Whatever objective you select, the event you are charged for is the impression. The technical field that sets it accepts one value, and there is no other option to choose.
The help centre says the same thing in words: regardless of an ad set’s goal, advertisers are billed when impressions are delivered. What the objective changes is how your bid is read, and that distinction decides what a budget on this channel can and cannot promise.
The second thing worth settling early is who is actually here, because the usual objection and the usual rebuttal are both lazy. The British regulator measures this as the most youth-skewed audience of the top ten platforms, with two thirds of its adult visitors under 35. So the objection is largely right. The twist is that the youngest part of that audience, the one the sales pitch is built on, is exactly the part Europe and Britain have put beyond profile targeting.
This page covers what the channel bills, what budget makes its numbers readable, who you reach according to two regulators, what the rules on minors removed, which growing format a European advertiser cannot buy at all, and what its library shows.
The objective does not change the billing, it changes the bid
This is the point that distorts the most decisions on this channel, and it is written plainly in two separate documents.
The help centre states that you will be billed on impressions, and that regardless of an ad set’s objective, advertisers are charged when impressions are delivered.
The technical documentation goes further: the field that sets the billing event accepts one value only, impression. There is no other option to choose.
What the objective actually determines is how your bid is read, and the distinction matters to a budget.
On the impressions objective, your bid is a maximum cost per thousand: the channel will not exceed it.
On every other objective, whether clicks, installs, two or fifteen second views, purchases reported by the tag, or form fills, your bid becomes a cost goal. The documentation is explicit that this is not a maximum and can be exceeded.
The consequence is concrete, and the channel gives the worked example itself. With a bid of five currency units on a conversion objective, you can see ten units of impressions delivered in a day with no conversion occurring at all. Your cost per acquisition for that day is not five. It is infinite.
This is not a malfunction. It is how an optimised auction normally works and other channels share it. But it has to be understood before a budget is set, because it rules out reasoning in guaranteed cost per result.
One real exception, and it is worth naming. The rule above governs the auction. Reserved and fixed-fee products work differently: Snap Select is bought at a fixed cost per thousand, First Lens is a flat fee on a rate card priced by expected reach per country, and reach-and-frequency lenses are measured on actual plays rather than impressions. The company’s own quarterly filing describes its advertising contracts as based either on impressions delivered or on a fixed fee over a period. So the honest formulation is: in the auction you are billed per impression whatever the objective, and the exceptions are the products you cannot buy without a salesperson.
And the objective list is shorter than most guides say. The channel moved from twelve objectives to five: awareness and engagement, traffic, leads, app promotion, and sales. Three bid strategies sit underneath, automatic, lowest cost with a maximum bid, and target cost, with the minimum-return strategy withdrawn in February 2025. A guide listing a dozen objectives is describing the previous product.
Two useful details. On reserved formats, the cost per thousand is fixed at the moment of booking. And a video view here is defined as an impression plus at least two seconds of viewing, or a click. That is not a completion, and we have covered the point across platforms in what you buy when you buy video.
The floor is low, and that is the problem
The published minimum is genuinely small. Read on its own, it is misleading.
What is published. The minimum daily budget for an ad set is five currency units, the channel’s own pricing page states it, and the technical documentation confirms it as the minimum accepted value. At campaign level the documented minimum is twenty. The technical minimum bid is one cent.
What the channel recommends, on the same page: between twenty and fifty a day.
And here is what makes the floor inoperative. The channel states elsewhere that an ad set should be judged after 30 to 50 conversions aligned with the bid objective, and that a learning phase precedes stabilisation.
Put the two numbers side by side. At five a day on a purchase objective, reaching thirty conversions takes longer than any reasonable decision window. The technical floor lets you launch a campaign. It does not let you conclude anything from one.
A documentation inconsistency we flag rather than resolve. The length of the learning phase is given in three places with three values: four days or more on the pricing page, one to seven days in the help centre, at least fifteen days on the targeting page. Do not build a decision calendar on any one of them without checking it in your own account.
And an audience floor. An ad set cannot target fewer than a thousand people, and the channel advises against going above twenty million.
On the premium formats, no amount is published at all. Exclusive takeover placements, first-position augmented reality lenses and curated packages go through a salesperson on a rate card that is not public. We found no budget threshold for those formats, and none for reach-and-frequency buying either.
Who you reach, and here the objection is largely right
On the neighbouring channel we argued that the teen-network objection does not survive the measurement. Here it mostly does, and saying so is the only way the first argument stays worth anything.
