TikTok ads for small business in 2026: is it worth it?

Every year somebody tells small business owners that TikTok is where the cheap attention is. Some of them move a Meta campaign across, run it for two weeks and conclude the platform does not work for their business. Usually the platform worked fine. What failed was a set of assumptions carried over from a different auction, a different creative format and a different floor on what it costs to start. Everything below is either printed on a platform's own help pages, with the date it was read, or it is arithmetic we are doing on top of those numbers and saying so. Where a widely repeated claim has no source we could find, it is named as such rather than repeated.

Who TikTok actually suits, and who it does not

The question is not whether your customers are on TikTok. At the scale the platform operates, they probably are. The question is whether you can pay the entry price twice: once in money, once in video. TikTok is the rare channel where the second currency is the binding one, and it is the one people forget to budget for. Five situations where the answer tends to be yes:

• –You already make video, or somebody in the business will happily be on camera. Not a studio and not a script doctor: a phone, a face, and a willingness to shoot again next week. TikTok's own creative guidance asks for a DIY rather than an overly polished style, so the quality bar is lower than it sounds and the frequency is higher.
• –What you sell is watchable. Food, clothing, hair, nails, renovation, fitness, pets, anything with a visible before and after. If seeing it happen is the argument, the format is doing half your persuading for you.
• –Your customers do not know your category exists yet. Nobody searches for a thing they have not heard of, and a feed that shows people what they did not ask for is the only kind of media that can create that demand.
• –You already have organic posts that earned attention without being paid for. This is the cheapest entry there is, because Spark Ads run an existing post as an ad, and TikTok states that views, comments, shares, likes and follows gained while it is boosted are attributed to the organic post itself.
• –You can leave a campaign alone for a week. TikTok documents that pausing and editing hold a campaign in its learning phase. An owner who cannot resist adjusting on day two is paying for exploration and then throwing the result away.

The other list is shorter and worth reading twice, because getting this side wrong costs a quarter rather than a week.

• –One video and no plan for the second. TikTok suggests three to five creatives per ad group and refreshing them when results decline. A single asset is not a campaign here, it is a test that ends the moment the asset tires.
• –A total budget under roughly $50 a day that is already producing results somewhere else. TikTok's published campaign floor is above $50 a day on its own. Moving half a small budget across leaves two channels, both underfed.
• –Sales that hinge on a specific searched-for term. Emergency plumbing, a named part, a comparison somebody makes deliberately. That demand exists at a moment TikTok has no way to sell you.
• –Nobody who can watch the account for the first fortnight. The opening weeks want attention without meddling, which is a specific and slightly awkward combination of available and disciplined.
• –A regulated or restricted category. Read TikTok's advertising policies before producing anything rather than after. The rules are not Meta's rules, and a rejection after a shoot is an expensive way to learn the difference.

One question sorts most readers faster than the two lists above: can you commit to shooting new video every week for the next three months? If the answer is no, the money is better spent making the channel that already works work harder. TikTok does not reward a campaign you set once, and no amount of budget substitutes for a supply of creative.

What it costs to get in, in numbers TikTok publishes

This is where TikTok is unusually helpful, and it is worth knowing before anyone quotes you a starting budget. TikTok prints its minimum budgets in dollars on its own help page, at both campaign and ad group level. Every figure in the table below came from About budgets for ads in TikTok Ads Manager, a page TikTok marks as last updated in July 2026, read on September 10, 2026.

Two things follow from that table and only one of them is TikTok's. Theirs first: once an ad group spends its allocated budget, delivery pauses automatically until you raise the budget or extend the schedule. The second is our arithmetic on top of their numbers, so treat it as ours. TikTok's creative guidance recommends three to five ad groups per campaign with three to five creatives in each. At just over $20 an ad group, that structure is $60 to $100 a day before you have learned anything, which is a different conversation from the $50 headline. You can start narrower than TikTok recommends. You should know that you are doing it, and why.

One cost arrives later and surprises people. TikTok requires mandatory business verification, an identity check confirming your company is real and legally registered, once total ad spend crosses a threshold. That threshold follows local requirements, so TikTok does not publish a single figure for it: you are notified when you cross it and given a deadline, and missing the deadline pauses your campaigns. Verifying at Business Center level covers every ad account beneath it. That page is marked last updated August 2026 and was read on September 10, 2026.

How the auction and the learning phase differ from Meta

TikTok sells impressions in an auction, the same as Meta, and describes ranking ads on bid price and relevance in much the same language. Where the two diverge is what you are billed for. TikTok Ads Manager offers four billing methods: CPM, oCPM, CPV and CPC. oCPM, priced per thousand impressions but delivered to people judged likely to act, is the default on the conversions and app install objectives, and it is the closest analogue to how a Meta lead or sales campaign charges you.

CPV is the one with no Meta equivalent, and it is worth understanding even if you never select it, because it tells you what the platform counts as watched. TikTok charges a CPV bid when the first of three things happens: somebody views the ad for at least six seconds, somebody using accelerated view watches an ad originally at least six seconds long, or somebody interacts with the ad inside the first six seconds. Six seconds is TikTok's own line between a scroll and a view, and it should shape the first six seconds of everything you shoot.

