[{"data":1,"prerenderedAt":391},["ShallowReactive",2],{"guide-facebook-ad-campaign-structure":3},{"id":4,"title":5,"answer":6,"authorId":7,"body":8,"category":294,"ctaVariant":295,"dataset":294,"description":296,"examples":297,"extension":298,"faqs":299,"heroImage":324,"intro":325,"meta":326,"navigation":327,"path":328,"publishedAt":329,"seo":330,"sources":331,"stats":362,"stem":389,"updatedAt":329,"__hash__":390},"blog\u002Fblog\u002Ffacebook-ad-campaign-structure.md","Facebook Ad Campaign Structure (2027)","A Facebook ad account is built in three levels: the campaign holds the objective, the ad set holds the audience, placements and budget, and the ad holds the creative. For structure, fewer ad sets usually beat more, because each ad set needs about 50 optimization events a week to exit Meta's learning phase, so splitting one budget across many audiences starves them all and can make them bid against each other in the same auction. The modern default is to consolidate into one to a few broad ad sets and test variety at the ad level, which is exactly how Advantage+ campaigns (past a $20 billion annual run rate in Q4 2024) simplified the classic three-tier setup.","likit-sae-lee",{"type":9,"value":10,"toc":283},"minimark",[11,16,20,23,79,88,92,100,103,106,109,122,125,128,132,135,138,141,144,148,151,154,157,164,168,171,174,177,248,251,255,258,261,264,267,271,274,277,280],[12,13,15],"h2",{"id":14},"the-three-levels-and-which-decision-lives-where","The three levels, and which decision lives where",[17,18,19],"p",{},"Every Facebook ad account is built in the same three-level shape, and structure is really just the discipline of putting each decision on the level where it belongs. Get that right and Ads Manager stops feeling like a maze. Get it wrong and you end up fighting the delivery system instead of feeding it.",[17,21,22],{},"The top level is the campaign. It owns one thing above all: the objective, which is the instruction Meta optimizes toward (Sales, Leads, Traffic, Awareness, Engagement, App promotion). It can also hold a shared budget if you choose to set the budget here rather than lower down. Below it sits the ad set, which owns the delivery conditions: who can see the ad (the audience), where it can appear (placements), the schedule, and the optimization event the ad set is chasing. At the bottom sits the ad, which owns the message: the image or video, the primary text, the headline, and the call-to-action button. A campaign can hold many ad sets, and each ad set can hold many ads.",[24,25,26,42],"table",{},[27,28,29],"thead",{},[30,31,32,36,39],"tr",{},[33,34,35],"th",{},"Level",[33,37,38],{},"What it owns",[33,40,41],{},"The structural question it answers",[43,44,45,57,68],"tbody",{},[30,46,47,51,54],{},[48,49,50],"td",{},"Campaign",[48,52,53],{},"The objective, and the shared budget if you enable it",[48,55,56],{},"What are we trying to achieve, and is this a separate goal?",[30,58,59,62,65],{},[48,60,61],{},"Ad set",[48,63,64],{},"Audience, placements, schedule, optimization event, and (in a manual setup) the budget",[48,66,67],{},"Under what conditions should Meta deliver, and to whom?",[30,69,70,73,76],{},[48,71,72],{},"Ad",[48,74,75],{},"The creative: image or video, primary text, headline, button",[48,77,78],{},"What do we actually show a person?",[17,80,81,82,87],{},"The reason this matters for structure, and not just for finding a setting, is that Meta's learning and delivery happen at the ad-set level. The ad set is the unit that gathers conversion data and optimizes. So the number of ad sets you create, and how you carve up your audiences between them, is the single decision that most determines whether your account performs. This guide is about that: how many ad sets to run, how to keep them from cannibalizing each other, and how the Advantage+ era changed the answer. If you need the underlying launch mechanics first, the ",[83,84,86],"a",{"href":85},"\u002Fblog\u002Fhow-to-run-a-facebook-ad","step-by-step guide to running a Facebook ad"," walks through them end to end. Where the budget physically lives, at the campaign or ad-set level, is a related but separate question with its own trade-offs, so it is not re-litigated here.",[12,89,91],{"id":90},"how-many-ad-sets-per-campaign-and-the-math-behind-the-answer","How many ad sets per campaign, and the math behind the answer",[17,93,94,95,99],{},"Ask this question in any advertising forum and you will get a number: three, five, one. The honest answer is that there is no fixed count, but there is a hard constraint that tells you the ceiling for your account. It is ",[83,96,98],{"href":97},"\u002Fblog\u002Ffacebook-ad-learning-phase","the learning phase",".",[17,101,102],{},"When you launch or significantly edit an ad set, Meta's delivery system explores who responds before it settles into stable, optimized