The United Kingdom, measured by the regulator. Ofcom’s Online Nation 2025 report, on the Ipsos iris passive panel of UK online adults for May 2025, puts Snapchat at 11.1 million monthly visitors, 23% of online adults, up 10% year on year, with 7.3 million visitors on an average day.
And then the composition, which is the part that matters. The same report states that Snapchat has the most youth-oriented audience profile of the top ten platforms, with 66% of its users aged 18 to 34 and only 7% aged 55 or over. For comparison, in the same table TikTok has 39% in that 18 to 34 band.
So among British adults, two thirds of this audience is under 35. That is not an objection to be rebutted, it is a targeting fact to plan around.
Below adult age, the figure is starker still. Ofcom measures that among 8 to 14 year olds, Snapchat takes an average of 45 minutes a day, just behind YouTube at 48, and that the two together account for 52% of all time those children spend online. That is not your advertising audience, and the next section explains exactly why.
The United States, measured by survey. The Pew Research Center finds 26% of American adults saying they ever use Snapchat, in a survey of 5,511 adults with a margin of error of plus or minus 1.8 points. By age: 56% of 18 to 29 year olds, 33% of 30 to 49, 13% of 50 to 64, and 4% of those 65 and over. By gender, 30% of women against 21% of men.
One Pew page to avoid, because it is easy to land on: a chart giving Snapchat at 18% overall with 35% among 18 to 34 year olds is a median across 37 countries, not a United States figure.
What the channel says about itself, and why the numbers do not line up. Its commercial pages claim 971 million monthly active users, footnoted to internal data for the fourth quarter of 2025. Its own filed results give 946 million for that quarter and 971 million for the second quarter of 2026. The sales page is quoting a real figure against the wrong period. The same pages state that 50% of Snapchatters are over 25, sourced to internal data of 14 March 2024 with no method published.
And the only regional figure that is audited is about to disappear. In its third-quarter 2026 results the company reported 92 million daily actives in North America, down 6% year on year, and 98 million in Europe, down 1%, against 305 million in the rest of the world, up 9%. Note what North America means here: it includes Mexico, the Caribbean and Central America, so it is not a United States figure. Europe includes Turkey. No figure is published for the United States alone or for the United Kingdom.
From the first quarter of 2027, global monthly actives become the headline metric, global daily actives are replaced, and regional daily actives and regional revenue per user are discontinued. The North America and Europe split leaves the public record after the fourth quarter of 2026. If you want a regional baseline for this channel, take it now.
What Europe and Britain took out of your targeting
This is the most structural chapter for anyone planning on this channel, and it is documented with unusual precision. It is also where the United Kingdom lands on the opposite side of the line from where it landed on other channels.
Accounts aged 13 to 17 no longer receive personalised advertising in a named list of countries, and the documentation is precise about which: the channel does not serve teen accounts ads that are personalised based on things like certain historic content profiles or inferred interests, in the United Kingdom, Switzerland, Norway and the European Union. Brazil joined the same regime in March 2026.
Nor does the channel match your data to them. Personal data you upload, through the tag, the conversions interface or a customer list, is not matched against teen accounts in that zone.
What becomes unavailable for that age band: gender, predefined audiences, custom audiences, dynamic product ads, and targeting by app installation state. On the optimisation side, every conversion objective disappears, from clicks to installs through purchase events and form fills.
What remains: placement, age, language, location and device information. Context rather than profile.
And two automatic rules to know before you build anything. Ad sets targeting only people under 18 will not deliver. And a mixed ad set is automatically adjusted to a minimum age of 18.
A technical trap that follows from it. In the programmatic interface, an unset minimum age means thirteen. If you drive the account through the API, set 18 explicitly rather than leaving the field empty.
For adults, European users can refuse profiling-based advertising. They keep receiving ads, contextual ones: age, language, location, device, and the content currently being viewed.
And in the United States, none of this applies. The programmatic interface still lists 13 to 17 as a targetable age band. The channel’s own page on teenagers says additional restrictions apply to the content and targeting of ads for teens, and names only age-gated categories such as gambling and alcohol. We found no American equivalent of the European rule, and we state that as what it is: an absence in the pages we could open, rather than a confirmation from the channel.
Two further targeting capabilities are out of reach in this zone: advanced demographics, which the documentation declares unavailable in the European Union, the United Kingdom and Switzerland, and two of the four predefined audience families, which are reserved to selected advertisers in the United States.