The learning phase is where the numbers genuinely differ, and where the comparison is most often reported wrongly. TikTok's learning phase page says volatility typically starts to decline after about 25 campaign results, or seven days from when the campaign enters the phase. Meta's documented benchmark is roughly 50 optimization events per ad set per week. Half the events looks like an easier platform, and it may well be, but the two statements are not the same shape: TikTok counts results at campaign level and describes the point where volatility starts to fall, while Meta counts events at ad set level and describes a threshold. Use each platform's number against its own account rather than against the other platform.

TikTok also publishes something Meta does not: explicit percentage limits on how hard you may raise a budget, and how often. Its budget page advises increases of no more than 40% per adjustment while a campaign is in the learning phase, no more than 30% once it has left, and no adjustments more often than every two days. Its Campaign Budget Optimization page adds that you should wait at least three days or 50 conversions before adjusting, keep each adjustment within 30% of the current daily budget, run three to five active ad groups, and that campaign budget optimization does not function at all when only one ad group in the campaign is active. Those are guidelines rather than hard blocks, but they are the platform's own, and they are more specific than the guidance Meta publishes on the same question.

TikTok and Meta, side by side

One boundary first, because it decides whether this is the table you need. If the question you are really asking is which kind of demand to buy, the kind that already exists and gets typed into a search box or the kind an ad has to create from nothing, that is a different comparison with a different answer and it is written up separately at aevin.dev/blog/facebook-ads-vs-google-ads. The table below assumes you have already decided to create demand rather than harvest it, and are choosing where to do that.

Every figure in that table came from a platform's own help page, read on September 10, 2026: TikTok's budget, bidding methods, learning phase, in-feed ad specification and Search Ads pages, and Meta's minimum budgets page. Two rows deserve a pause. The minimum budget row is the sharpest practical difference between the two products and it runs against expectation, because the newer platform is the one telling you the price of entry up front while the older one asks you to discover it from an alert in the interface. And the caption row is not cosmetic. On Meta the copy above the video does a great deal of the selling. On TikTok it is a fixed, unstyled line that cannot carry a link, which pushes the entire argument into the video itself.

Why the ad that works on Meta usually fails here

The most common way a small business tests TikTok is to take the Meta ad that works and upload it. It is a reasonable instinct and it usually produces a bad result, for reasons that have nothing to do with the audience. Here is what TikTok's in-feed ad specification actually asks for, on a page marked last updated June 2026 and read on September 10, 2026:

• –Vertical 9:16 at 540 by 960 pixels or larger is the recommended dimension. Horizontal at 16:9 and square at 1:1 are accepted, at 960 by 540 and 640 by 640, and accepted is not the same as native: a square in a vertical frame announces itself as an import before anyone has heard a word.
• –File requirements are generous and rarely the problem: .mp4, .mov, .mpeg, .3gp or .avi, up to ten minutes long, 500 MB or less, at a bitrate of 516 kbps or more.
• –Sound is not optional. TikTok's ad review best practices state plainly that clear audio is required for video ads. An ad built to be understood with the sound off is not merely weaker here, it is a policy problem.
• –Captions are shown in white in a uniform font you cannot customize, and they support no clickable links, no @ symbols and no hashtags. Everything your Meta primary text was doing has to move inside the video.
• –The safe zone is real, and TikTok publishes downloadable templates for it, with separate files for ads carrying anchors and for right-to-left regions. A re-cropped Meta creative reliably puts its own text underneath the interface.

Then there is timing, and this is where the format is genuinely unlike anything else. TikTok's creative guidance for performance ads recommends introducing your content proposition inside the first three seconds and prioritizing the hook in the first six, shooting at least 720p, using captions or text overlays at five to ten words per second, and featuring actual people, whether creators, employees or customers, rather than the product alone. It also recommends a DIY, not overly polished style so the ad sits comfortably beside the organic content around it. That last instruction is the hardest one for a business that has just paid for a beautifully finished Meta asset.

Volume is the other half of it. TikTok suggests three to five different creatives per ad group and three to five diversified ad groups per campaign, with big differences between creatives rather than small ones, and it recommends adding fresh creative into an existing ad group instead of building a new one so the ad group keeps what it has learned. Read that next to the ad group budget floor and it becomes a real commitment, which is why the honest question about TikTok is a production question before it is a media one.

The cheapest possible test avoids production altogether. Spark Ads run an organic post as an ad, your own or a creator's with their authorization, and TikTok states that the views, comments, shares, likes and follows gained while the post is boosted are attributed to the organic post. Spark Ads take .mp4 or .mov with no duration restriction, and the caption comes from the original post, up to four displayed lines. If you already have a post that earned attention without being paid for, that is the piece of evidence worth buying more of, and it costs nothing to shoot.