delivery. An ad set exits the learning phase once it gathers roughly 50 optimization events (a purchase, a lead, or whatever you optimize for) within about a week. Below that, the system does not have enough signal to optimize confidently, and delivery stays exploratory and expensive. An ad set that never reaches 50 events a week sits in learning indefinitely, spending your money at its least efficient.",[17,104,105],{},"That single number turns \"how many ad sets\" into arithmetic. Take your weekly budget, divide by your cost per result, and you get the events each ad set can produce. Then check whether that clears 50.",[17,107,108],{},"Here is a worked example in round numbers. Say your cost per purchase is about $20, and you have $2,000 a week to spend.",[110,111,112,116,119],"ul",{},[113,114,115],"li",{},"One ad set funded with the full $2,000 produces about 100 purchases a week. It clears the 50-event bar comfortably and optimizes well.",[113,117,118],{},"Two ad sets at $1,000 each produce about 50 purchases each. Both just reach the threshold, with no margin.",[113,120,121],{},"Five ad sets at $400 each produce about 20 purchases each. None of them ever exits the learning phase, so all five deliver unoptimized, and you have quietly built five permanently-learning ad sets out of a budget that could have powered one strong one.",[17,123,124],{},"At a $20 cost per result, an ad set needs roughly $1,000 a week just to reach 50 events. That is the real answer to \"how many ad sets can I run\": as many as your budget can fund past the threshold, and no more. A larger budget or a cheaper result buys you more ad sets; a smaller budget or a pricier result forces you to consolidate. The popular \"three to five ad sets\" heuristic is fine as a starting shape, but it is downstream of this math, not a substitute for it. Never create an ad set you cannot feed.",[17,126,127],{},"Meta's own ceilings are far higher than anything performance calls for. A standard ad account allows up to 50 ads in a single ad set and up to 5,000 ad sets and 5,000 campaigns per account (managed accounts go to 10,000 each and up to 50,000 ads). Those are storage limits, not recommendations, and they count both active and paused assets, so a sprawling account fills them faster than you would think. The number that governs good structure is 50 events a week, not 5,000 ad sets.",[12,129,131],{"id":130},"audience-fragmentation-why-more-ad-sets-means-less-data-per-ad-set","Audience fragmentation: why more ad sets means less data per ad set",[17,133,134],{},"Fragmentation is the failure mode the math above describes, and it is worth naming because it is so counterintuitive. The instinct, especially early on, is that more ad sets means more coverage and more chances to win. In practice, more ad sets built on the same budget means less data in each one, and data is the fuel the whole system runs on.",[17,136,137],{},"Splitting a budget across audience segments feels like precision. You separate men from women, or 25-34 from 35-44, or one interest from another, telling yourself you will \"see which one works.\" But you have not created more opportunity, you have diluted the opportunity you had. The same pool of likely buyers now has to prove itself five times over on a fifth of the budget each, and the delivery system gets a thin, noisy signal from every one of them. None reaches the volume it needs to optimize, so instead of one confident ad set you get several uncertain ones.",[17,139,140],{},"This is also why demographic slicing has fallen out of favour as a way to structure ad sets. Meta's delivery already reads age, gender, and interest signals inside a broad audience and finds the responsive people on its own. Pre-splitting those segments into separate ad sets does the system's job for it, badly, and pays the learning-phase tax on each slice. The segmentation that still earns its own ad set is the segmentation the algorithm cannot infer or should not cross: prospecting kept apart from retargeting, or a genuinely different offer, not a demographic cut of the same offer.",[17,142,143],{},"The tell that you are fragmented is an account full of ad sets stuck in \"Learning\" or \"Learning limited,\" each spending a little and none of them clearly winning. The fix is not to wait longer. It is to consolidate, to pour the budget that was feeding five starving ad sets into one or two that can actually eat.",[12,145,147],{"id":146},"audience-overlap-when-your-ad-sets-bid-against-each-other","Audience overlap: when your ad sets bid against each other",[17,149,150],{},"Fragmentation wastes data. Overlap wastes money, and it is a distinct problem. Overlap happens when two or more of your ad sets are eligible to show ads to the same person. Because those ad sets belong to you, they end up competing in the same auction for the