Put this next to the other channels and the pattern breaks. On Microsoft Advertising the United Kingdom loses the professional targeting and gains the consent duty. On LinkedIn it keeps the fuller targeting and publishes nothing. Here it sits squarely with Europe. Each platform draws its own line, and none of them follows the political map. If you plan across these three markets, check the line per platform rather than assuming one applies.
The formats, and the fastest-growing one Europe cannot buy
Where ads appear. Single image or video in Stories and Spotlight. Story ads and collection ads. Commercials, which run inside the channel’s curated content and are non-skippable for up to six seconds. Augmented reality lenses in the camera, and filters on users’ own snaps. And sponsored snaps, which arrive in the chat tab.
Formats are not available on every objective. Commercials sit under awareness and engagement only. The leads objective takes image or video, story and collection, with no augmented reality. The others add the AR formats.
Several products are announced and not self-serve. Snap Select is described as available for select clients in Canada, France, Saudi Arabia, the United Kingdom and the United States. First Lens, First Commercial and First Story are booked through a partner team, and reach-and-frequency lenses require approval. None of these has a published price.
And here is the one that matters commercially. Sponsored snaps, the format that lands in the chat tab, is not served in the European Union or Norway. Meanwhile the company’s own quarterly filing attributes its advertising revenue growth to spending across native surfaces, naming sponsored snaps and promoted places specifically.
Put those two sentences together and you get the position plainly: the surface driving this channel’s growth is one a European advertiser cannot buy. An American or British advertiser can. That is the reverse of the pattern on most of the other channels in this series, where Europe loses targeting and gains disclosure. Here Europe loses the inventory itself.
The ad library, the most open of the series
Having now examined the libraries of LinkedIn, TikTok, Meta, Bing and Amazon, we can say it: technically, this is the most accessible of them.
What it covers, and the stated window understates it. The page says ads delivered in the European Union and other jurisdictions over the last twelve months, refreshed every twenty-four hours. We queried it. The country menu offers the 27 member states and Turkey, and the oldest ad we surfaced carried a start date of 23 May 2023, which is three and a half years rather than twelve months. This is the only library in the series whose published claim is more modest than what it actually holds, which is a pleasant change from the ones that overstate. Search works by exact paying-advertiser name only, so you need the legal entity, not the trading name.
What sets it apart is that its programmatic interface answers without authentication. We tested it. Where the other platforms require an approved application or an account, this one responds directly. For anyone building competitive monitoring, that difference is the whole cost of the exercise.
What it exposes, for each ad: the identifier, the ad account, the status, the creative type, the languages, the headline, the call to action, a media preview, the start date, the total impressions and their breakdown by country, the payer’s name, the review status and any rejection reasons. Alongside that sits a targeting object with the minimum age, the age brackets, the languages, whether each is an inclusion or an exclusion, and the devices.
What it does not expose, on our sample: no spend, no budget, no bid, no custom or lookalike audience, no interests, and no breakdown of reach by age band.
A reading nuance we flag without settling it. The regulation speaks of the number of recipients reached. The library publishes impressions. Those are not the same quantity, since one person can receive several impressions. We note the vocabulary gap; we do not rule on whether it complies.
And the limit that decides whether this is useful to you. We tested it directly: querying the interface for the United Kingdom returns an invalid-country error, and so does the United States. The phrase about other jurisdictions is not visible in what the tool accepts. So this library covers European delivery and nothing else, exactly as we found on TikTok’s, on LinkedIn’s and on Microsoft’s. Four libraries, one border.
What you can actually do with it. Pull your competitors’ live creatives, their impression volumes by country, and the name of the paying entity, which often differs from the trading name. And since the window is twelve months, a dated record beats an occasional look. We ran the comparative exercise on the Meta ad library.
The proceedings, and the difference between an investigation and a fine
Three jurisdictions, and in none of them has a penalty been imposed on this channel. That matters, because the coverage rarely separates the two.
The European Union. The channel was designated a very large online platform on 25 April 2023. On 26 March 2026 the Commission, with the Dutch coordinator, opened formal proceedings under the Digital Services Act, on the protection of minors and on risk assessment. The Commission’s own wording is that the service may have breached the regulation by exposing minors to grooming attempts, and it says the investigation focuses on five areas. Requests for information preceded it across five dates from November 2023 to October 2025.
No fine and no finding of non-compliance appears on the Commission’s supervision list. An open proceeding is not a conviction, and presenting it as one would be dishonest. The exposure, for scale, is up to 6% of worldwide turnover.
