Meta's Advantage+ automation has quietly rewritten how Facebook and Instagram ads work. Audience settings are now suggestions, not instructions — the algorithm decides who sees your ad based on how people respond to your creative. Here is what changed and how advertisers in Jordan and the GCC should restructure their accounts.
Quick Answer
Under Meta's Advantage+ system the algorithm decides who sees your ad based on how people respond to it — not the audience settings you enter. Your creative is now the primary targeting mechanism: a specific hook draws the right audience faster than any interest stack. Consolidate campaigns, build 6–10 genuinely distinct creative concepts, and maintain hard guardrails (geography, exclusions, CAPI) rather than narrow audience constraints.
What Actually Changed Inside Meta's Delivery System
If you opened Ads Manager five years ago, the skill that separated a good media buyer from an average one was audience construction. Lookalike percentages, layered interest stacks, exclusion audiences built like legal contracts. You could spend a full afternoon on targeting alone, and it showed in the results.
Open Ads Manager today and most of that work is either gone or actively working against you. Meta has spent the last two years rebuilding its ad delivery system around one idea: the advertiser tells the machine what outcome they want, and the machine figures out who should see the ad. Advantage+ started as a shopping campaign type. It is now the direction of the entire platform — Meta reports advertisers using Advantage+ Shopping campaigns see an average 32 percent improvement in cost per acquisition versus manual campaigns.
Meta's older delivery model treated your targeting settings as instructions. The current system treats them as suggestions. When Advantage+ audience options are on — and Meta keeps making them harder to turn off — the audience you define is a starting signal, not a boundary. The algorithm will deliver outside your defined audience the moment it believes it can find conversions there.
Why 'The Creative Is the Targeting' Is Literally True
In practice, the system decides who sees your ad based on how people respond to it. Watch time, saves, shares, comments, click quality, and post-click conversion behaviour teach the algorithm what kind of person your ad is for. Your image, your video, and your first line of copy are doing the job that interest targeting used to do.
Think about what an ad communicates in its first few hours of delivery. If your video opens with a wide lifestyle shot and generic copy, Meta gets a weak signal — the people who stop on it are a mixed group, and the system takes longer and spends more of your budget figuring out who converts. If your video opens with 'If you run a restaurant in Riyadh, your delivery margins are lying to you,' the algorithm gets a razor-sharp signal within hours. Restaurant owners stop. Everyone else scrolls. The engagement data draws a clean outline around your actual customer far faster than any interest stack could manage.
Calling out your audience in the creative itself — in the first line, in the visual, in the voiceover — is no longer a copywriting nicety. It is the most reliable targeting mechanism you have left. Meta will override your audience settings. It cannot override your hook.
Creative Diversity Is Now a Delivery Requirement
There is a second consequence that gets less attention: when one ad concept carries the whole account, the algorithm builds one narrow model of your customer. Different concepts — different formats, different angles, different problems addressed — let the system discover pockets of your market that a single ad would never reach. Accounts running the same visual treatment on repeat are being punished for it, and creative fatigue sets in faster than most teams realise.
Campaign Consolidation: Fewer Campaigns, More Signal
The old best practice was segmentation. Separate campaigns by audience temperature, by product line, by placement, sometimes by device. Every segment felt like control. Under the current delivery model, segmentation is mostly signal starvation. Ten campaigns splitting 3,000 JOD a month are each learning from a trickle of conversions, exiting and re-entering the learning phase, and competing against each other in the same auctions.
The consolidated structure that performs best for most accounts now looks unfashionably simple: one or two campaigns per real business objective, broad or lightly-constrained audiences, budget concentrated so the system gets enough conversion volume to learn properly, and the complexity moved into the ad level — where six to ten genuinely different creative concepts do the segmentation work. The 32 percent CPA improvement from consolidation is believable precisely because most fragmented accounts were paying a hidden tax: auction overlap, perpetual learning phases, and budget spread too thin for any ad set to reach statistical significance.
