[{"data":1,"prerenderedAt":449},["ShallowReactive",2],{"guide-facebook-ads-product-launch-strategy":3},{"id":4,"title":5,"answer":6,"authorId":7,"body":8,"category":346,"ctaVariant":347,"dataset":346,"description":348,"examples":349,"extension":350,"faqs":351,"heroImage":376,"intro":377,"meta":378,"navigation":379,"path":380,"publishedAt":381,"seo":382,"sources":383,"stats":417,"stem":447,"updatedAt":381,"__hash__":448},"blog\u002Fblog\u002Ffacebook-ads-product-launch-strategy.md","Facebook Ads for a Product Launch (2027)","Launching a brand-new product on Facebook means starting with no pixel history and no lookalike seed, so the sequence differs from a normal campaign. Spend the weeks before launch building warm audiences from teaser content and a waitlist (a video engagement custom audience holds viewers for up to 365 days), then cold-start delivery on broad and interest targeting because you have no data to model yet. Optimize for cheaper upper-funnel actions first, let each ad set gather Meta's roughly 50 optimization events in 7 days before you judge it, and only build your first lookalike once launch traffic gives you a real seed of 1,000 or more buyers.","xanny-lee",{"type":9,"value":10,"toc":334},"minimark",[11,16,26,34,37,41,44,47,50,53,57,65,68,71,74,78,81,195,198,201,205,208,211,283,286,289,293,296,299,302,305,309,312,315,318,322,325,328,331],[12,13,15],"h2",{"id":14},"why-a-launch-breaks-the-normal-facebook-playbook","Why a launch breaks the normal Facebook playbook",[17,18,19,20,25],"p",{},"Most Facebook advice, and every ",[21,22,24],"a",{"href":23},"\u002Fblog\u002Fhow-to-run-a-facebook-ad","step-by-step guide to running a Facebook ad",", assumes you already have history. Retarget your site visitors, seed a lookalike off your best customers, optimize for purchases, let the algorithm lean on the conversion data you have been feeding it for months. A launch takes every one of those levers away. On day one you have no pixel history to optimize against, no customer list to upload, and nothing to model a lookalike from. You are asking a data-hungry system to perform with an empty plate.",[17,27,28,29,33],{},"That matters because of how delivery works now. Meta's system reads your creative and your conversion signal to decide who sees an ad, and it needs volume to do it. An ad set has to gather roughly 50 optimization events within about 7 days to exit ",[21,30,32],{"href":31},"\u002Fblog\u002Ffacebook-ad-learning-phase","the learning phase",", the window in which delivery is still unstable and your reported costs are mostly noise. A brand with a busy pixel clears that bar in a day. A launch with zero sales can sit in learning indefinitely, spending the whole time exploring and never settling. Optimize a fresh campaign straight for Purchase against broad targeting, with no history behind it, and you often burn the first two weeks watching Meta guess.",[17,35,36],{},"None of this reach is free, either. Gupta Media's tracker put the blended Meta CPM at about $8.19 across 2025, which is the price of putting an ad in front of a thousand people whether or not any of them was ever going to buy. You are prospecting into an enormous cold pool (Facebook ads could reach about 2.28 billion people worldwide as of January 2025, per DataReportal), and paying real money for every impression you spend flailing. The launch playbook, then, is not a different set of buttons in Ads Manager. It is a sequence built around one idea: manufacture the signal you are missing, deliberately and in phases, so that by the time you ask the system to optimize for sales it finally has something to learn from.",[12,38,40],{"id":39},"pre-launch-build-warm-audiences-before-you-have-anything-to-sell","Pre-launch: build warm audiences before you have anything to sell",[17,42,43],{},"The cheapest data you will ever collect is the data you gather before the product exists. You do not need something to sell in order to start building the audiences you will sell to. The weeks before launch are where a good launch is quietly won, because they convert an unknowable cold audience into a warm pool and a usable seed while the meter is still cheap.",[17,45,46],{},"Two asset types compound during this window. The first