Measuring Offline Campaigns Without Guesswork

A flyer campaign produces twelve bookings. A newspaper advert brings three callers who mention it. Your shop is busier after sponsoring a local event. These are useful observations, but they answer different questions. Bookings show a response; mentions suggest a connection; a busy shop might reflect the weather, payday or a competitor closing for refurbishment.

Measuring offline marketing means building a credible chain from exposure to action to commercial value. You will rarely identify every customer journey. You can, however, collect enough evidence to decide whether to repeat a campaign, change the offer or stop spending. The important work happens before anything is printed.

Start with one outcome and a decision

Choose the result that makes the campaign commercially worthwhile. For a plumber, that might be completed first-time customer jobs. For a café, it could be first visits that produce a paid transaction. For a workshop provider, it is usually paid places rather than enquiries.

Write down the decision the measurement must support: “We will repeat this leaflet drop if it acquires new customers below our £35 cost ceiling.” Set that ceiling from your margins and cash constraints, not an industry benchmark.

Build the minimum measurement brief

• Audience: 2,000 households within your delivery radius.
• Offer: a £10 discount on a first booking worth at least £80.
• Primary outcome: completed, paid first bookings.
• Response window: bookings made within 30 days of distribution.
• Completion window: jobs completed within 60 days.
• Decision rule: repeat only if acquisition cost and contribution meet your agreed targets.

Keep supporting measures, such as scans and enquiries, but do not let them replace the primary outcome. Two hundred scans with no profitable bookings are a diagnostic clue, not a commercial success.

Give each placement a visible response route

A campaign needs a response mechanism customers can use and staff can record. Match that mechanism to the action you want: a booking page for appointments, a phone number for urgent services or a till code for walk-in purchases.

Use separate identifiers for separate placements. A leaflet delivered in North Ward might carry NORTH10; an event handout might carry FAIR10. If both use the same code, you can measure the combined response but cannot compare the placements.

Test the whole journey before launch

Scan the printed proof on two phones. Type the short address manually. Call the tracking number. Submit a test booking and check that its source reaches your booking system or spreadsheet. A working landing page is not enough if the campaign label disappears at checkout.

For web analytics, use a consistent naming scheme such as source=community_newsletter, medium=print and campaign=autumn_service. Keep personal information out of tracking links, follow applicable consent requirements and collect only what you need.

Connect responses to paid outcomes

A simple spreadsheet is often sufficient. Record an enquiry or order ID, date, placement, response method, new-or-existing customer status, sale value, direct costs and final status. Record cancellations and refunds too. Otherwise, campaigns that generate weak bookings can look better than they are.

Ask customers, “Where did you first hear about us?” Then, where useful, ask, “What prompted you to contact us today?” The first question captures discovery; the second captures the trigger. A customer may have known your shop for years but visited because of yesterday’s leaflet.

Allow “don’t know” and multiple influences. Do not force staff to select a campaign when the customer cannot remember. Count each order once in your sales total, even if it has a QR scan, a phone call and an offer code attached.

Attribution tells you which sales carried a campaign signal. Incrementality asks how many sales would not have happened without it.

Calculate the economics, not just the response rate

Consider an illustrative leaflet campaign: 2,000 delivered, £600 spent on design, printing and distribution, and 24 completed first-time customer jobs carrying its code. Those jobs generated £2,880 in collected revenue after discounts.

• Tracked customer response rate: 24 ÷ 2,000 = 1.2%.
• Attributed acquisition cost: £600 ÷ 24 = £25 per new customer.
• Attributed revenue return: £2,880 ÷ £600 = 4.8 times campaign spend.

That revenue return does not mean £4.80 of profit for every £1 spent. If materials, payment fees and delivery labour total £1,440, the jobs contribute £1,440 before campaign costs. Subtract the £600 campaign cost and £840 remains towards overheads and profit.

Include campaign staffing, agency fees and tracking charges where relevant. Treat your own time consistently, even if you show it separately from cash spending. If sales revenue is already recorded after discounts, do not subtract the discount again.

Future repeat purchases may improve the result, but do not use hoped-for lifetime value to rescue an unprofitable test. Track repeat contribution from this customer cohort over a defined period, such as 90 days.

Test whether the campaign created additional demand

Some coded customers would have bought anyway. Others may respond without using the code. To estimate additional demand, compare what happened with a reasonable estimate of what would have happened without the campaign.

Use a comparable area where practical

Choose two areas with similar household counts, customer profiles and previous sales. Distribute in one and hold the other back temporarily. Keep pricing and other promotions consistent, and measure all relevant sales by area, not only coded orders.

Suppose the campaign area rises from 40 to 58 orders while the comparison area rises from 30 to 36. Using an absolute-change comparison, the estimated lift is (58 − 40) − (36 − 30) = 12 orders.

That estimate assumes both areas would otherwise have experienced a similar change. Check several earlier periods for comparable trends, and watch for local events, delivery problems or offers crossing area boundaries. With small numbers, treat one test as directional evidence rather than proof.

Review on a fixed schedule

Check tracking during the first few days, but judge performance only after the response and completion windows close. Keep a weekly view of spend, enquiries, completed sales and contribution. Mark late responses separately so they remain visible without silently changing the original test.

Use the pattern to choose the next action. Many enquiries but few completed sales suggest problems with qualification, availability or the offer. Few responses despite confirmed distribution suggest the audience, message or response route needs work. Change one major variable at a time when possible.

Conclusion: measure enough to make the next decision

Start with one outcome, distinct tracking routes and a complete cost record. Then add a comparison that tests whether sales actually increased. You do not need perfect attribution; you need evidence strong enough to guide the next pound you spend.