Counting the leads: pixel, Events API and attribution

TikTok's measurement stack looks familiar and behaves slightly differently, so it deserves setting up deliberately rather than by analogy. The TikTok Pixel is a snippet on your site that reports events back: ad and event information, a timestamp, IP address, user agent, cookies, and page metadata and button clicks. One detail in TikTok's own description catches people out. Third-party cookies are on by default with the pixel while first-party cookies are optional, and TikTok says performance is boosted when both are paired with advanced matching. That is the opposite default from what most people assume in 2026, and it is worth looking at your consent setup before switching anything on.

The Events API is TikTok's server-side connection, sending events from your own systems rather than from the browser. One sentence on that page changes the order of your setup work: TikTok states that pixel or Events API set-up is a prerequisite for the Web Conversions objective. A website conversion campaign is therefore not something you launch on Monday and instrument on Friday. TikTok recommends running both together for websites, for the same reason Meta does. Browsers and connectivity lose events that a server can still report.

Attribution is where TikTok's number and your own number will disagree, and neither of them is lying. Windows are set per ad group. Click-through can be 1, 7, 14 or 28 days, view-through can be off, 1 day or 7 days, and engaged view-through, for app campaigns, offers 1 and 7 days. A 7-day click and 1-day view setting will count conversions your CRM attributes somewhere else entirely, and a 28-day click window will count a great many more of them. Choose the setting once, write it next to the campaign, and compare weeks under the same window instead of trying to reconcile a platform total against a CRM total. Those two numbers are answering different questions.

If you have no website form worth sending traffic to, TikTok's Instant Form keeps the whole thing in the app, with form templates, instant lead capture and, on TikTok's own page, real-time sync into a CRM. The trade is identical to Meta's lead forms and worth naming out loud: a form that takes two taps produces more leads and colder ones. That is a fine trade when somebody follows up within the hour, and a poor one when the leads sit in an export for a week.

Why there is no average TikTok CPM in this article

Two claims travel further than any others on this subject: that the average CPM on TikTok is some specific dollar figure, and that engagement on TikTok is some specific percentage higher than on Meta. We could not trace either one to a TikTok page, and we are not going to repeat a number whose source is another blog repeating a number. There is a structural reason no such figure can exist for you anyway. An impression costs whatever it takes to outrank the other advertisers bidding for the same person at the same moment, in your category, in your city, in that particular week. TikTok publishes budget minimums, billing definitions, creative specifications and learning phase guidance, all of which are checkable and all of which are above. It does not publish an average price for your account, and neither does anybody else, however confidently the table is formatted.

When the right answer is to stay on Meta

This section matters more than the rest of the page if you are choosing a first platform, which is why it is here rather than buried at the end. Five situations where opening a TikTok account is the wrong next move:

1. 1.Meta is working and you have not finished with it. Retargeting the traffic you already earned, a second creative angle, a landing page that answers the ad it came from, a cheaper conversion event further up the funnel. Every one of those is cheaper than a second platform and none of them needs a camera.
2. 2.The budget cannot carry two floors. TikTok's published campaign floor is above $50 a day, and a Meta ad set still needs enough optimization events to leave its own learning phase. Splitting a fixed budget across both means neither gathers enough results to tell you anything, and you pay for two exploration phases while buying one set of conclusions.
3. 3.Your customers are searching rather than scrolling. If people type a specific term when they want what you sell, the money belongs where the typing happens. That is a decision about which kind of demand you are buying, and it is worked through properly in our piece on Facebook ads against Google ads rather than here.
4. 4.There is no supply of video and no realistic plan to create one. This is the single most common reason a TikTok test fails, and it is knowable before you spend a dollar. Three months of weekly shooting is the commitment. If that sentence produced a wince, believe the wince.
5. 5.You want to buy the audience without changing the ad. It is technically possible, since 16:9 and 1:1 are accepted formats, and it is the most reliable way to conclude that TikTok does not work. The platform is not rejecting your business. It is rejecting a rectangle.

If none of those apply and you want a real test rather than a gesture, the cheap version looks like this. Install the pixel and confirm it fires before you build anything. Run one campaign at the published floor with a single ad group, so the budget is not divided. Use a Spark Ad on an organic post that already earned attention, so you are testing distribution rather than production. Leave it alone for the first week, because pausing and editing hold it in the learning phase. Then judge it at about 25 results or seven days, against your own definition of a lead rather than the platform's, and decide with that in front of you.

Where Aevin fits, and what it will not do here

Aevin, our own product, connects Meta, Google Ads and TikTok in one panel at $97, $197 and $497 a month, with a landing page builder, image generation for creative, a CRM for the leads and the MAD training programme around it. On a connected TikTok account it does exactly three things: pause, resume, and change a daily budget, at campaign, ad group or ad level. Nothing happens on its own. Every change arrives as a before and after that you approve, and nothing reaches TikTok until you approve it. Three limits belong in the same paragraph as the promise. If the object carries a lifetime budget rather than a daily one, the change is declined with a note sending you into TikTok Ads Manager, because editing a lifetime budget means something different from editing a daily one. Campaign creation, audience edits and creative swaps are Meta-only today, and bid strategy changes are Google-only. And sending conversions back to a platform is a Meta feature here, so your TikTok Events API setup stays your own work. If what you want is a tool that builds TikTok campaigns for you, this is not that, and knowing it now is cheaper than finding out in month two.