is a video engagement custom audience. Run a low-cost teaser: a short clip of the product in use, a founder talking about the problem it solves, a \"coming soon\" tease. Optimize a cheap video-views or engagement campaign around it, and every person who watches becomes someone you can retarget later. Meta lets you retain video viewers for up to 365 days and lets you set the threshold (say, everyone who watched at least fifteen seconds or 75% of the clip), so you are filtering for genuine interest, not accidental scrolls. That audience sits there, growing, ready for launch day.",[17,48,49],{},"The second is a waitlist. A simple landing page or a lead form collects emails from people who want to know when you go live. Those addresses upload as a customer-list custom audience, which Meta will build once the list clears its 100-person minimum. In practice you want more than the floor: upload several hundred raw contacts, because match rates lose some of them, and a list of 300 to 500 gives you a durable audience that survives the shrinkage. These waitlist people are the highest-intent humans in your entire launch. They asked to hear from you before you ever ran a sales ad.",[17,51,52],{},"The objective in this phase is never Sales. It is awareness, engagement, video views, or leads: cheap actions that fill the pool. Start three to six weeks out, keep the budget modest, and let the audiences accumulate. The payoff is twofold. On launch day you are not fully cold, because you have a warm pool to convert first at a far better cost than any prospecting can manage. And you have the beginnings of a seed for the lookalike you cannot build yet, forming from engagement and video views that pile up much faster than purchases ever will.",[12,54,56],{"id":55},"cold-start-prospecting-targeting-when-you-have-nothing-to-model","Cold-start prospecting: targeting when you have nothing to model",[17,58,59,60,64],{},"Warm audiences run out fast, so a launch still needs net-new reach, and here is where the missing history bites hardest. You cannot ",[21,61,63],{"href":62},"\u002Fblog\u002Ffacebook-lookalike-audiences","build a lookalike",", because a lookalike needs a source audience to imitate and Meta wants 1,000 to 50,000 people in that source. You have neither the buyers nor, on day one, the volume. So prospecting cold comes down to two moves.",[17,66,67],{},"The first is broad and Advantage+ targeting. Hand Meta as few constraints as you can and let the system find likely buyers from the signal it does have, which at launch is mostly your creative. Broad targeting paired with strong creative now routinely outperforms hand-built interest stacks, and it is the only approach that scales without a seed. It is also where Meta is steering everyone: Advantage+ shopping campaigns passed a reported $20 billion-plus annual run rate, and the Sales objective now defaults toward that automation. For a launch, broad is not a compromise you settle for until you have data. It is frequently the best cold-start structure there is.",[17,69,70],{},"The second is a small hedge of interest and behavior audiences: two to four tightly themed sets that describe the obvious buyer for your category. Not fifteen stacked interests, and not a replacement for broad, but a diversification, so you are not betting the entire launch on one delivery approach while it is still finding its feet. Keep the count low on purpose. Every extra ad set splits your launch budget thinner, and a launch budget spread across a dozen ad sets guarantees that none of them ever reaches the 50 events it needs to leave the learning phase.",[17,72,73],{},"Two rules keep the cold start clean. Exclude your warm audiences from every prospecting ad set, so you are paying cold CPMs only for genuinely new people and not re-buying the pool you already built for free. And run several distinct creative angles rather than one ad, because when you have no audience data, the creative is the signal the system reads to decide who to show it to. A single execution gives Meta almost nothing to work with. Three or four real angles give it room to find the message that lands.",[12,75,77],{"id":76},"the-week-by-week-launch-ramp","The week-by-week launch ramp",[17,79,80],{},"A