
The campaign spend must equal the cost export: 18,000 + 12,000 + 6,000 = $36,000. Conversion counts must also reconcile: 120 + 80 + 30 = 230.
Revenue needs a separate bridge because not every customer has a reliable campaign assignment:
| Revenue reconciliation through April 30 | Amount (USD) |
|---|---|
| Search | 54,000 |
| Webinar | 40,000 |
| Partner | 30,000 |
| Unassigned cohort customer revenue | 16,000 |
| Cohort revenue control total | 140,000 |
The check is 124,000 + 16,000 = $140,000. Keep the unassigned amount visible; spreading it across campaigns would create evidence the source does not contain. Finance should confirm the cohort revenue control against the underlying revenue lines, while marketing operations owns the campaign mapping.
Search and Webinar both cost $150.00 per conversion. Partner costs $200.00 but shows the highest revenue per dollar of spend. That makes customer quality, conversion definitions and revenue timing useful follow-up questions before reducing Partner's budget.
These figures do not establish profitability. They omit the cost of fulfilling the customer sale and other marketing costs. Apply the separate marketing ROI and ROAS method before treating a revenue multiple as a return on profit. If a campaign produces inquiries that sales has not yet closed, the sales-team analytics overview helps frame the handoff without counting pipeline as revenue.
Record each exception with the campaign, disputed source rows, responsible team and next review date. Marketing operations resolves identifiers; finance resolves revenue classification; the marketing manager decides spending after those checks.
Bring a sample of your own campaign cost, conversion and customer revenue exports to Covirage, with the agreed reporting dates and attribution rule. Covirage maps the columns; its tools compute and check the comparisons, and the external AI model explains the results with every figure traced to its rows. Create a free Covirage account, then set up your business and bring your files.
See how this fits self-service analytics for business teams.
Bring campaign costs, recorded conversion events and customer revenue, with campaign identifiers and dates. Keep event identifiers to check duplicates and customer or invoice identifiers to trace revenue. Record the attribution rule, reporting currency and cutoff so the comparison can be repeated.
A comparison of exports shows recorded associations under an attribution rule. It does not by itself establish that advertising caused a sale. Keep causal claims separate from campaign reporting and use a suitable experiment when the decision depends on incremental impact.
Review both because they answer different questions. Cost per conversion measures the cost of a defined event. Revenue per dollar of spend connects that activity with recorded customer outcomes. A campaign can have expensive conversions and still produce stronger revenue per dollar.
Use a cadence that matches spending decisions and the sales cycle. Keep the reporting cutoff and outcome maturity consistent between reviews. A recent campaign with incomplete customer revenue should be labeled provisional rather than ranked directly against a fully matured campaign.