Est. reading time: 8 minutes
The most common mistake we see businesses make with TikTok advertising isn’t targeting, budgeting, or even creative quality. It’s treating TikTok like a second Meta account, porting over the campaigns that work on Facebook and Instagram with minimal changes and wondering why performance disappoints. It’s an understandable instinct, both platforms are paid social, both run auction delivery, both offer similar objectives, and on the surface the mechanics feel interchangeable. They’re not. TikTok is a fundamentally different environment with different user behavior, different content expectations, and a different algorithm, and the businesses that succeed there build for it specifically rather than copy-pasting a Meta playbook and hoping the platform adapts. Here’s what we’ve learned running TikTok campaigns across multiple verticals.
The content expectation is completely different
This is the foundation everything else stems from. Someone scrolling TikTok is watching content, short, entertaining, personality-driven, full-screen vertical video with sound on, no static images, no carousels, and the user swipes past anything that doesn’t earn attention in the first second. On Meta, there’s an established visual language for ads, people recognize them and tolerate polished, commercial-looking content, so a well-produced product photo with benefit copy performs because it fits the environment. That same ad on TikTok feels like a commercial interrupting a conversation, and users don’t just skip it, they resent it, because the platform’s culture has a low tolerance for anything corporate, over-produced, or disconnected from how real people communicate there.
TikTok’s own phrase is “make TikToks, not ads,” and it’s the most practical advice for the platform, not marketing speak. The ads that perform look and feel like organic content: shot on phones, real people talking to camera, native editing styles, text overlays, trending sounds, fast cuts, and value or entertainment delivered before any ask. Production quality isn’t irrelevant, it’s measured differently. A slightly shaky video of someone genuinely excited about a product beats a studio commercial almost every time, because the metric isn’t “does this look expensive,” it’s “does this feel real.”
The hook window is brutal
Meta gives you a few seconds to capture attention. TikTok gives you closer to one, the swipe is faster, the content volume higher, the thumb already in position. Across dozens of campaigns, the single biggest predictor of ad performance we’ve found is what happens in the first one to two seconds, not the offer, not the product, not the call to action. The opening hook.
That changes creative production. On Meta you might lead with the brand or a clean establishing visual; on TikTok you lead with disruption, a surprising statement, an unexpected visual, a question that stops the scroll. Hook structures we’ve seen work consistently: “I can’t believe nobody talks about this,” curiosity-triggering, strong for products solving problems people don’t know they have. “Here’s what happened when I tried [product],” a story opening that promises a payoff, effective for DTC and visible-results products. “Stop scrolling if you [specific audience identifier],” a direct callout that self-selects the audience. “POV: you just discovered [product],” a native format that reads organic rather than promotional.
The hook doesn’t need to be clever, it needs to buy the next three seconds. TikTok creative works as a series of small commitments, the first second earns the next three, the first three earn the next ten, and somewhere in those ten you’ve made your case, the craft we broke down fully in the 3-second hook.
The algorithm rewards watch time, not clicks
Meta optimizes heavily around whatever action you select. TikTok cares about that too, but carries an additional priority that shapes delivery: watch time and completion. The platform’s core product is keeping people watching, so it evaluates every piece of content, ads included, partly on how long people watch, rewarding completions and replays with more delivery at lower cost and penalizing early skips fast.
The practical implications start with length. Shorter isn’t automatically better, a 15-second ad watched to completion often beats a 30-second ad people abandon halfway, even with a stronger pitch, while a genuinely engaging 45-second ad that holds attention can beat both because total watch time is what’s being rewarded. We typically test in the 15-to-30-second range first, then extend to 45 or 60 for concepts with strong first-half retention, and the data settles the question quickly. And watch video completion rate in Ads Manager, the metric Meta-first advertisers habitually ignore: below roughly 15 to 20 percent on a 15-second ad, the creative isn’t holding attention and the algorithm will bury it regardless of targeting.