launch is a funnel that fills from the top down, and the sequence below moves down that funnel only as fast as the data lets it. The table is the skeleton; the logic underneath it is what matters.",[82,83,84,106],"table",{},[85,86,87],"thead",{},[88,89,90,94,97,100,103],"tr",{},[91,92,93],"th",{},"Phase",[91,95,96],{},"Primary objective",[91,98,99],{},"Audience",[91,101,102],{},"What you are buying",[91,104,105],{},"What to watch",[107,108,109,127,144,161,178],"tbody",{},[88,110,111,115,118,121,124],{},[112,113,114],"td",{},"Weeks -6 to -1 (pre-launch)",[112,116,117],{},"Video views, engagement, leads",[112,119,120],{},"Broad, plus a waitlist landing page",[112,122,123],{},"A warm pool and an email list",[112,125,126],{},"Cost per view, list size, audience growth",[88,128,129,132,135,138,141],{},[112,130,131],{},"Launch week (week 0)",[112,133,134],{},"Retarget warm first, then broad prospecting",[112,136,137],{},"Warm pool, then broad and Advantage+",[112,139,140],{},"First site visitors and first sales",[112,142,143],{},"CPM, early CTR, whether events fire",[88,145,146,149,152,155,158],{},[112,147,148],{},"Weeks 1-2",[112,150,151],{},"Sales, optimized for a mid-funnel event",[112,153,154],{},"Broad plus 2-4 interests, warm excluded",[112,156,157],{},"Enough optimization events to learn",[112,159,160],{},"Events per week versus 50, cost per event",[88,162,163,166,169,172,175],{},[112,164,165],{},"Weeks 3-4",[112,167,168],{},"Sales, optimized for Purchase",[112,170,171],{},"Broad, plus retargeting site visitors",[112,173,174],{},"Consistent, trackable purchases",[112,176,177],{},"Learning status, cost per result trend",[88,179,180,183,186,189,192],{},[112,181,182],{},"Week 5 and beyond",[112,184,185],{},"Scale, plus first lookalike",[112,187,188],{},"Broad, 1% lookalike, retargeting",[112,190,191],{},"Efficient, stable growth",[112,193,194],{},"Cost per result while scaling, frequency",[17,196,197],{},"Launch week has a deliberate order: convert the warm pool before you spend a cent prospecting cold. Those pre-launch viewers and waitlist sign-ups are the people most likely to buy immediately, and their sales seed your pixel with the first real Purchase events, which every later phase depends on. Only once the warm pool is working do you open broad prospecting to bring in strangers.",[17,199,200],{},"Weeks one and two are about volume of the right kind. You almost certainly are not selling enough units yet to feed a Purchase-optimized ad set, so you optimize for something that happens more often (covered in detail in the next section) and let the ad set gather enough events to stabilize. Weeks three and four are when you step down to Purchase optimization, assuming sales have grown enough to support it, and lean on retargeting the site visitors the launch has now produced. Week five is the earliest point most launches can honestly scale and introduce a first lookalike, because that is roughly when the seed becomes large enough to be worth modeling. Push any of these phases earlier than the data allows and you are back to asking the system to optimize for an event it cannot find.",[12,202,204],{"id":203},"how-much-data-to-gather-before-you-scale","How much data to gather before you scale",[17,206,207],{},"The gate on the whole ramp is the learning phase. An ad set exits it after roughly 50 optimization events in about 7 days, and below that threshold the delivery is unstable and the cost figures you are staring at are mostly noise. For a launch, that single rule drives two decisions that people get wrong constantly.",[17,209,210],{},"First, pick an optimization event you can realistically hit 50 of per week. This is the most important lever you control at launch and the one most often ignored. If you sell ten units a week early on, optimizing for Purchase strands the ad set in learning forever, because it will never see 50 purchases in seven days. Optimize instead for an event that occurs several times per sale, so the ad set gets the volume it needs to stabilize, then step down toward Purchase once real sales support it.",[82,212,213,226],{},[85,214,215],{},[88,216,217,220,223],{},[91,218,219],{},"Optimization event",[91,221,222],{},"Roughly how often it happens",[91,224,225],{},"Good