same impression, which means you are, in effect, bidding against yourself and pushing up the price you pay.",[17,152,153],{},"Meta tries to protect you from the worst of this with auction filtering. When several of your ad sets could serve the same person in the same auction, the system generally enters only one of them (typically the one predicted to perform best) rather than letting them all fight. That prevents the crudest form of self-competition, but it does not make overlap free. The filtered-out ad sets still lose delivery opportunities, your spend concentrates unpredictably in whichever ad set wins the internal filter, and reporting gets muddier because impressions shuffle between ad sets in ways you did not design. Heavy overlap also feeds back into fragmentation, since the ad sets that keep losing the filter gather even less data.",[17,155,156],{},"You do not have to guess whether overlap is happening. Ads Manager has a built-in tool: open the Audiences section, select 2 to 5 saved audiences (each needs at least 1,000 people to compare), and choose Show Audience Overlap. It reports the percentage of people shared between each pair. A small overlap is normal and rarely worth acting on; a large one is a signal that those ad sets should either be merged or separated with exclusions.",[17,158,159,163],{},[83,160,162],{"href":161},"\u002Fblog\u002Ffacebook-ad-audience-exclusions","Audience exclusions"," are the surgical fix when you want to keep audiences distinct. If you run a prospecting ad set and a retargeting ad set, exclude your retargeting audience (and recent purchasers) from the prospecting one, so the two never chase the same person. That keeps each ad set aimed at a clean, non-overlapping slice and stops your own campaigns from inflating each other's costs. With blended Meta CPM sitting near $8.19 across 2025 per Gupta Media's tracker, the price of reach is already high enough without paying a self-inflicted premium on top of it.",[12,165,167],{"id":166},"consolidation-the-fix-for-both-problems-at-once","Consolidation: the fix for both problems at once",[17,169,170],{},"Consolidation is the move that solves fragmentation and overlap together, and it has become the default advice for a reason. The idea is simple: use fewer, broader ad sets, fund each one past the learning threshold, and move all of your variety down to the ad level where it belongs.",[17,172,173],{},"Concretely, that means collapsing several narrow audiences into one broader one, letting the delivery system find the responsive people inside it, and then loading that single ad set with a small set of genuinely different creatives so the ad set can optimize across them. You are not giving up testing. You are moving the test from the ad-set level (where it costs you a full learning phase per variant and creates overlap) to the ad level (where it costs you nothing extra and creates none). The audience stays consolidated; the messages compete.",[17,175,176],{},"The contrast is easiest to see side by side.",[24,178,179,191],{},[27,180,181],{},[30,182,183,185,188],{},[33,184],{},[33,186,187],{},"Fragmented structure",[33,189,190],{},"Consolidated structure",[43,192,193,204,215,226,237],{},[30,194,195,198,201],{},[48,196,197],{},"Ad sets",[48,199,200],{},"Many narrow ones (age, gender, interest slices)",[48,202,203],{},"One to a few broad ones",[30,205,206,209,212],{},[48,207,208],{},"Budget per ad set",[48,210,211],{},"Thin, often below the learning threshold",[48,213,214],{},"Enough to clear ~50 events a week",[30,216,217,220,223],{},[48,218,219],{},"Overlap",[48,221,222],{},"High; ad sets compete in the same auction",[48,224,225],{},"Low; audiences are broad or cleanly excluded",[30,227,228,231,234],{},[48,229,230],{},"Where you test",[48,232,233],{},"At the ad-set level (a learning phase per variant)",[48,235,236],{},"At the ad level (creatives compete inside one ad set)",[30,238,239,242,245],{},[48,240,241],{},"Typical result",[48,243,244],{},"Ad sets stuck in learning, uneven delivery",[48,246,247],{},"Ad sets exit learning, delivery stabilizes",[17,249,250],{},"The rule of thumb that falls out of this: change the ad set when you are changing who or where (a new audience type, a new placement strategy, prospecting versus retargeting). Change the ad when you are changing what you say (a new hook, a new format, a new angle). Most of the testing that actually moves results is the second kind, so most of your account's variety should live at the ad level, not spread across a dozen ad sets. Keep a handful of distinct creatives (commonly three to five) inside a well-funded ad set and let cost per result pick the winner.",[12,252,254],{"id":253},"how-advantage-changed-the-classic-three-tier-structure","How Advantage+ changed the classic three-tier structure",[17,256,257],{},"For years the craft of Facebook