The United Kingdom. The company states in its own quarterly filing that it was designated a Category 1 service by Ofcom under the Online Safety Act in July 2026, and it records the statutory maximum as 18 million pounds or 10% of the previous year’s global revenues. A regulatory tracker reports binding information requests of 12 March 2026 and a report on the response in May. We flag that we could not open the regulator’s own pages, which refuse automated access, so for the British proceedings we are relying on the company’s filing and on a tracker rather than on a primary regulatory document.
The United States. The quarterly filing refers to a Federal Trade Commission consent decree without detail, and to state attorney general actions brought since 30 January 2024 under consumer protection laws, with Nevada and New Mexico trials listed for August 2027. Separately there is private litigation on social media addiction, with four bellwether cases settled during 2026. Private litigation and settlements are not regulatory penalties, and they should not be quoted as such.
What an advertiser should take from it. Not that the channel is about to vanish, and not that nothing is happening. The live question is whether the rules around minors tighten further, because the targeting those rules remove is precisely the targeting this channel’s audience profile would otherwise make valuable.
What does not hold up
“Snapchat is for teenagers.” Mostly true for the adult audience, and the honest version is more useful than either the claim or the rebuttal. Ofcom measures 66% of British adult visitors aged 18 to 34 and 7% aged 55 or over, the most youth-skewed profile in the top ten. Pew finds 56% of American 18 to 29 year olds using it against 13% of 50 to 64 year olds. So yes, it skews young. The twist is that the part of that audience you would most want to reach on this argument, the 13 to 17s, is exactly the part Europe and Britain have put beyond profile targeting. The sales pitch and the regulation are now pointing in opposite directions.
“Fifty per cent of Snapchatters are over 25.” That is the channel’s own figure, footnoted to internal data of 14 March 2024, about the United States, with no method published. It is not contradicted by anything we found, and it is not a measurement either.
“971 million monthly users.” The number is real and the label is wrong. The commercial page footnotes it to the fourth quarter of 2025, while the company’s own filed results give 946 million for that quarter and 971 million for the second quarter of 2026. A sales page quoting a correct figure against the wrong period is a small thing that tells you how much the other figures on it have been checked.
“North America means the United States.” Not in these filings. North America here includes Mexico, the Caribbean and Central America, and Europe includes Turkey. Neither figure can be read as a national one, and no United States or United Kingdom number is published at all.
“Ninety per cent of 13 to 24 year olds are on the platform.” An internal figure computed by dividing the channel’s own user count by a population estimate across more than twenty-five countries. Not a measurement of your market.
“Four point four trillion dollars of spending power.” A study commissioned by the channel in 2021, aggregating average discretionary spending country by country. A real method, aggregating something that is not what this audience will spend with you.
“Thirty-four per cent more likely to purchase” and “friend influence is four times stronger.” Commissioned studies, the second resting on a survey of 787 American respondents in 2020. Sales material measuring stated intention.
“Cost per thousand is 35% cheaper than Meta.” No independent audited comparison exists. The one tracker with a published method covers tens of millions of impressions without stating a country, an advertiser count or a sample, and its own page is dated before the data it reports. Not citable.
“The European investigation means a fine is coming.” It means proceedings were opened on 26 March 2026. No fine and no finding of non-compliance appears on the Commission’s list. The two are not the same thing.
“From five a day.” True as a technical floor, misleading as a working budget, because the channel itself asks for 30 to 50 conversions before an ad set can be judged.
How we work on this
We come to this channel in two situations: a company that has ruled it out on an idea of its audience, and a company spending on it believing it pays for results.
We start from the billed event. Everything here is paid per impression. That rules out reasoning in guaranteed cost per result, and it changes how a test budget is built.
We calculate the budget that makes the numbers readable. Not the five a day floor, but the amount that reaches the 30 to 50 conversions beyond which the channel itself says an ad set can be judged.
We put your target against the regulators’ measurement, looking at coverage and time spent separately. They do not say the same thing, and it is the second that determines how often you will actually be seen.
We check the minimum age is set explicitly, particularly where the account is driven through the programmatic interface, in which an empty field means thirteen.
And we pull your competitors’ library entries. On this channel the interface answers without authentication, which makes it the cheapest competitive monitoring of the series to set up, where your market is covered.
Where to start
Stop reasoning in cost per result here. You are buying impressions, whatever the objective says.
Size your budget from the conversion volume you need, not from the published floor.
Look at coverage and time spent as two separate questions. The first tells you whether your audience is present. The second tells you how often you will meet it.
Set the minimum age explicitly to 18 if you drive the account through the programmatic interface, where an empty field means thirteen.
And query the ad library for your competitors. It answers without authentication, which is rare, where your market is covered at all.