to optimize for when",[107,227,228,239,250,261,272],{},[88,229,230,233,236],{},[112,231,232],{},"Purchase",[112,234,235],{},"Once per sale (rarest)",[112,237,238],{},"You already do 50 or more sales a week",[88,240,241,244,247],{},[112,242,243],{},"Initiate Checkout",[112,245,246],{},"A few per sale",[112,248,249],{},"Sales are climbing but still under 50 a week",[88,251,252,255,258],{},[112,253,254],{},"Add to Cart",[112,256,257],{},"Several per sale",[112,259,260],{},"Early launch, sales still sparse",[88,262,263,266,269],{},[112,264,265],{},"Landing page views",[112,267,268],{},"Many per session",[112,270,271],{},"Week one, almost no funnel yet",[88,273,274,277,280],{},[112,275,276],{},"Video views, engagement",[112,278,279],{},"Cheapest, needs no product",[112,281,282],{},"Pre-launch warm-up",[17,284,285],{},"Second, budget to the event, not to a vanity number. Work backward from the 50-event floor. Suppose your launch-week cost per Add to Cart is about $6 and the ad set needs 50 of them in 7 days to leave learning. That is roughly 50 times $6, or about $300 a week, which is close to $43 a day for that one ad set. Starve it below that and it never learns, no matter how long you wait. Spread the same $43 across five hopeful ad sets and, again, none of them learns. Concentrate the budget so at least one ad set clears the phase, and expand only after it does.",[17,287,288],{},"The last piece is patience. Do not judge a launch ad on a day or two. Wait until the ad set is out of learning and you have a couple of weeks of stable cost per result behind you. This matters more now than it used to, because reported conversions run structurally thinner since Apple's App Tracking Transparency changes, which a University of Maryland study estimated cut measured ad click-throughs by about 37%. Your early numbers under-report reality, so give the launch room to prove itself before you kill an ad or crank a budget.",[12,290,292],{"id":291},"building-your-first-custom-and-lookalike-audiences-from-launch-traffic","Building your first custom and lookalike audiences from launch traffic",[17,294,295],{},"As launch traffic flows, the data assets you were missing finally start to exist, and the order you build them in matters. Do it top down, from the audiences that pay off immediately to the ones that need scale.",[17,297,298],{},"Website custom audiences come first, because the pixel starts logging the moment ads run. Create separate audiences for all visitors, for people who viewed the product page, for people who added to cart, and for purchasers. Each is a distinct level of intent, and together they power your retargeting straight away. Retargeting the warm pool and these fresh site visitors, with purchasers excluded, is usually the cheapest, highest-return spend of the entire launch, and it is the engine that funds your prospecting.",[17,300,301],{},"Next comes a customer-list custom audience built from your actual buyers, once you have at least 100 real emails and ideally a few hundred to survive match-rate loss. This is a higher-quality asset than a page visitor, because these people actually paid.",[17,303,304],{},"The lookalike comes last, and only when a source is genuinely large enough. Meta wants 1,000 to 50,000 people in the seed, and quality beats size: seed off your best customers (purchasers, or high average-order-value buyers) rather than all visitors, and start with a 1% lookalike, the closest match, before you widen it. Realistically you may not have 1,000 buyers for weeks, and this is exactly where the pre-launch warm-up earns its keep. The video engagement and waitlist audiences you started building before launch accumulate far faster than purchases, so they can seed a workable lookalike sooner than your buyer list can. Do not force a lookalike off 150 buyers just because the button is there; a thin seed produces a weak model that broad targeting would beat. Let broad and creative carry the prospecting until the seed is real.",[12,306,308],{"id":307},"phasing-from-awareness-into-conversion-as-signal-accumulates","Phasing from awareness into conversion as signal accumulates",[17,310,311],{},"The through-line of the whole launch is a funnel that fills over time, and each phase