A restaurant campaign attribution window is the period after a campaign interaction during which an order can be credited to that campaign. A 30-day window will usually capture more eligible orders than a seven-day window, but that does not mean the campaign caused those extra sales.
Choose the window by looking at when your guests actually order and whether the report can explain each credit. Before buying software, ask the vendor to show the same campaign under both windows, identify overlapping credits, and export the underlying orders.
Define the dates and the decision
Start with four dates. The send date is when the message goes out. The click date is when a guest clicks its link. The order date is when the order is placed. The attribution date is the interaction date from which the reporting window starts. A vendor might use the send, a click, or another event as that starting point.
Those choices matter. A guest could receive an email on Monday, click it on Thursday, and order the following Wednesday. That order is nine days after the send but six days after the click. It falls outside a seven-day send-based window and inside a seven-day click-based window.
Ask which timestamp the report uses, which time zone applies, and whether the boundary includes an order exactly seven days after the starting event. Also ask what happens if a guest clicks more than once: does the clock start at the first click, the latest click, or another event?
The inputs should be identifiable records, not just a campaign total. Ask for the campaign ID, interaction timestamp, order ID, order timestamp, order value and matching customer identifier or other matching method. Check the source of each field. Campaign interactions may come from a marketing tool; order records may come from an ordering system or restaurant POS system. Ask how records are matched and what happens when a guest cannot be matched.
An order timestamp is an observed record. The decision to credit an order because it occurred within a window is a reporting assumption. Record that distinction before comparing vendors or campaigns.
Compare seven-day and 30-day results
Use the following hypothetical example as a worksheet. A campaign has one tracked click at the start of day zero. Assume every listed order is matched to that click, no other campaign is eligible, and the window includes orders placed up to exactly seven or 30 elapsed days after the click. Both reports cover the same send cohort and include all listed orders. Values are order sales, not profit.
| Possible outcome | Orders by elapsed day | Seven-day sales | 30-day sales |
|---|---|---|---|
| A: meaningful later orders | Day 3: $40; day 6: $30; day 12: $50; day 25: $80 | $70 | $200 |
| B: few later orders | Day 3: $40; day 6: $30; day 12: $10 | $70 | $80 |
Read each row as a separate possible outcome, not as two reports to add together. In outcome A, the longer window adds $130 in reported sales: $50 on day 12 plus $80 on day 25. In outcome B, it adds only $10. The orders themselves have not changed; only their eligibility for campaign credit has.
These figures exclude fixed costs, variable costs and exception costs. They are not an all-in operating cost or a return calculation. No discount, message charge, processing fee, refund or canceled order is included. Ask whether a real report uses gross order value, net sales after discounts, or another definition, and how it handles refunds and cancellations.
If your aim is to judge a short-lived lunch offer, seven days may be the more useful primary view. In outcome A, the large later difference makes a 30-day secondary view worth investigating, but it does not establish that the message drove those later orders. If your campaign is intended to encourage a less frequent visit, ask for both views and review the timing of later orders before choosing one as your main measure. In outcome B, the choice changes the reported total only slightly.
Keep the chosen rule consistent when comparing campaigns. If you change it, label the old and new reports rather than presenting the difference as growth. A loyalty program may introduce other reasons for a guest to return, so ask whether a campaign report can distinguish an offer redemption from an order merely placed within the window.
Find overlapping campaign credits
Return to outcome A. Suppose a second campaign produces a tracked click for the same guest on day 10, before the $50 order on day 12. Under a 30-day window, both clicks could be eligible for that order. If each campaign independently reports the full $50, adding their campaign totals would count that order twice.
Ask the vendor to show its credit rule. Does it give full credit to every eligible campaign, assign the order to one campaign, split credit, or offer more than one view? If the rule assigns all credit to the most recent eligible click, the $50 goes to the second campaign. Assuming the other outcome A orders remain credited to the first campaign, its 30-day total falls from $200 to $150. Its seven-day total stays $70 because the day 12 order was outside that window already.
That is a reporting-rule change, not a $50 loss of sales. The restaurant still has the same orders. A campaign manager should review overlapping order IDs before combining campaign totals. If the report gives full credit to multiple campaigns, keep campaign figures separate or use a deduplicated order-level view for the combined total. Ask the vendor how to produce that view; do not assume the dashboard has one.
Other exceptions deserve the same attention. A shared device or unmatched guest can make a link between click and order uncertain. A canceled order can remain in a report until its data is updated. Ask how the system flags unmatched records, duplicate IDs and later order adjustments, and who can review them.
Set a pass-or-fail vendor demonstration
Give the vendor a small fictional dataset before the demo. Include the outcome A orders, the second campaign click and a canceled or refunded order as a separate exception. State the credit rule you want to inspect and the expected treatment of that exception. The exercise should be simulated: it must not send a guest message, take a real payment or dispatch a courier.
Have the campaign manager own the result, with the person responsible for order reconciliation checking the order records. Use this checklist:
- Show the settings. Identify the starting event, elapsed-time boundaries, time zone, matching method and sales definition.
- Switch the window. With the same fictional cohort and order data, show $70 for seven days and $200 for 30 days under the single-campaign assumptions in outcome A.
- Show overlap. Add the second eligible click and explain whether the day 12 order is credited once, split, or shown in both campaign reports. If using the stated most-recent-click rule, show $150 for the first campaign's 30-day total.
- Inspect exceptions. Show how the separate canceled or refunded order is treated and whether an unmatched order is visible for review.
- Reconcile the export. Provide order-level records with enough dates, identifiers, values and credit information for the reconciliation owner to reproduce the displayed totals.
Pass: the vendor demonstrates the agreed rules, produces the expected totals for the applicable scenario, explains the overlap and exception treatment, and provides an export that reconciles. Fail: a total cannot be reproduced, an overlap is hidden, or the vendor cannot explain the rule behind a credit. Record any unresolved exception with the campaign manager as owner and ask the vendor for a written resolution before relying on the report.
After a successful simulation, a limited live service pilot can check how real records arrive and how staff handle exceptions. Agree on its scope and permissions first. Do not treat either the simulated result or the pilot's attributed sales as proof that a campaign caused those orders; a holdout test addresses a different question: what changed because guests received the campaign.
Choose a reporting rule your team can maintain
Write down one primary window, its starting event, the order-value definition and the overlap rule. Decide when staff will review late orders and adjustments. If you use a second window for context, label it clearly and avoid adding the two totals together.
For your next buying conversation, compare these requirements with Fleksa's restaurant marketing tools and ask which attribution settings, exports and exception reviews are currently supported. If order data comes from other restaurant management software, include the matching and reconciliation steps in that conversation.
Frequently asked questions
Does a longer window increase actual sales?
Not by itself. It can assign more existing orders to a campaign report. Establishing whether the campaign caused additional sales requires a different comparison, such as a suitable holdout test.
Should the window start at the send or the click?
That depends on the question and the interactions the system can record. A click-based rule measures from an observed click; a send-based rule starts earlier. Ask the vendor to demonstrate both if they are important to your decision.
Can I add attributed sales from several campaigns?
Only after checking the credit and deduplication rules. If one order receives full credit in more than one campaign, adding those reports overstates the combined order value.
