How to Launch a Meta Advantage+ Campaign With AI UGC Creative

Advantage+ needs recent purchase data to learn from. Before you build anything, check Events Manager for your pixel and Conversions API combined purchase count over the last 7 days. If you're consistently seeing 50 or more attributed purchases a week, you're in good shape. Below that, the algorithm has too little to work with and results will be volatile for weeks rather than days.

If your store is new or purchase volume is thin, don't skip straight to Advantage+. Run a simpler manual conversion campaign with slightly broader targeting for two to three weeks first, spending enough to accumulate purchase events. That manual phase isn't wasted effort — it's what feeds the pixel the signal Advantage+ will later rely on, and it also gives you a baseline CPA to judge the Advantage+ campaign against once you do launch it.

Also check Conversions API is actually connected, not just the browser pixel. Browser-only tracking has been degraded by cookie restrictions and ad blockers for years, and campaigns relying solely on it under-report purchases, which starves the algorithm of exactly the data it needs most. Conversions API sends events server-side and typically recovers a meaningful share of the purchases the browser pixel alone would miss.

Advantage+ shopping campaigns pull from your product catalogue in Commerce Manager, so sync it via your Shopify app or a scheduled feed before you touch Ads Manager. Every active SKU needs an accurate price, live availability status and at least one clean product image — items with an outdated price or 'out of stock' flag get silently deprioritised by the delivery system.

Open Commerce Manager and run the diagnostics check on your catalogue before building the campaign. Look specifically for items flagged with missing GTIN or brand fields, mismatched prices between the feed and your live site, and images under the minimum resolution — these are the errors that quietly shrink your eligible inventory without ever throwing an obvious error in Ads Manager.

If you sell fewer than roughly 10 SKUs, catalogue quality matters even more, because there's less inventory for the algorithm to rotate through if a handful of items get suppressed. For small catalogues, manually review every listing rather than trusting the automated feed sync, and re-check after any price change or restock — feed delays of a day or two are common and can leave stale data live longer than you'd expect.

In Ads Manager, click Create, choose the Sales objective, and select 'Advantage+ shopping campaign' as the campaign type. This locks the campaign into automated targeting and placements — you won't see the usual detailed targeting fields at ad set level because Advantage+ campaigns run on a single ad set structure with the budget set at the campaign level.

Resist the urge to duplicate an old manual campaign into Advantage+ structure — build it fresh. Manual campaigns carry ad set-level exclusions, custom audiences and placement restrictions that don't translate cleanly, and a messy migration is a common reason teams see worse-than-expected results in week one and wrongly conclude Advantage+ underperforms for their account.

Load a minimum of three ad variations, ideally six or more. Advantage+ tests creative across a much wider audience than a manual campaign, so a single hook and angle gets shown to people it doesn't fit. Mix formats: two or three AI UGC videos with different hooks, one or two static product shots, and a catalogue-based dynamic creative if you have a large SKU range.

Think of the creative set as covering different reasons to buy, not just different edits of the same idea. If your product solves a pain point for one audience and is a gift idea for another, make sure at least one video speaks to each — Advantage+ will find the pockets of audience that respond to each angle far faster than you could by hand, but only if you give it more than one angle to test.

This is where an AI UGC workflow pays off, because generating four or five distinct hook variants from one product brief takes minutes rather than a full creator shoot per angle. Klip Kanvas lets you paste your product link, generate several avatar-led scripts with different hooks, and export all of them in the 9:16 and 1:1 ratios Advantage+ placements actually use, so the creative side of setup doesn't become the bottleneck.

Set a campaign budget you're comfortable testing with for at least a week — as a rough floor, budget for around 50 purchases at your historical CPA so the campaign can exit learning. Then find the 'existing customers' budget control, usually under audience settings: this caps what share of spend can go toward people who already purchased or engage with your business.

For a campaign whose job is finding new customers, leave the existing-customer cap at its default low setting, typically in the 20% range, so most of the budget prospects rather than re-serves people who already bought. If retention and repeat purchase are part of the goal for this specific campaign, you can raise the cap, but then you're deliberately blending prospecting and remarketing spend into one number, which makes the resulting CPA harder to read at a glance.

Watch the 'new vs existing customer' breakdown in reporting weekly rather than setting the cap once and forgetting it. If a growing share of purchases is coming from existing customers despite a low cap, that's usually a sign your excluded customer list needs refreshing rather than a sign to fight the algorithm further.

Advantage+ intentionally limits manual targeting to broad guardrails: age range, gender and geography. Don't try to recreate detailed interest targeting through workarounds — the entire point of the campaign type is letting the algorithm find audience segments you wouldn't have thought to target manually.

The one control worth using deliberately is geography, especially if you can't fulfil orders in certain regions or if a region has meaningfully different unit economics because of shipping cost. Set geography once at launch based on real operational constraints, not based on a guess about who your customer is — that guess is exactly what the algorithm is built to test faster than you can.

Once live, leave the campaign alone. Meta's delivery system needs roughly 50 optimisation events to exit the learning phase, and every meaningful edit — budget change, new creative, audience tweak — resets that count. Give it three to four days minimum before you touch anything, even if early results look mixed.

Set a calendar reminder rather than checking the dashboard daily out of anxiety. Checking is harmless; editing isn't. The most common self-inflicted problem we see is a budget bump on day two because CPA looked high, which restarts the learning clock and produces exactly the volatile results the change was meant to fix, turning a normal three-day rough patch into a two-week one.

If you must make a change before the campaign exits learning — a genuine catalogue error, for instance — make it once, make it complete, and then commit to leaving it alone again. Multiple small edits spread across a week are worse than one larger edit made early, because each edit re-triggers the same signal-gathering process from a lower starting point.

Once the campaign exits learning, judge it against the CPA benchmark you set in step one, not against a single day's numbers. Scale budget in increments of roughly 20% every two to three days rather than doubling it, which tends to reset learning again. Refresh creative when frequency climbs past roughly 2.5–3, a common signal of fatigue.

Accounts we work with usually see Advantage+ campaigns settle into a stable CPA range within 10 to 14 days of consistent, uninterrupted delivery. If CPA is still more than double your manual-campaign baseline after a full two weeks at stable spend, that's a legitimate reason to pause and rebuild — either the catalogue, the creative set or the budget cap needs a rethink, not just more patience.