optimizes for the deepest action the data can currently support. Early on, in pre-launch and launch week, you buy cheap upper-funnel actions: views, engagement, traffic. These fill the warm pool and light up the pixel. In the middle weeks you shift to conversion optimization on an achievable event, an Add to Cart or an Initiate Checkout, because those happen often enough to feed the learning phase while true purchases are still thin. Late in the launch you move to Purchase optimization and lean hard on retargeting, because by then you have both the sales volume and the audiences to make it work.",[17,313,314],{},"Sequencing the retargeting is where a lot of launch profit hides. Your pre-launch warm audiences and your launch-week site visitors are the highest-converting people you have access to, so a dedicated retargeting ad set (with recent purchasers excluded, so you stop paying to sell to people who already bought) usually posts the best return of anything in the launch. Prospecting fills the top of the funnel with strangers; retargeting harvests the ones who leaned in. Run both, and let the retargeting profit subsidize the more expensive cold prospecting.",[17,316,317],{},"One measurement caution runs under all of it. With signal thinner than it used to be, the platform-attributed numbers you see in Ads Manager can under-count the sales your ads actually caused, especially early. Send your conversion events server-side through the Conversions API so the data is as durable as possible, and read blended results (total revenue against total spend over a couple of weeks) rather than reacting to a single attributed spike or dip. A launch judged hour by hour looks like chaos. Judged week by week, the trend is what tells you whether the sequence is working.",[12,319,321],{"id":320},"launch-pitfalls-that-waste-the-first-month","Launch pitfalls that waste the first month",[17,323,324],{},"Most failed launches do not fail on the creative. They fail on the sequence, and the same handful of mistakes shows up again and again.",[17,326,327],{},"Optimizing for Purchase on day one with zero sales is the classic one. The ad set never sees 50 purchases in a week, so it never leaves learning, and you spend the launch in permanent exploration. Start higher up the funnel and step down. Skipping the pre-launch window is the next: launching fully cold, with no warm pool and no seed, is the most expensive way to begin, because you pay full cold CPMs to build from nothing what a few weeks of cheap teaser spend would have handed you. Splitting the budget across too many ad sets kills launches quietly, since a launch budget divided a dozen ways can never push any single ad set past the learning threshold.",[17,329,330],{},"Then there are the self-inflicted resets. Editing the audience, the optimization event, or the creative, or jumping the budget sharply, all count as significant changes that restart the learning phase and waste the spend that got you partway. A launch is fragile; batch your edits and change one thing at a time. Forcing a lookalike off a tiny seed makes a weak model that underperforms the broad targeting it replaced. Judging results in days instead of weeks leads you to kill ads before the signal to evaluate them even exists. And shipping a single creative gives the system almost nothing to read at the exact moment it is relying on creative to do the targeting.",[17,332,333],{},"That last point is also the opportunity. When you have no audience data, creative variety is the fastest lever you have, so the teams that launch well are the ones who can research the angles already working in their category, generate several on-brand variations, launch them to Meta, and feed each read straight into the next round. A platform like AdPlay.ai keeps that loop in one place, but the sequence holds with any workflow. Warm the audience before you have a product, cold-start on broad while the creative does the targeting, optimize for events you can actually generate, and build your custom and lookalike audiences the moment the launch produces enough data to make them real. Do that in order, and the cold start stops being a handicap and becomes a