Separate what you can prove from what dies inside the app
Start by separating what you can prove from what you cannot. Direct orders through your own site are fully trackable: you see the source, the device, the KD value. WhatsApp catering enquiries are trackable if you use one number and log every quote. Reservations are trackable if the host asks one question at the door. Promo codes are trackable anywhere, including inside Talabat and Deliveroo. Everything else — someone seeing a Reel on Thursday and opening the aggregator app on Friday — is invisible. That is not a reporting failure. It is how the platforms are built, and no agency dashboard fixes it.
So build the channel you can actually measure. A restaurant that takes its own delivery and pickup orders through its own store on Shopify with KNET at checkout keeps the customer's number, the order history and the full attribution trail — and keeps the 25–35% the aggregator takes. You will not move all your volume there. Moving 15–25% of it inside a year is realistic, and that slice is the only part of your business where you can honestly say one dinar of ads returned four. Price direct the same as the app so switching is easy for the customer.
Promo codes are the cheapest bridge between an ad and an order. Give each channel its own code — one for Snapchat, one for Instagram, a separate one per influencer — and make it work inside the aggregator too. Redemptions undercount real impact, usually by a lot, because plenty of people order without typing anything. Treat the code count as a floor, not a total. If a Snapchat code pulls 60 redemptions in a week on 200 KD of spend, the true influence is higher. But you now hold a number that is real instead of a screenshot of reach.
Ready to start your Shopify store?
Start a free trial and try the platform for yourself.
Read covers, aggregator orders and spend on one weekly sheet
Put four numbers on one weekly sheet: ad spend, covers or dine-in receipts, total aggregator orders, and average ticket. Weekly, not daily — restaurant demand in Kuwait swings hard by day, and Friday alone will lie to you. Then look at the shape of it. Spend went from 0 to 300 KD, aggregator orders went from 180 to 240, average ticket held steady. That is a real 60-order lift and you can put a value on it. If order count held flat and only the ticket moved, your ads did nothing and your menu pricing did everything.
Check the calendar before you blame the campaign. Ramadan flips everything: iftar volume, the suhoor tail, catering enquiries. Eid week is a spike, then a hole. Hala February brings footfall to Salmiya and the Avenues that has nothing to do with your ads. In July and August half your regulars are out of the country, and the same 300 KD buys 30–40% fewer orders. Compare a week to the same week last year, or at minimum to the four weeks before it. Never compare it to the peak you had during Ramadan.
Give any change three to four weeks before you judge it. A new campaign, a new photo set, a new code — the first week is noise, the second is learning, the third and fourth are the answer. Owners who kill a campaign on day five and try a new agency every month never accumulate one clean read. Freeze a single variable, hold spend steady, and let the weeks stack up. Four weeks at a flat 250 KD tells you more than twelve weeks of jumping between 100 and 600 KD.
Convert the return to fils, then check the P&L before the campaign
Here are honest numbers. On direct orders and catering, a healthy return is 3–6 KD of revenue per 1 KD of ad spend once the account is past its first month. On aggregator-heavy accounts the measured return usually reads 1.5–3 KD, and the real figure is higher than that because attribution is broken. Below 2 KD on direct orders, something is wrong with the offer, not the targeting. Above 8 KD sustained and you are either under-spending or someone is counting organic orders as paid ones.
Now convert that into profit, because revenue return is the number agencies quote and the number that misleads owners. A shawarma plate at 2.5 KD with 32% food cost, 30% aggregator commission and packaging leaves you a few hundred fils. At 4 KD of revenue per dinar spent, that same dinar returns roughly 60–90 fils of real contribution — genuine, but not the picture ROAS paints. Do this arithmetic once for your three best sellers and you will make better marketing decisions than any dashboard will hand you.
This is why marketing takes blame it did not earn. If food cost sits at 38% instead of 30%, if you pay 30–35% to the aggregator on your highest-volume item, if the kitchen comps ten meals a week for late tickets — no ad account fixes that. The tell is simple: orders are up and cash is not. When that happens, stop reviewing the campaign and review the P&L. Marketing can bring more of the wrong order faster. It cannot make an unprofitable dish profitable.