For a D2C launch in the UAE, start at roughly 55% Meta, 30% Google, 15% Snapchat, and expect that split to be wrong within six weeks. Meta buys you demand you don't have yet. Google catches the demand Meta creates, and almost nothing else in month one, because nobody is searching for a brand they've never heard of. Snapchat is the cheap-reach lever that works brilliantly for some categories and quietly incinerates budget for others. The point of the first 45 days of digital advertising for a UAE ecommerce brand isn't profit. It's finding out which of those three deserves 60% of your money by quarter two.
We run paid media alongside Shopify builds for brands in Dubai, Abu Dhabi and Sharjah, and the pattern that costs founders the most money is treating the UAE as one market with one language and one buyer. It isn't. It's four or five audiences stacked in a country the size of a mid-sized Indian state, each with different platform habits, different price tolerance and different reasons to buy.
Why Google gets less than you expect at launch
Search volume in the UAE is thin compared to the US or India, and it splits across English and Arabic, with a fair amount of transliterated Arabic typed in Latin script. A new brand in, say, clean skincare or modest activewear will find that non-brand search inventory is either expensive (dominated by Noon, Amazon.ae and Sephora-type retailers bidding on category terms) or so low-volume that you can't gather enough clicks to learn anything.
So we split Google budget at launch roughly like this: 55–60% into Shopping or Performance Max with a properly built feed, 25% into a tight non-brand search campaign on high-intent long-tail terms, and 15% held for brand defence. Brand clicks in the UAE are cheap, often under AED 2, and you will need them by week three, because Meta and Snapchat push people to Google to check whether you're real. If you skip brand protection, a marketplace listing of a competing product will take that click.
The feed matters more than the bidding. Products need Arabic titles where the audience searches in Arabic, correct GTINs if you're reselling, AED pricing with VAT handled consistently, and no size variants collapsing into one listing. We've seen PMax campaigns written off as "not working" when the actual problem was 40% of the catalogue disapproved for a missing attribute.
Meta carries the launch, but only if you segment the expats properly
The single biggest lift we get on UAE Meta accounts isn't from creative volume or from some bidding trick. It's from stopping the brand talking to "UAE, 18–55, all languages" as if that were an audience.
Practically, these are the segments worth separate ad sets, separate creative and separate landing pages:
- Western expats and higher-income professionals, clustered in Dubai Marina, JLT, Downtown, Al Reem. English creative, higher AOV tolerance, respond to brand story and ingredient or material detail. Buy on card, rarely COD.
- South Asian expats — Indian, Pakistani, Bangladeshi, Sri Lankan, plus Filipino buyers who behave differently again. Price-anchored, heavy WhatsApp users, far more likely to want cash on delivery, strong response to bundle and family-pack framing. Sharjah and Ajman skew heavily here and shipping costs are different.
- Emirati and Gulf Arab nationals. Arabic-first creative that was written in Arabic, not translated from English. Highest AOV in most categories we've worked in, and the least tolerant of a storefront that reads like a bad localisation.
- Levantine and North African Arab expats. Arabic creative works, but the register and the references differ from Gulf Arabic. Often the cheapest CPMs of the four.
You don't need all four live on day one. Pick two, based on who the product is actually for, and give each about AED 250–400 a day so the ad set can exit learning. Splitting AED 33,000 across nine ad sets is the most common self-inflicted wound in a UAE launch account.
Snapchat: real reach, unreliable attribution
Snapchat's usage in the UAE and the wider Gulf is strong enough among under-30s, particularly Arabic-speaking under-30s, that ignoring it is a genuine miss for fashion, beauty, F&B, phone accessories and anything gifted at Eid. In the accounts we manage, Snapchat consistently delivers cheaper reach than Meta for that demographic. It also, consistently, over-reports conversions.
Two things follow. First, don't judge Snapchat on platform-reported ROAS; judge it on blended CPA across the whole account when you turn it on and off. Second, creative built for Meta will not perform there. Vertical, fast, face-forward, sound-on, Arabic voice or Arabic text, and the offer visible inside three seconds. Repurposed 4:5 Instagram assets get skipped.
Where Snapchat earns nothing: high-consideration, high-ticket products bought by 35+ professionals. B2B. Anything requiring a long explanation. If your AOV is above roughly AED 600 and your buyer is a working parent, put that 15% into Google Shopping or into a second Meta creative angle instead. We tell clients this before they ask us to test it, and about a third of them test it anyway. Fair enough — it's cheap to find out.
TikTok deserves a mention as the fourth option. In the UAE it sits somewhere between Meta and Snapchat: better creative-driven discovery than Snapchat, weaker purchase signal than Meta, and it rewards brands that can produce eight to ten new videos a month. If you can't feed it, don't start it.
The arithmetic, with real numbers
Take a first-month budget of AED 60,000. At 55/30/15 that's AED 33,000 Meta, AED 18,000 Google, AED 9,000 Snapchat.
