Advantage+ Shopping (ASC): The E-commerce Guide

Advantage+ Shopping Campaigns, ASC for short, stopped being an experiment a long time ago. On Meta s Q4 2024 earnings call, CFO Susan Li said the product had passed a 20 billion dollar annual run rate, growing 70 percent year over year. By mid 2025 it was the default way most e-commerce brands spend on Facebook and Instagram, and adoption had pushed past 80 percent of e-commerce advertisers according to figures cited from Meta s own earnings commentary. That scale creates a problem of its own. The advice around ASC is now a swamp of agency folklore, half remembered settings, and recycled myths repeated until they sound like fact. This guide cuts through it. We will cover the campaign structure, the catalog, the existing customer budget mechanics that Meta has quietly changed, the creative volume that actually matters, how to measure ASC honestly, and the specific cases where it wins or loses. Where a number comes from an agency rather than Meta, it is flagged plainly so you can weigh it for what it is.

What ASC actually is, and what it replaced

ASC is a campaign type where Meta s machine handles almost every lever you used to pull by hand. You set a budget, an objective tied to purchases, a catalog or a set of creatives, and the system decides who sees the ad, when, on which placement, with which product, and at what bid. There are no separate prospecting and retargeting ad sets in the classic ASC. The audience is the entire eligible pool, and the algorithm splits prospecting from re-engagement internally, without you drawing the line. This is the opposite of the old playbook where you built a dozen ad sets stacked with interest stacks and lookalikes, each one a guess about who might convert. The promise is simple and, when the fit is right, often delivered: feed the machine clean signals and good creative, and it will outperform your manual structure because it sees more of the auction in real time than you ever could from the dashboard.

There is a naming trap worth clearing up first. In early 2025 Meta began rebranding Advantage+ Shopping Campaigns as Advantage+ Sales Campaigns. The rename was not cosmetic. The scope widened to cover sales, leads and app installs, not just online shopping, and the structure changed: the new version lets you build multiple ad sets, much like a manual sales campaign, as long as Advantage+ Budget stays on to keep the Advantage+ status. Beta testing started around February 2025 with a wider rollout planned through the year. So when you read older guides talking about the single ad set ASC and newer ones talking about multiple ad sets, both are right for their moment, and conflating them is how people end up following instructions for an interface that no longer exists. In this guide ASC refers to that underlying engine, whatever Meta calls the button this quarter.

The catalog is the foundation, not a detail

ASC leans hard on your product feed. If the feed is wrong, every downstream decision is wrong. Meta needs accurate inventory, correct prices, clean titles, detailed descriptions, proper categories and high quality images. A product that shows out of stock when it is available kills delivery on your best seller, silently, with no error to warn you. A price that lags a sale shows the old number and breaks trust at the worst moment, right as the click lands. Practitioners running ASC at scale repeatedly land on a baseline of at least 50 active SKUs in a verified Meta catalog before the dynamic side has enough to work with, alongside a properly configured pixel and Conversions API. Treat the 50 SKU figure as practitioner guidance rather than a hard Meta rule, but the principle holds without argument: a thin or stale catalog starves the algorithm of the very inputs it needs to make good calls.

One catalog beats many. Meta s own help documentation strongly recommends a single catalog for all your advertising, because learnings consolidate inside one catalog and fragment across several. If you spin up a new feed, that catalog starts with none of the conversion history of the old one, and you pay for that reset in worse early performance while it relearns from scratch. A generic example: an outdoor gear retailer that kept separate catalogs per supplier found ASC oscillating between bestsellers because no single catalog held enough aggregated signal to settle. Merging into one feed, with supplier handled as a field rather than as a separate catalog, stabilised delivery within two weeks and stopped the daily lurching between products. The takeaway is structural, not clever: give the machine one well groomed source of truth, not five fragmented ones competing for the same learnings.

The existing customer budget cap, and why it changed

For a long time ASC s signature control was the existing customer budget cap. You set the maximum share of spend allowed to go to people already in your custom audiences of buyers, so you could push the campaign toward new customer acquisition. A cap of zero meant pure prospecting. A cap of 30 percent let the machine spend up to three tenths of the budget re-engaging existing buyers. It was the one dial that genuinely mattered for brands obsessed with new customer growth rather than reselling to the same list they already owned, and it gave a clean, single number to argue about in a meeting. Its simplicity was the point, and its loss changed how the whole campaign has to be thought about.

Meta has been retiring that cap. The replacement uses ad sets instead of a percentage slider. To run pure prospecting, you build a sales campaign with an ad set that explicitly excludes the custom audiences representing your existing customers, which reproduces the old zero percent cap. To split spend deliberately, Meta suggests two ad sets: one targeting existing customers, one targeting broad audiences while excluding those same buyers. With Advantage+ Budget on, the system optimises distribution between the two ad sets automatically, and you keep control through ad set spending limits. Much of this is documented in practitioner write-ups rather than a single tidy Meta help page, so verify the exact flow in your own account before you rebuild a campaign around it, because the interface has been moving and what was true last quarter may already have shifted.