schedule.",{"title":335,"searchDepth":336,"depth":336,"links":337},"",2,[338,339,340,341,342,343,344,345],{"id":14,"depth":336,"text":15},{"id":39,"depth":336,"text":40},{"id":55,"depth":336,"text":56},{"id":76,"depth":336,"text":77},{"id":203,"depth":336,"text":204},{"id":291,"depth":336,"text":292},{"id":307,"depth":336,"text":308},{"id":320,"depth":336,"text":321},null,"neutral","Launch a new product on Facebook with no pixel data: pre-launch warm-up, cold-start targeting, a week-by-week ramp, and when to build your first lookalike.",[],"md",[352,355,358,361,364,367,370,373],{"question":353,"answer":354},"How do I run Facebook ads for a product launch with no pixel data?","Treat the launch as a data-manufacturing exercise, not a normal campaign. Start weeks early with cheap teaser campaigns (video views, engagement, a waitlist) to build warm audiences you can retarget on launch day. On launch, prospect on broad and Advantage+ targeting plus a few interest sets, because you have no conversion history to seed a lookalike. Optimize for an event you can actually generate 50 of a week, then step down toward Purchase as real sales accumulate.",{"question":356,"answer":357},"Can I use lookalike audiences for a brand-new store or product?","Not on day one. A lookalike needs a source audience to model, and Meta recommends 1,000 to 50,000 people in that seed. A new store has none, so you prospect on broad and interest targeting first. Build the seed from launch traffic and, faster, from the video engagement and waitlist audiences you gather before launch, then create your first 1% lookalike once the seed is genuinely large enough to matter.",{"question":359,"answer":360},"How long before launch should I start warming up an audience?","Give yourself roughly three to six weeks. That is enough time to run low-cost teaser video and engagement campaigns, grow a video engagement custom audience (Meta retains viewers for up to 365 days), and collect a waitlist you can upload as a customer-list custom audience once it clears 100 people. The earlier you start, the larger the warm pool you can convert cheaply on launch day, and the sooner you have a seed for a lookalike.",{"question":362,"answer":363},"What objective should I use to launch a product with no conversion history?","Do not jump straight to Purchase optimization when you have zero sales, because the ad set will sit stuck in the learning phase. Before launch, use video views or engagement to build audiences. At launch, optimize for a mid-funnel event that happens often, like Add to Cart or Initiate Checkout, so the ad set can gather enough events to stabilize. Move to Purchase optimization once you are reliably generating dozens of sales a week.",{"question":365,"answer":366},"How much should I budget for a Facebook product launch?","Budget to the optimization event, not to a round number. An ad set needs about 50 optimization events in 7 days to exit the learning phase, so multiply 50 by your cost per event to find the weekly floor for one ad set. If your cost per Add to Cart is about 6 dollars, that is roughly 300 dollars a week, or about 43 dollars a day, for a single ad set to learn. Concentrate budget so at least one ad set clears the phase rather than starving five that never do.",{"question":368,"answer":369},"How much launch data do I need before I can scale?","Wait until the ad set has exited the learning phase (about 50 optimization events in 7 days) and you have a couple of weeks of stable cost per result, not a good day or two. Early launch numbers are noisy, and reported conversions run thinner than they used to since Apple's tracking changes, so read a multi-week trend. Once cost per result holds, scale in small steps rather than doubling budget, which would reset learning.",{"question":371,"answer":372},"How do I build my first custom audience from launch traffic?","As soon as ads run, the pixel starts logging visitors. Create website custom audiences for all visitors, viewers of your product page, people who added to cart, and purchasers, each as its own audience. Upload your actual buyer emails as a customer-list custom audience once you have at least 100. These power your retargeting immediately and become the seed for your first lookalike as the numbers