advertising was largely the craft of building ad sets: stacking interests, layering demographics, and hand-carving audiences into separate ad sets to test against each other. Advantage+ took most of that apart, and it did so at enormous scale. Meta reported its Advantage+ shopping campaigns passed a $20 billion annual run rate in the fourth quarter of 2024, up roughly 70% year over year, and by late 2025 said its broader AI-powered ad tools, including the Advantage+ suite, had surpassed a $60 billion annual run rate. By 2026 Advantage+ is the default campaign setup for the Sales, Leads, and App objectives.",[17,259,260],{},"The structural change is that Advantage+ collapses much of the ad-set layer up into the campaign. Instead of you building several ad sets with distinct hand-built audiences, an Advantage+ campaign leans toward one broad, largely automated audience and asks you to supply strong creative instead. Advantage+ audience does not delete the classic building blocks (Core audiences you define, Custom Audiences of people who know you, Lookalikes of people who resemble them). It sits over them, treating what you provide as a suggestion and a starting point rather than a hard boundary, then looking more widely for people likely to convert. You keep hard controls where you need them, such as location, minimum age, and explicit exclusions, but you stop micro-segmenting into separate ad sets.",[17,262,263],{},"Why does broad-and-automated now beat hand-split-and-narrow? Because the machinery underneath got dramatically better at finding the right person inside a big pool. Meta detailed its Andromeda retrieval engine in December 2024: it scans from tens of millions of candidate ads down to a few thousand relevant ones for each person, and Meta reported it lifted retrieval recall by 6% and ad quality on selected segments by 8%, with model capacity scaled up to 10,000 times the prior system. In November 2025 Meta added GEM, a ranking model that scores the retrieved ads, which it reported raised conversions by 5% on Instagram and 3% on Facebook Feed. When the system is that good at picking the responsive person out of a broad audience, splitting that audience into narrow ad sets yourself removes the very freedom the system needs. It is worth being precise about scope here, though: all of this is Meta's own delivery on Facebook and Instagram, not a cross-platform standard.",[17,265,266],{},"The practical translation for structure is a two-part rule. Push the audience decision up: let one broad ad set (or an Advantage+ campaign) do what you used to split across five. Push the testing decision down: put your effort into several strong, distinct creatives at the ad level, because that is now the lever the system reads and the one you still control. The classic three tiers are all still there. What changed is how much work each one is supposed to do.",[12,268,270],{"id":269},"a-structure-that-scales-without-fragmenting","A structure that scales without fragmenting",[17,272,273],{},"Put the pieces together and a sane default structure emerges, one that works whether you run a manual campaign or lean on Advantage+. Start with one campaign per genuine objective. If you sell products, that is usually a single Sales campaign; add a second campaign only when the goal is truly different (a Leads campaign, or a clearly separate prospecting-versus-retargeting split you want to report on independently). Resist the urge to open a new campaign for every product or idea. Products and ideas are ad-level decisions, not campaign-level ones.",[17,275,276],{},"Inside that campaign, run as few ad sets as your budget can feed past 50 events a week. For many accounts that is one broad ad set; for larger budgets it might be two or three, perhaps one broad prospecting ad set and one retargeting ad set with recent purchasers excluded so the two never overlap. Then fill each ad set with a small set of genuinely different creatives and let them compete. That is your testing engine, and it costs you no extra learning phases.",[17,278,279],{},"Scaling this structure is where discipline pays off, because the instinct that breaks it is the instinct to add. When something works, do not celebrate by spinning up eight new narrow ad sets; that re-fragments the account you just consolidated. Scale the winner by feeding it more budget (in measured steps, so a large jump does not reset the learning phase) or by duplicating the proven ad set and letting the copy find its own footing while the original keeps running. When something decays, the fix is almost never a new audience, it is a new creative, because fresh creative is the fastest lever on cost and the audience layer is already doing its job. That is the loop that compounds: read the result, make the next ad, feed it back into a structure that stays lean. Keeping