Why does this matter beyond a settings change? Because the cap was how serious brands fought the lazy reflex of ASC: spend where conversions are cheapest, which is almost always your warmest existing buyers. Without a deliberate exclusion, a campaign labelled growth can quietly become a coupon machine for people who would have bought anyway. A generic example: a coffee subscription brand saw its ASC report a strong blended ROAS while new customer count flatlined for two months straight. The spend had drifted toward repeat buyers because nothing stopped it from chasing the cheap conversion. Rebuilding with an explicit existing customer exclusion dropped the reported ROAS but lifted genuine new customer volume, which was the actual goal the whole time. The lesson is that the easy number and the right number are rarely the same.

Creative volume: the lever you still own

ASC took targeting away from you and handed you creative in return. The auction sorts itself out; the ad is what you bring. Meta s 2025 best practice guidance points to roughly 6 to 10 creative variations per ASC campaign, mixing static, video and catalog formats. Agencies report that accounts running fewer than 4 creatives fatigue faster and see CPMs climb 20 to 30 percent within two weeks. Treat those percentages as agency observation rather than a Meta benchmark, but the direction is consistent across every source: too few creatives starve the machine of options and let fatigue set in. The point of volume is not to A/B test colours or swap a caption. It is to give the algorithm enough genuinely different angles that it can find the rare ad it actually wants to spend on, the one outlier that carries the account. Most of your creatives will get little spend, and that is not failure, it is the system doing its job: sorting the many ordinary ads from the handful it can scale, which only works if you hand it real variety to sort through in the first place.

Different angles beat near duplicates. Six versions of the same hero video with swapped captions is not six creatives, it is one creative wearing six hats, and the machine sees through it instantly. A useful mix for an e-commerce ASC pulls from distinct concepts: a founder talking to camera, a problem then solution skit, a user generated review montage, a clean product demo, a static with social proof, and a catalog format that lets Meta pick the product dynamically per person. A generic example: a kitchenware brand that had been refreshing one polished studio video every week plateaued, then shipped four genuinely different concepts in a single drop, and one of them, an unscripted customer talking about a single pan, became the top spender for the quarter. Variety, not polish, is what feeds ASC, and the brands that grasp this stop treating creative as a finishing touch.

Budget and the learning phase

ASC obeys the same delivery physics as any Meta campaign: it needs conversion volume to optimise. The widely cited threshold is around 50 optimisation events per week to exit the learning phase cleanly, and that is a Meta concept, not an invention of the blogs. Below that the system never gathers enough signal to settle, and performance stays jumpy and unreliable. Some practitioners report ASC reaching stability at a lower floor, around 25 conversions, faster than a manual campaign would; treat the 25 figure as practitioner observation rather than an official number. The practical reading is simple and unforgiving: if your store is not producing roughly 50 purchases a week, ASC will struggle to settle, and you may be better served building volume first with a broad manual campaign before migrating once the data is there.

On budget size, agency consensus lands on 50 dollars a day as a technical minimum and 100 dollars a day as a practical floor, with 150 to 300 dollars a day a common starting band for small to medium stores and 500 to 2000 plus for established brands. None of those are Meta published minimums, so read them as field ranges rather than rules. The deeper point is that budget and conversion volume are linked at the hip: a budget too small to generate 50 weekly conversions at your real CPA traps the campaign in permanent learning, burning money without ever stabilising. Do the arithmetic backwards. If your CPA is 40 dollars and you need 50 conversions a week, you need roughly 285 dollars a day just to feed the learning phase, regardless of what a generic blog minimum suggests is enough to start.

How many ASC campaigns? Almost certainly fewer than you think

A persistent myth says you need a stack of ASC campaigns, one per product line, one per audience theme, ten in total to cover your bases. It is wrong, and it actively hurts. If you run ten ASC campaigns competing for the same eligible audience, you create internal auction competition: your own campaigns bid against each other and drive your costs up while fragmenting the conversion signal each one needs to exit learning. Most accounts perform best with one or two ASC campaigns plus a handful of manual campaigns for specific jobs. Consolidation is not laziness, it is how you stop paying a tax on your own structure. The instinct to subdivide comes from the old manual world where granularity gave you control; in ASC, granularity just splits your signal and starves every fragment, which is the exact opposite of what you want.

Measuring ASC honestly: the incrementality problem

ASC reports a ROAS that looks great, and that is exactly where it gets dangerous. Because the machine spends where conversions are cheapest, it gravitates toward people with high purchase intent, including those who would have bought without ever seeing the ad. The platform takes full credit for those sales anyway. The number you should actually care about is incrementality: the conversions that happened because of the ad and not merely alongside it in time. One apparel brand running ASC found in a test that only 17 percent of reported conversions were truly incremental, a sobering reminder that platform attributed ROAS and real lift are different animals entirely. Treat that 17 percent as a single brand result, not a universal rate, but the gap it reveals between credited and caused is real, common, and quietly expensive when you ignore it.

