Home › Guides › Facebook ads for ecommerce
Facebook ads for ecommerce: a set-up playbook for UAE online stores
Facebook ads for ecommerce work when Meta can see real sales. Send a Purchase event with its value in AED through the Pixel and the Conversions API, connect a clean product catalogue, then run one Advantage+ sales campaign with enough budget for about 50 purchases a week. Scale only when store orders roughly match what Meta reports.
If you run an online store in the UAE, you may already have tried Facebook ads for ecommerce: a few boosted posts, a traffic campaign to the homepage, or a structure copied from a 2022 tutorial. The ads got clicks and some orders, and nobody could say whether Meta paid for itself.
The problem is rarely the creative. It is what Meta is told about sales, and how much room the campaign has to learn. Meta has also changed catalogue campaigns, so much of the advice online describes switches that no longer exist. Here is the order I set up a store selling in AED.
Start with tracking: a Purchase event that carries the value
A sales campaign optimises towards the event you choose, so Purchase has to fire once per order and carry the order value in the right currency. Otherwise every number in Ads Manager, from cost per purchase to ROAS, is built on bad data.
- Value and currency. Meta's page on setting the value and currency of Pixel events says its standard event code defaults to $0.00 USD. Send the real order total with the currency code AED, as a plain decimal with no symbol or commas: 1499.00, not AED 1,499.
- The three catalogue events. To show products from your catalogue, Meta requires ViewContent, AddToCart and Purchase, each with content_ids (or contents) matching the product IDs in the catalogue.
- The Conversions API next to the Pixel. In the browser, ad blockers, browser settings and patchy connections can lose Pixel events; the Conversions API sends them from your server or store platform, and many platforms connect it without a developer. If you send Purchase from both, give both copies the same event name and event ID so Meta counts the order once. My Meta Conversions API setup guide covers the options.
Place a real test order and check it in Events Manager before you spend. A store that reports AED 0 purchases, or every order twice, trains Meta on the wrong customers.
What I check first in a UAE store account. The currency on the Purchase event, because a snippet copied from Meta's example still says USD. Whether Purchase fires once per order, not on every refresh of the thank-you page. And cash on delivery: Purchase usually fires on the order confirmation, so a refused parcel still counts as a sale. If refusals are common, compare Meta's purchase count with delivered orders weekly, and consider sending a separate server event for paid orders through the Conversions API, so Meta learns from real sales.
Build the product catalogue before the ads
A catalogue is the product list Meta uses to show each person relevant items; it lives in Commerce Manager inside your business portfolio and connects to your Pixel. Then check the catalogue match rate, which shows whether your events match catalogue products. A low rate often means the IDs on the site and in the feed disagree.
- Arabic and English. One catalogue can hold country and language feeds, so each person sees product details in their language. If your site has an Arabic version, add an Arabic feed rather than a second catalogue. One ad can carry headline and text for up to 48 more languages.
- AED prices that match the site. If the ad shows a price the shopper cannot find at checkout, you lose the sale and their trust.
- Product sets. A product set controls which products can appear in an ad, so group what you want to push: bestsellers, high-margin lines or a seasonal range.
What changed in catalogue campaigns
Meta has renamed Advantage+ catalog ads to "ads with product media" and moved the catalogue choice to the campaign, according to its page on changes to sales campaign setup using a catalog. A tutorial that tells you to find the Advantage+ catalog ads switch is out of date. In new Sales campaigns:
- Product data is on at campaign level by default when you have a catalogue. Keep it on.
- Product media is a creative source you pick per ad. Choose it to show catalogue products, or upload your own media; both still use the catalogue's data.
- One catalogue per campaign. If you mixed catalogues before, split the campaigns or merge the catalogues.
- Product set per ad, dataset per ad set. If several Pixel datasets connect to the catalogue, pick the right one for each ad set.
Older campaigns keep running as they were, but Meta moves a duplicate of one to the new set-up where possible, so check the settings on any copy.
The Facebook ads for ecommerce campaign structure to start with
For a store with working tracking, I start with one Advantage+ sales campaign. Meta calls Advantage+ sales campaigns the most efficient way to drive online sales with the Sales objective, with its most advanced optimisation when Advantage+ audience, placements and campaign budget are all on. It reports a 9% average improvement in cost per conversion; that is Meta's figure, and the page gives no method or sample (checked 22 November 2026).