grow.",{"question":374,"answer":375},"Should I use Advantage+ for a new product launch?","Broad and Advantage+ targeting is often the right cold-start move, because with no audience data the creative is the main signal Meta reads, and broad targeting paired with several strong creatives scales without a seed. Advantage+ shopping campaigns now run at a reported 20 billion dollar-plus annual run rate and default the Sales objective toward automation. Feed it multiple creative angles, exclude your warm audiences from prospecting, and let it find buyers.","\u002Fimages\u002Fblog\u002Ffacebook-ads-product-launch-strategy-hero.webp","You are about to launch a product nobody has bought yet, on an ad platform that runs on data you do not have. No pixel history, no customer list to model, no lookalike seed. Meta's delivery system is happiest when you hand it a pile of past conversions to learn from, and a launch is the one moment you have none of them. This playbook is the cold-start sequence: how to manufacture signal before launch day, prospect without a lookalike, and phase from awareness into conversion as the data finally starts to accumulate.",{},true,"\u002Fblog\u002Ffacebook-ads-product-launch-strategy","2027-05-26",{"title":5,"description":348},[384,388,391,394,397,400,404,407,410,413],{"label":385,"url":386,"year":387},"Meta Business Help Center, About the Learning Phase","https:\u002F\u002Fwww.facebook.com\u002Fbusiness\u002Fhelp\u002F112167992830700","2026",{"label":389,"url":390,"year":387},"Meta Business Help Center, About Lookalike Audiences","https:\u002F\u002Fwww.facebook.com\u002Fbusiness\u002Fhelp\u002F164749007013531",{"label":392,"url":393,"year":387},"Meta Business Help Center, Create a Customer List Custom Audience","https:\u002F\u002Fwww.facebook.com\u002Fbusiness\u002Fhelp\u002F170456843145568",{"label":395,"url":396,"year":387},"Meta Business Help Center, Create a Video Engagement Custom Audience","https:\u002F\u002Fwww.facebook.com\u002Fbusiness\u002Fhelp\u002F1099865760056389",{"label":398,"url":399,"year":387},"Meta Transparency Center, Advertising Standards and Ad Review","https:\u002F\u002Ftransparency.meta.com\u002Fpolicies\u002Fad-standards\u002F",{"label":401,"url":402,"year":403},"Gupta Media, Social Media Ads Cost and CPM Tracker","https:\u002F\u002Fwww.guptamedia.com\u002Fsocial-media-ads-cost","2025",{"label":405,"url":406,"year":403},"WordStream \u002F LocaliQ, Facebook Ads Benchmarks 2025","https:\u002F\u002Fwww.wordstream.com\u002Fblog\u002Ffacebook-ads-benchmarks-2025",{"label":408,"url":409,"year":403},"DataReportal, Essential Facebook Statistics and Trends","https:\u002F\u002Fdatareportal.com\u002Fessential-facebook-stats",{"label":411,"url":412,"year":403},"AdExchanger, Meta Q4 Earnings and Advantage+ Growth","https:\u002F\u002Fwww.adexchanger.com\u002Fplatforms\u002Fq4-meta-minted-money-and-improved-its-monetization\u002F",{"label":414,"url":415,"year":416},"University of Maryland, Robert H. Smith School of Business, Small Businesses Take Big Hit from Apple's Privacy Regulation","https:\u002F\u002Fwww.rhsmith.umd.edu\u002Fresearch\u002Fsmall-businesses-take-big-hit-apples-privacy-regulation","2024",[418,422,425,428,431,435,439,443],{"label":419,"value":420,"source":421},"Optimization events an ad set needs to exit Meta's learning phase","~50 in 7 days","Meta, 2026",{"label":423,"value":424,"source":421},"Recommended lookalike audience source size","1,000-50,000",{"label":426,"value":427,"source":421},"Minimum size for a customer-list custom audience","100 people",{"label":429,"value":430,"source":421},"Maximum retention window for a video engagement custom audience","365 days",{"label":432,"value":433,"source":434},"Blended Meta (Facebook and Instagram) CPM, full year","$8.19","Gupta Media, 2025",{"label":436,"value":437,"source":438},"All-industry Facebook CPC, Leads objective","$1.92","WordStream, 2025",{"label":440,"value":441,"source":442},"Annual run rate of Meta's Advantage+ automated ads","$20B+","AdExchanger, 2025",{"label":444,"value":445,"source":446},"People Facebook ads could reach worldwide, January 2025","2.28 billion","DataReportal, 2025","blog\u002Ffacebook-ads-product-launch-strategy","5pDuLXWa90VJhasRCAQzrrQx_R8CIH59NLWgqu8CVmQ",1786093700389]