There are honest ways to measure it. Meta offers Conversion Lift Studies, where users are randomly split into test and control groups, the control sees no ads, and the difference in conversions estimates true lift over roughly three to four weeks. Meta has also added Incrementality Optimization for sales campaigns, which uses historical holdout data to push delivery toward incremental conversions rather than easy ones the buyer would have made anyway. Outside the platform, geo based methods work well: an inverse holdout turns ASC off in some regions and keeps it on in others, then compares the gap; a causal impact test turns all ASC off for 15 to 20 days while holding everything else constant and measures the revenue lost. None of these is free or instant, but all of them answer the one question platform ROAS structurally cannot answer on its own, and the cost of running one is trivial next to the cost of scaling a campaign you wrongly believed was profitable.

When ASC wins, and when it does not

ASC is built for a specific shape of business and it shines there. The brands that win run an established store with steady conversion data, ideally 50 plus purchases a week, a clean catalog connected to Meta, pixel and Conversions API both firing, and products with relatively short purchase cycles and prices that do not demand weeks of deliberation. Reported wins are real when the fit is right. Marpipe s analysis put ASC at 4.52 ROAS against 3.70 for manual, a 22 percent edge; that is an agency figure, not Meta s. A widely cited fashion DTC case reported 2.8 times blended ROAS, 41 percent lower CPA and a 19 percent lift in new customer rate after moving from a manual structure. Read those as vendor reported results, but the pattern, lower CPA and less management overhead for the right profile, recurs too often across independent sources to dismiss as cherry picking.

ASC loses in the cases the success stories quietly skip. A store doing 10 purchases a week will never feed the learning phase and will thrash. A high consideration, high price product, think a 3000 dollar piece of furniture or a B2B equipment sale, has a purchase cycle measured in weeks, and ASC s preference for cheap near term conversions fits it badly. A brand chasing genuine new customer acquisition has to fight the platform s drift toward existing buyers, now via explicit exclusions rather than a budget cap. And rising costs are real: some Advantage+ segments saw new customer acquisition cost climb from 257 dollars in May 2024 to 528 dollars in May 2025 per practitioner reports, which is a brand level observation rather than a Meta figure but a clear warning against assuming ASC stays permanently cheap as more advertisers crowd into the same auction.

A practical setup that holds up

Put it together and the setup is unglamorous, which is the point. Clean the catalog first: one verified feed, accurate stock and prices, decent images, sensible categories. Confirm the pixel and Conversions API are both firing on purchase and add to cart, because ASC optimises only on what it can measure. Launch with 6 to 10 genuinely different creatives spanning static, video and catalog, not six recolours of one asset. Set a budget large enough to clear roughly 50 weekly conversions at your real CPA, not a generic blog minimum. Run one or two ASC campaigns, not ten. If new customers are the goal, build the existing customer exclusion deliberately rather than hoping the machine respects your intent. Then leave it alone long enough to exit learning before you judge it, and judge it on incrementality, not on the flattering ROAS the dashboard hands you for free.

ASC is not magic and it is not a scam. It is a very capable engine with a clear bias: it spends where conversions look cheapest, which serves established stores with strong signal and undercuts brands that need patient, incremental, new customer growth. The advertisers who win with it are not the ones who trust it blindly, nor the ones who refuse it on principle. They are the ones who feed it a clean catalog and varied creative, constrain it where its instincts conflict with their goals, and measure it against the truth rather than against its own marketing. Do that, and ASC becomes the strongest single tool in an e-commerce account. Skip the discipline, and you get a confident dashboard reporting a number that may not really be yours, and a budget quietly optimising for the wrong outcome.

Sources

Meta Q4 2024 earnings call, CFO Susan Li on Advantage+ Shopping passing a 20 billion dollar annual run rate (afaqs.com, storyboard18.com, adexchanger.com). Meta Business Help Center, About Advantage+ Sales Campaigns and About Meta Advantage+ Catalog Ads (facebook.com/business/help). Meta for Developers, Advantage+ Shopping Campaigns documentation (developers.facebook.com). Jon Loomer Digital, No More Existing Customer Budget Cap and 83 Changes to Meta Advertising in 2025 (jonloomer.com). Madgicx, Advantage+ Shopping Campaigns Farewell Customer Budget Cap (madgicx.com). Marpipe, ASC ROAS analysis 4.52 versus 3.70 (marpipe.com). Stella Hey Stella, incrementality study on ASC (stellaheystella.com). Triple Whale and Hunch, Meta Conversion Lift guides (triplewhale.com, hunchads.com). Foxwell Digital on the new ASC system (foxwelldigital.com). Agency benchmarks on creative volume, budget floors and CPA trends from Adverge Media, Alex Neiman, OptiFOX and David Tamachi, flagged as agency rather than Meta figures.