On the Meta side, assume a CPM of AED 30 — plug your own once you have two weeks of data. AED 33,000 at AED 30 CPM buys 1.1 million impressions. At a 1.2% CTR that's 13,200 clicks, so a CPC of AED 2.50. At a 1.4% site conversion rate, 185 orders. CPA = 33,000 ÷ 185 = AED 178.
Now the other side of the ledger. AOV AED 320, gross margin 60% = AED 192. Subtract AED 20 of subsidised last-mile shipping and AED 8 in payment gateway fees, and contribution per order is AED 164. Against a CPA of AED 178, you're down AED 14 an order before you've paid anyone a salary.
That gap is normal in month one. What closes it is not a lower CPM. It's AOV. Add a two-item bundle that lifts AOV from AED 320 to AED 400 and contribution goes to 240 − 28 = AED 212, against the same AED 178 CPA. AED 34 positive per order, on the same media spend. Or lift site conversion from 1.4% to 1.8% and CPA drops to 33,000 ÷ 238 = AED 139.
Which is why we'd rather spend the first two weeks of a launch fixing the product page, the bundle offer and the checkout than rewriting ad copy. The media math is usually not solvable inside the ad account.
Payments and returns change the ROAS you can afford
UAE COD is nothing like India's, but it exists, and it's meaningful for the Sharjah and Ajman segments. Where we've enabled it, undelivered or refused orders run high enough that COD orders need their own CPA target, not the blended one. Offering Tabby or Tamara instalments does lift AOV in fashion and electronics, at a merchant fee well above card. Build both into the contribution number before you set a ROAS target, or you'll scale a campaign that looks profitable in Meta and loses money in the bank account.
VAT at 5% is a smaller distortion than Indian GST, but decide early whether prices display inclusive of VAT. Switching later breaks your feed, your comparison shopping and half your creative.
Set up measurement before you spend, not after
Platform-reported conversions across Meta, Google and Snapchat will add up to more orders than Shopify recorded. That's not a bug you can fix; it's overlapping attribution windows. What you can do:
- Run one honest number as the scoreboard: total ad spend ÷ Shopify orders for the same period. Blended CPA. Everything else is diagnostic.
- Add a one-question post-purchase survey asking where the customer first heard about you. In launch months this is the most useful attribution data you'll get, and Snapchat is where it most often disagrees with the ad platform.
- UTM everything, including influencer and WhatsApp links, before the first riyal or dirham goes out.
- Get server-side conversions live for Meta and Google. On iOS-heavy audiences, and the UAE is very iOS-heavy, the difference in signal quality is not marginal.
Landing page speed is a media cost
Paid traffic lands on one or two pages, usually over mobile data, often on a page carrying four review widgets, a currency switcher, a WhatsApp button and an upsell app. When LCP on that page sits at 4 seconds, you're paying full CPC for visitors who leave before the hero image renders. Cutting a 3.8s LCP to under 2s on a launch landing page has moved conversion rate more reliably for us than any audience test.
Check the actual pages your ads point to, on a throttled 4G profile, not the homepage on office wifi. If you want the boring version handled automatically, our SwiftStore app scans and monitors PageSpeed and fixes a chunk of the common theme and script issues; if the problem is structural, that's a deeper speed engagement rather than an app job.
When to move the money
Reallocation rules we use, roughly:
Day 14. Kill any ad set that has spent AED 3,000 with no purchases. Don't touch the platform split yet — two weeks isn't enough on Google, where PMax needs conversion volume to stabilise.
Day 30. Compare cost per purchase across the three, then sanity-check it against the post-purchase survey. If Snapchat's platform CPA looks best but almost no survey respondents mention it, cut it to 5% and hold. If branded search volume has grown week on week, increase Google's share; that growth is the clearest signal your upper funnel is working.
Day 60. Now commit. Most of the UAE accounts we've taken from launch to steady state end up somewhere between 45/45/10 and 65/25/10, with the Google share rising over time as brand search compounds. Beauty and fashion aimed at young Arabic-speaking buyers is the exception — Snapchat can justify 25% or more there.
Seasonality will overrule all of this twice a year. Ramadan and the run-up to Eid shift both CPMs and buying hours dramatically, White Friday in November turns November into the most expensive auction of the year, and July–August goes quiet as expat families travel. Don't launch a brand in the second week of August and conclude that digital advertising in the UAE doesn't work.
The unglamorous prerequisite
None of the above matters if the store can't take the traffic. Arabic pages that break right-to-left layout, a checkout that only accepts one card type, product pages with a single image, no size guide in centimetres, no visible delivery timeline for Abu Dhabi versus Dubai. We've audited launch accounts where the media plan was sound and the CPA was ruined by a shipping calculator that failed at the last step on mobile Safari.
If you're building or replatforming ahead of a paid launch, our team in Dubai handles both sides of that — the storefront build and the media — which mainly means we can't blame each other. If you already have a store and want a second pair of eyes on where the paid budget is leaking, take the free audit; it takes about a week and you'll get the blended CPA math for your own numbers rather than the example ones above.