One important caveat: if you sell to genuinely distinct markets — a brand selling in both Jordan and Saudi Arabia with different pricing and different Arabic dialects in the creative — forcing everything into one campaign muddies your signal instead of strengthening it. Separate by real business difference, never by hypothetical audience difference.
What Still Deserves Manual Control
Exclusions still matter enormously. Advantage+ will cheerfully re-serve ads to people who bought yesterday if you let it, because a recent purchaser is statistically likely to convert again — even when that conversion is worthless to you. Excluding recent purchasers, existing customers where appropriate, and irrelevant geographies remains manual work worth doing.
Geographic boundaries matter more in our region than almost anywhere else. An e-commerce store that only ships within Jordan cannot afford delivery drifting into audiences it cannot serve. Broad targeting inside a hard geographic fence is the right configuration.
Budget pacing around regional commercial moments still beats the algorithm. Meta's system optimises continuously, but it does not know that Ramadan changes everything about consumption behaviour in this region, that White Friday compresses a quarter of e-commerce demand into two weeks, or that Saudi National Day briefly rewrites the auction. Anticipating these moments with budget and creative prepared in advance is human work — and it is where regional advertisers quietly outperform global templates.
Building a Creative Pipeline That Feeds the Machine
If creative is the targeting, then creative production is media buying. A working pipeline needs genuine variety: a founder talking to camera, a customer testimonial format, a product demonstration, a problem-agitation hook, a static image with a hard offer, a native-feeling vertical video. Each concept is a different fishing net thrown into a different part of the water. Six ads that are the same concept in different colours are one ad.
Hooks deserve disproportionate effort. Write ten hooks for every concept and let delivery data — not internal opinion — decide the winner. Maintain a kill-and-replace rhythm: review frequency and performance weekly, retire what is decaying, and keep a bench of fresh concepts ready. The accounts that struggle most are not the ones with bad creative — they are the ones with a slow pipeline.
Arabic-First Creative: The GCC Edge
Creative conceived in Arabic — with the dialect, humour, and rhythm of the market it targets — consistently outperforms English creative run through translation. The gap is not small, and the algorithm amplifies it: stronger resonance means stronger engagement signals, which means cheaper delivery. For advertisers targeting Saudi Arabia and the UAE specifically, this is the most reliably underexploited advantage in the market.
The Data Layer: What Feeding the Algorithm Actually Means
Start with the Conversions API, implemented server-side and deduplicated properly against your browser pixel. Deduplication is where most regional implementations quietly fail: events fire twice, Meta's models learn from inflated conversion counts, and the account optimises toward a distorted picture of reality. If your Events Manager shows poor deduplication rates, fixing that will likely do more for performance this quarter than any creative or structural change.
For lead generation, sending downstream signals back — lead qualified, meeting booked, deal won — via offline conversions turns Meta from a machine that finds you form-fillers into one that finds you customers. Most accounts in our market stop at the form fill, then wonder why lead quality decays. Customer lists also deserve more respect: uploading and maintaining your customer file gives the system a ground-truth definition of who actually pays you, sharpening everything Advantage+ does with its internal audience expansion.
Where This Is Heading
Meta is not going to reverse course. Every quarterly update moves more control from the advertiser to the system. The advertisers who thrive are the ones who moved their effort upstream: from managing settings to managing inputs. Clean first-party data, a fast and genuinely varied creative pipeline, hard business guardrails, and measurement that catches the algorithm when it optimises for the wrong thing.
For advertisers in Jordan and the GCC, there is a genuine edge available right now. Regional market knowledge, dialect-correct Arabic creative, and an understanding of local commercial rhythms are exactly the inputs the automated system rewards — and exactly the things a global template cannot supply. The machine levelled the playing field on settings. It raised the stakes on everything else.
Frequently Asked Questions
Ready to act?
Get a free audit of your ad accounts
No sales pitch, no commitment. We'll review your current setup and show you exactly where the budget is leaking.
Talk to an Expert