Spark Ads change the game for credibility
TikTok offers a format with no true Meta equivalent: Spark Ads, which boost existing organic posts, from your account or a creator’s with authorization, as paid ads that keep their real engagement and link back to the organic profile. This solves the authenticity problem directly, you’re promoting content that is organic rather than imitating it, users interact with it like any feed video, and the genuine engagement signals further improve delivery.
The strategic play we recommend: build a creator pipeline where creators produce organic content featuring your product, let organic performance identify the winners, then amplify those proven posts as Spark Ads. Organic becomes your creative testing layer, and paid budget only backs content that’s already demonstrated resonance, an approach that consistently outperforms traditional ad creative because the content was built for the platform, validated by the audience, and arrives carrying real social proof. It’s also cheaper than producing everything in-house, since you’re borrowing creator production styles that already work.
Target broader than you’re comfortable with
As on Meta, TikTok’s algorithm has gotten dramatically better at finding the right people inside broad audiences, and narrow targeting tends to underperform because it restricts the learning. We’ve tested it repeatedly, interest stacks of three to five layers against broad targeting with only age, gender, and geo, same creative, and broad wins more often than not once the pixel has conversion data to work with. The exception is early account life before the pixel knows anything, where light interest targeting helps the algorithm start, with the goal of graduating to broad within the first few weeks of spend.
Custom audiences and lookalikes keep their role, particularly retargeting, but prospecting should lean on broad targeting with the creative doing the filtering. A hook that calls out a specific pain point for a specific kind of person makes the right audience self-select, no micro-targeting settings required.
The budget threshold is real
TikTok demands more upfront commitment than most advertisers expect. The learning phase needs roughly 50 conversion events per ad group per week, and TikTok CPMs run higher than Meta’s in many verticals, so the arithmetic is blunt: at a $30 target CPA, exiting learning takes about $1,500 per week per ad group, and with two or three ad groups running, that’s $3,000 to $4,500 a week before scaling even starts.
We’re straightforward with clients about this. A total paid social budget of $2,000 a month means TikTok probably isn’t the right channel yet, and the money is better concentrated on Meta, where the learning phase is more forgiving. The businesses that should be testing TikTok have Meta performing well and want diversification, sell something visually demonstrable, can fund the algorithm’s learning properly, and can produce or source platform-native creative. TikTok can absolutely be a high-performing channel, but it isn’t a cheap testing ground, and going in underfunded is the reliable way to conclude “TikTok doesn’t work for us” when the test never had a fair chance.
Creative volume is the biggest constraint
More than targeting, bidding, or structure, the number one bottleneck on TikTok performance is creative volume, because the platform burns through creative faster than Meta, what works this week can fatigue by next. Our starting guidance is five to ten new creative assets every two weeks, which sounds like a lot until you define the term: on TikTok, assets means variations, different hooks on the same core message, different creators presenting the same product, different formats, talking head, demo, reaction, text-on-screen, and different lengths, not polished productions.
The production model that works combines in-house phone video, even rough, creator partnerships, micro-creators included, and systematic variation, taking a winning concept and cutting five versions with different hooks or pacing. The brands that struggle are almost always the ones that produce three polished videos, launch, and have nothing new for six weeks, by week three the ads are fatigued and by week six the account is running on fumes, the decay curve we mapped in creative fatigue on TikTok.
When TikTok makes sense and when it doesn’t
We don’t recommend TikTok for every client, and that’s worth saying plainly. It tends to work for DTC ecommerce with visually interesting products, beauty, fashion, home, fitness, food, for app installs and digital products where conversion happens quickly, and for brands targeting younger demographics, though the audience has broadened well past the Gen Z narrative. It tends to be harder for B2B, high-consideration purchases with long cycles, products that can’t be demonstrated visually, and businesses that can’t produce video at the volume the platform demands. None of these are absolute, we’ve seen exceptions both directions, but they’re a reasonable starting framework.
The core question is simple: can you produce authentic, platform-native video at a consistent volume, and can you fund the learning phase long enough to get real data? If yes, TikTok is worth testing. If either answer is no, the budget is better spent deepening your Meta strategy until those conditions change.
