Billboard advertising ROI: measure campaign outcomes

ROI on billboard advertising is the contribution associated with the campaign after campaign costs, divided by those campaign costs. Start with matched orders and full costs, rather than traffic counts or impressions. Unless the evidence supports a causal estimate, call the answer an attribution-based return.

Set the measurement boundary before buying space

Write down the campaign dates, locations, response window and outcome definition. A roadside board, transit display and indoor screen can use the same financial method, but their response routes and audience evidence differ. Keep each format identifiable in the campaign register.

Audience delivery is useful for checking the placement you purchased. Geopath's glossary distinguishes circulation, impressions, reach and frequency; these are audience measures, not a sales ledger. Ask the supplier which definition, geography and period its report uses.

For this calculation, contribution means net revenue less product and variable fulfillment costs, before advertising. Campaign investment includes placement, creative production, installation and the other agreed campaign costs. The marketing ROI versus ROAS guide owns the underlying denominator distinction.

Attribution-based billboard ROI = (matched contribution before advertising − full campaign cost) / full campaign cost

Sales and finance should approve the scope together. Decide whether repeat customers count, when refunds become final and how shared creative is allocated. Retain the same policy when comparing campaigns.

Build a response-to-order file

Export the placement register, response log, order register, credit memos and campaign invoices. Include campaign ID, format, location, response date, order ID, order date, net revenue, variable costs and campaign cost category. Record the approved observation cutoff alongside the file.

Use a memorable URL or phone number for a roadside board. A QR code can suit a pedestrian or indoor placement; it should not depend on a driver using a phone. Ask sales staff to record the campaign identifier consistently when customers call.

Deduplicate repeated calls and form submissions before matching orders. If an order has several advertising touches, apply one agreed allocation rule; the marketing attribution guide explains that separate job. Leave unmatched responses visible instead of assigning an estimated sale to each one.

Worked example across out-of-home formats

A fictional US commercial flooring company reviews one completed campaign. Response capture runs from September 1 through September 30, 2026, and matched orders are observed through October 15, 2026. Each order is assigned to one response route, so the rows do not claim the same revenue twice.

Put these columns in A:F, with the format in A and the three format rows in rows 2:4. In Excel or Google Sheets, calculate contribution in G, campaign cost in H, contribution after advertising in I and ROI in J:

G2 = B2-C2
H2 = SUM(D2:F2)
I2 = G2-H2
J2 = IF(H2=0,NA(),I2/H2)

Fill down, then format J as a percentage. Dollar amounts are exact; displayed percentages are rounded to one decimal place.

The total is =SUM(I2:I4)/SUM(H2:H4), not the average of the row percentages. Reconcile $120,000 to the matched order and credit memo register, and $29,000 to campaign invoices. The identity $48,000 − $29,000 = $19,000 proves the numerator.

Add geography without overstating the evidence

A geographic comparison can reveal a question that tracked responses miss. Keep it separate from the response-based return because all-area revenue includes orders already counted above and orders with no tracked response.

The difference in changes is $12,000 − $6,000 = $6,000. It is an observational revenue contrast, not automatically incremental billboard revenue. Weather, promotions, territory changes and different underlying trends can affect the comparison. Do not add it to the tracked revenue table.

Random assignment and suitable design strengthen causal evidence, although research on randomized geographic advertising experiments also identifies challenges from heterogeneous regions and interference. See association and cause before labeling an area difference as lift. Covirage is not affiliated with Geopath, Google or Microsoft.

Where it goes wrong

• Impressions become orders. Audience delivery cannot replace matched sales evidence.
• Installation disappears. Excluding production, permits or installation can make an expensive placement look attractive.
• Commuters become local residents. The people passing a board may buy outside the store's assigned territory.
• Response windows differ. A mature campaign and a recently installed board have had different time to generate orders.
• The best average gets the whole budget. A past attributed return does not measure the return on additional locations.

Use advertising ROI by medium when comparing this result with print or digital campaigns on a consistent cost and observation basis.

Check your own campaign file in Covirage

Bring your placement invoices, response-to-order records and matched revenue and cost file to Covirage; it maps the columns, and its tools compute and check the supplied campaign and geographic comparisons. The external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.

See Covirage for sales teams for the business context behind those comparisons.