the research, the creative, and the launch in one place (a platform like AdPlay.ai is built for exactly that loop) makes the cadence easier to sustain, but the structural discipline is what does the work.",[17,281,282],{},"The through-line is that good campaign structure is subtraction, not addition. The delivery system rewards concentration: enough budget and enough conversions in one place for it to learn, one clean audience per ad set so nothing competes with itself, and all your variety pushed down to the creative where it can be tested for free. Every extra ad set you add divides your data and risks overlapping the ones you already have. So the question to ask before creating anything new is not \"could this be its own ad set\" but \"can I feed it, and does it overlap what I already run.\" Most of the time the answer sends you back to the ad level, which is exactly where the modern platform wants your attention anyway.",{"title":284,"searchDepth":285,"depth":285,"links":286},"",2,[287,288,289,290,291,292,293],{"id":14,"depth":285,"text":15},{"id":90,"depth":285,"text":91},{"id":130,"depth":285,"text":131},{"id":146,"depth":285,"text":147},{"id":166,"depth":285,"text":167},{"id":253,"depth":285,"text":254},{"id":269,"depth":285,"text":270},null,"neutral","How to structure Facebook campaigns, ad sets and ads: how many ad sets per campaign, avoiding audience overlap, and consolidating in the Advantage+ era.",[],"md",[300,303,306,309,312,315,318,321],{"question":301,"answer":302},"How many ad sets should I have per campaign?","As few as your budget can feed with data, which for most accounts is one to a few. The binding constraint is not a rule but the learning phase: each ad set needs about 50 optimization events a week to exit it, so divide your weekly budget by your cost per result and check whether each ad set clears that bar. If a $2,000 weekly budget split across five audiences leaves each getting only 20 purchases, none of them ever stabilize. Three to five is a common heuristic, but the honest answer is: do not create an ad set you cannot afford to fund past 50 events.",{"question":304,"answer":305},"What is the difference between a campaign, an ad set, and an ad?","They are the three levels of the hierarchy, and each owns one type of decision. The campaign sets the objective (what Meta optimizes for) and, if you turn it on, the shared budget. The ad set sets who sees it (audience), where (placements), the schedule, and the optimization event. The ad holds the creative itself: the image or video, the primary text, the headline, and the button. Structurally, the campaign is the goal, the ad set is the delivery conditions, and the ad is the message.",{"question":307,"answer":308},"What is audience fragmentation on Facebook ads?","Fragmentation is splitting one budget and one pool of conversions across too many ad sets, so each one gets too little data to optimize. Advertisers assume more ad sets means more coverage, but the opposite happens: the same total budget divided by more ad sets means fewer events per ad set, and ad sets that never reach about 50 events a week sit permanently in the learning phase delivering unoptimized. The fix is consolidation, fewer ad sets each funded past the threshold.",{"question":310,"answer":311},"What is audience overlap and how do I check for it?","Overlap is when two or more of your ad sets can show ads to the same people, so they compete in the same auction and you effectively bid against yourself, which can raise costs. Meta's delivery system tries to prevent this with auction filtering, entering only one of your overlapping ad sets into a given auction rather than all of them. You can measure it in Ads Manager: open Audiences, select 2 to 5 saved audiences (each needs at least 1,000 people), and choose Show Audience Overlap to see the percentage shared between each pair. Exclusions and consolidation are the fixes.",{"question":313,"answer":314},"Should I use one campaign or several campaigns?","Use separate campaigns only when they pursue genuinely different objectives or need separate reporting, for example a Sales campaign and a Leads campaign, or prospecting kept apart from retargeting. Do not spin up a new campaign for every audience or creative idea; that is what ad sets and ads are for. Meta counts both active and paused assets toward account limits (up to 5,000 campaigns and 5,000 ad sets on a standard account), so a tidy structure is also easier to manage at scale.",{"question":316,"answer":317},"How did Advantage+ change Facebook campaign structure?","It collapsed much of the ad-set layer into the campaign. In a classic setup you hand-built several ad sets with distinct audiences; an Advantage+ campaign pushes you toward one broad, largely automated audience and asks you to supply strong creative instead. By 2026 