GROWTH-SPEND GUIDE
How to evaluate restaurant discounts and advertising without mistaking sales for profit
A campaign can increase orders and still reduce contribution. The right question is not “How much campaign sales did we get?” but “How much additional contribution did the campaign create after discount and advertising cost?”
Separate who funds the discount
Classify every offer into restaurant-funded, platform-funded and shared-funded portions using the authorised report or agreement. A customer may see one discount while the settlement assigns the cost differently. Only the restaurant-funded portion directly reduces the restaurant’s order revenue, although other offer terms may influence fees.
Do not treat a platform-funded amount as restaurant revenue unless the settlement actually reimburses it under your accounting definition. Trace offer rows to order or settlement identifiers where available.
Measure campaign economics
The difficult part is the baseline. Comparing with the immediately previous day can be misleading because weekday, weather, season, menu availability and outlet uptime affect demand. Use comparable days or a simple test-control design when possible.
Worked campaign example
A campaign produces 300 attributed orders and ₹120,000 of restaurant revenue. After food, packaging, platform services, funded discounts and expected refunds—but before advertising—the orders contribute ₹24,000. Advertising cost is ₹15,000, leaving ₹9,000 campaign contribution.
Suppose comparable organic behaviour suggests 180 of those orders would have occurred anyway and would have contributed ₹12,000. Incremental contribution after advertising is therefore negative ₹3,000: ₹9,000 campaign contribution minus ₹12,000 expected baseline. Revenue and attributed orders looked strong, but the incremental result does not justify repeating the campaign unchanged.
| Measure | Value |
|---|---|
| Attributed restaurant revenue | ₹120,000 |
| Contribution before ad spend | ₹24,000 |
| Advertising spend | ₹15,000 |
| Campaign contribution | ₹9,000 |
| Expected baseline contribution | ₹12,000 |
| Estimated incremental contribution | −₹3,000 |
Look beyond first-order economics
A first order can be intentionally low-margin when it creates valuable repeat demand, but this must be measured rather than assumed. Track cohorts by first campaign exposure, then compare repeat order rate and contribution over a defined period. Use privacy-respecting aggregated data and do not fabricate customer identities when the source does not expose them.
Define a payback window. If an acquired customer has not recovered the initial subsidy within that window, the campaign needs a different audience, offer or menu bundle.
Campaign checklist
- Confirm the funded share of each discount.
- Use delivered orders, not only clicks or initiated orders.
- Deduct variable platform and packaging costs.
- Include the advertising invoice, not an incomplete dashboard snapshot.
- Compare like-for-like days and outlet uptime.
- Watch for cannibalisation of full-price organic orders.
- Set a contribution stop-loss before launch.
- Review repeat contribution by cohort after the campaign.
Frequently asked questions
Is ROAS enough?
No. Revenue divided by ad spend ignores food, packaging, commission, discounts and refunds. Contribution-based return is more useful for profitability.
What if later daily ad rows are missing?
Do not assume zero spend. Record the coverage gap and confirm against invoices or the authorised campaign dashboard.
Should every low-margin promotion be stopped?
Not automatically. It may support acquisition, launch or capacity utilisation, but define the objective and payback evidence in advance.
Illustrative analysis only. Campaign attribution and commercial terms vary by source.