- One or two ad sets, not ten. Meta's learning-phase guidance warns that many ad sets each learn less. Combine similar ones.
- A mix of creative. Product media ads for range, plus your own video and images for the story: delivery speed, returns, why buy from you and not a marketplace. Meta's best practices ask for a wide variety of creative.
- Audience segments. Meta also recommends building audience segments, so reports split results into new, engaged and existing customers. That shows whether spend is finding new buyers.
- Placements left open. Meta's developer documentation says any placement exclusion switches the campaign's Advantage+ state off, so removing Audience Network or Messenger out of habit costs you the set-up you chose.
Old Advantage+ shopping campaigns could cap spend on existing customers; Meta's developer documentation says new campaigns cannot, and suggests a separate existing-customer ad set with a spending limit. I add one only when the segment report shows too much spend going to people who already buy.
Budget: enough for about 50 purchases a week
Meta's learning-phase page says an ad set usually leaves learning after about 50 results in the week after its last significant edit, and costs run higher and less stable until then. That sets a floor for the budget.
As an illustration: at a cost per purchase of AED 60, 50 purchases a week needs about AED 3,000 a week, or AED 430 a day. At AED 150 per purchase it needs about AED 7,500 a week. Use your own cost per purchase from the first weeks; my guide to setting a Facebook ads budget works through the maths.
If the budget cannot reach that, accept a slower, noisier learning period, or optimise for a higher-volume event such as AddToCart while purchase data builds. That learns faster, but Meta then finds people who add to cart, who do not always pay, so I switch back to Purchase as soon as volume allows. Once it is learning, leave it alone: Meta advises against edits and frequent budget changes during the learning phase.
When to switch to value optimisation
Value optimisation makes Meta chase the highest total order value, not the most orders. Meta's page on maximizing the value of conversions says it needs value signals through the Conversions API (required) and the Pixel. Its requirements page recommends at least 100 website conversion events with at least 5 distinct values in the past 14 days, and says the option can take a few weeks to appear after you qualify.
It is worth testing when order values vary a lot, for example a fragrance store selling single bottles and gift sets. Meta's advice for the first test:
- Run it for at least 3 weeks with budget for 50 or more conversions a week.
- Start without a ROAS goal, and add one after the first test.
- Expect a higher cost per result and fewer purchases. Judge it on revenue and ROAS, not order count.
Judge results on store orders, not Ads Manager
Ads Manager reports the purchases it attributes to your ads; your store reports orders, and your bank what was paid. Compare all three weekly: a campaign that looks profitable in Ads Manager can lose money once refused parcels and returns come out.
Set the target first: your break-even ROAS from your margin, plus the profit you need. My guide to what a good ROAS is shows the maths; a store on a 30% margin needs a very different ROAS from one on 70%.
Run Meta next to Google, not instead of it: Google catches people already searching, Meta reaches people who were not looking yet. My Performance Max guide for e-commerce covers the Google side; the Skeidar e-commerce account saw +103% ROAS with Performance Max.
The set-up order
- Fix the Purchase event: one per order, real value, currency AED, tested with a real order.
- Add the Conversions API with deduplication, and check event counts in Events Manager.
- Build the catalogue with AED prices, an Arabic feed if your site has one, and product sets.
- Connect the Pixel to the catalogue and check the match rate.
- Launch one Advantage+ sales campaign: Product data on, product media plus your own creative, placements open.
- Budget for about 50 purchases a week, or optimise for an earlier event until you can.
- Let it learn, then judge it on delivered, paid orders against break-even ROAS.
- Test value optimisation once you have the volume and a spread of order values.
If you would rather have someone build and run this for your store, that is what my Meta Ads management covers.
Frequently asked questions
Are Facebook ads worth it for an ecommerce store in the UAE?+
What is the best Facebook campaign type for ecommerce?+
How much should I spend on Facebook ads for my online store?+
What happened to Advantage+ catalog ads?+
Related
Running Meta ads for your online store?
Message me on WhatsApp or request a free campaign audit. I'll check what your Purchase event is reporting, whether your catalogue matches your site, and whether your budget gives Meta enough sales to learn from.