Advantage+ is the default for the Sales, Leads and App objectives, and Meta reported its AI-powered ad tools passed a $60 billion annual run rate in late 2025. The practical shift: structure decisions move up to the campaign, and testing moves down to the ad level.",{"question":319,"answer":320},"How many ads can I put in one ad set?","Meta allows up to 50 ads per ad set, but that ceiling is not a target. Because the ad set optimizes across the ads inside it, a handful of genuinely different concepts (commonly three to five) gives the system room to find a winner without spreading impressions too thin. The ad level is where variety belongs: keep the audience consolidated in the ad set and let distinct creatives compete inside it.",{"question":322,"answer":323},"Does consolidating ad sets actually lower costs?","It can, in two ways. First, it concentrates conversions so ad sets exit the learning phase and deliver optimized rather than exploratory, and learning-phase spend is the least efficient spend. Second, it removes overlap, so your ad sets stop competing against each other and inflating the price you pay in the auction. With blended Meta CPM near $8.19 in 2025, paying more than you need per impression compounds quickly, so a leaner structure protects the budget as much as it simplifies the account.","\u002Fimages\u002Fblog\u002Ffacebook-ad-campaign-structure-hero.webp","You opened Ads Manager to launch one product and somehow ended up with nine ad sets, each with its own audience, each spending a trickle. Costs climb, delivery is uneven, and nothing seems to be learning. The problem is rarely the targeting itself. It is the shape of the account: how many campaigns, how many ad sets, and where the creative lives. This guide is about that architecture, the campaign, ad set and ad hierarchy, and how to organize it so the delivery system has enough data to actually work.",{},true,"\u002Fblog\u002Ffacebook-ad-campaign-structure","2027-02-19",{"title":5,"description":296},[332,336,339,342,345,349,352,356,359],{"label":333,"url":334,"year":335},"Meta Business Help Center, Campaign, Ad Set and Ad Limits Per Ad Account","https:\u002F\u002Fwww.facebook.com\u002Fbusiness\u002Fhelp\u002F652738434773716","2026",{"label":337,"url":338,"year":335},"Meta Business Help Center, About the Learning Phase","https:\u002F\u002Fwww.facebook.com\u002Fbusiness\u002Fhelp\u002F112167992830700",{"label":340,"url":341,"year":335},"Meta Business Help Center, Understand Auction Overlap","https:\u002F\u002Fwww.facebook.com\u002Fbusiness\u002Fhelp\u002F537699989762051",{"label":343,"url":344,"year":335},"Meta Business Help Center, About Overlapping Audiences","https:\u002F\u002Fwww.facebook.com\u002Fbusiness\u002Fhelp\u002F1679591828938781",{"label":346,"url":347,"year":348},"AdExchanger, Q4 Meta Earnings and Advantage+ Growth","https:\u002F\u002Fwww.adexchanger.com\u002Fplatforms\u002Fq4-meta-minted-money-and-improved-its-monetization\u002F","2025",{"label":350,"url":351,"year":348},"Marketing Dive, Meta's AI Bets Supercharge Marketing Efficiency and Costs","https:\u002F\u002Fwww.marketingdive.com\u002Fnews\u002Fmeta-ai-bets-supercharge-marketing-efficiency-costs\u002F804238\u002F",{"label":353,"url":354,"year":355},"Engineering at Meta, Andromeda Personalized Ads Retrieval Engine","https:\u002F\u002Fengineering.fb.com\u002F2024\u002F12\u002F02\u002Fproduction-engineering\u002Fmeta-andromeda-advantage-automation-next-gen-personalized-ads-retrieval-engine\u002F","2024",{"label":357,"url":358,"year":348},"Engineering at Meta, GEM Generative Ads Recommendation Model","https:\u002F\u002Fengineering.fb.com\u002F2025\u002F11\u002F10\u002Fml-applications\u002Fmetas-generative-ads-model-gem-the-central-brain-accelerating-ads-recommendation-ai-innovation\u002F",{"label":360,"url":361,"year":348},"Gupta Media, Social Media Ads Cost and CPM Tracker","https:\u002F\u002Fwww.guptamedia.com\u002Fsocial-media-ads-cost",[363,367,370,373,377,381,385],{"label":364,"value":365,"source":366},"Levels every Facebook ad account is built in",3,"Meta Business Help Center, 2026",{"label":368,"value":369,"source":366},"Maximum ads allowed in one ad set",50,{"label":371,"value":372,"source":366},"Optimization events an ad set needs to exit the learning phase","~50 per week",{"label":374,"value":375,"source":376},"Advantage+ shopping annual run rate, Q4 2024","$20B+","AdExchanger, 2025",{"label":378,"value":379,"source":380},"Meta AI-powered ad tools annual run rate, Q3 2025","$60B+","Marketing Dive, 2025",{"label":382,"value":383,"source":384},"Andromeda retrieval recall improvement","+6%","Engineering at Meta, 2024",{"label":386,"value":387,"source":388},"Blended Meta (Facebook and Instagram) CPM, full year","$8.19","Gupta Media, 2025","blog\u002Ffacebook-ad-campaign-structure","tQSdGmil1kFqkOV86ZFcJLHVCVze3vjfpeeXa4ra9GE",1786093699538]