For a store the meter is cost per purchase, not cost per lead
Cost per lead is the wrong meter for a store. On Snapchat and Instagram in Kuwait a message costs 0.300 to 1.200 KD, and it feels cheap until you count what happens after it. Roughly one in eight of those DMs turns into a paid, delivered order, so a 600 fils lead is really a 5 KD customer. Every store that optimises for cheap messages ends up with a full inbox and a flat bank account. The number you run the business on is cost per purchase: total ad spend divided by orders that were paid for, delivered, and not sent back.
Here is where Kuwait stores actually land, per delivered order, on Meta and Snapchat traffic. Fashion and abayas with a 15 to 35 KD ticket: 3 to 7 KD. Beauty, perfume and skincare: 2.500 to 6 KD. Supplements and coffee: 3 to 8 KD. Baby and kids: 4 to 9 KD. Home, kitchen and small furniture at a 40 to 150 KD ticket: 8 to 25 KD. Electronics and phone accessories, where margin is thin: 5 to 15 KD. Jewellery and gifting above 100 KD: 15 to 45 KD. Below those ranges you are usually being carried by an existing audience. Above them, something in the funnel is broken.
You cannot measure cost per purchase at all if the order arrives as a DM. The inbox gives you no purchase event, no basket value and no return rate, so you are spending on feel. Once orders run through a real store on Shopify with KNET and Tabby at checkout, every order is tied to the ad that produced it and you see the true cost per purchase instead of the cost of a conversation. The first month after moving orders out of DMs the number usually looks worse than you hoped. That is fine. You are finally seeing reality and can fix it.
Ready to start your Shopify store?
Start a free trial and try the platform for yourself.
Work your break-even CPA from margin, after the failures
Work the ceiling from your margin, not from somebody's benchmark. Say your average order is 25 KD and the goods cost you 9. That is 16 KD gross, and most owners stop right there and decide they can pay 8 KD per order. Now subtract what Kuwait actually charges you. The courier takes 1.250 to 2 KD per drop inside the country. Packaging is 0.300 to 0.600. KNET is a flat fee of around 0.100 per transaction, and Tabby takes 4 to 6 percent of the order value. You are down to roughly 12.500 KD before anything has gone wrong.
Then come the failures. On cash on delivery, 8 to 20 percent of orders are refused at the door, or the customer travelled, or nobody answers, and you pay the courier both ways on those. Fashion returns run 5 to 12 percent, higher on shoes and anything with a size. Spread those losses across every order and a 25 KD sale is contributing about 10 to 11 KD, not 16. That is your honest ceiling. Pay 10 KD per order and you have worked for free. Pay 6 to 7 and there is something left for salaries and restocking.
Now flip it into ROAS. Break-even ROAS is one divided by your contribution margin rate: 11 KD on a 25 KD order is 44 percent, so you break even at 2.3, not at the 1.5 people quote in meetings. The dashboard will never tell you this. Meta counts a purchase on a seven-day click and one-day view, so it claims orders that Snapchat and Google also claim, plus orders your repeat buyers would have placed anyway. Use blended instead: all store revenue divided by every dinar spent across every platform. In Kuwait it usually reads 30 to 50 percent below platform ROAS.
Pay more per order only when the second order is real
A higher cost per order is only justified when the second order is real, and you can check that in ten minutes. Pull every customer who bought 90 days ago and count how many bought again. Supplements, coffee, skincare, baby consumables and everyday abaya basics typically repeat at 25 to 40 percent inside 90 days. Furniture, wedding items, gold and gifting repeat at under 10 percent. If your 90-day repeat rate is 35 percent at the same margin, each acquired customer is worth about 1.35 orders, so a 7 KD ceiling moves to roughly 9.500 KD.
The catch is cash. Paying 9.500 today for margin that only lands in month three means you fund the gap yourself while stock has to be reordered before it arrives. Only stretch the CPA if you have the runway and the repeat rate is measured, not hoped for. Two rules keep this safe. Never let the stretched ceiling exceed the margin on the first order plus one repeat, and re-measure the repeat rate every quarter. If it drops under 20 percent, pull the ceiling back to the single-order number that same week, not at month end.
Season moves the ceiling too, in both directions. Ramadan and the two weeks before Eid push Kuwait CPMs up 30 to 60 percent, so cost per order rises even when everything is working. That one is worth paying, because basket sizes rise with it. July and August are the opposite. Half of Kuwait is out of the country, conversion rate drops, and a 7 KD order becomes a 12 KD order for no good reason. Hala February and back-to-school in September are the two cheap windows most stores underspend on every single year.