TikTok's US joint venture: what creators should do in 2026

A recommendation system is not a fixed object you can transfer intact. It is a set of learned weights and a stream of behavioural signal. Retraining on a narrower population changes what it surfaces, and nobody, including the people running it, can tell you in advance exactly how.

Forrester's analysis of the divestiture puts it plainly: the new venture must retrain the recommendation algorithm on US user data, and the jury is out on whether a US-only TikTok reproduces what made the original work. Its advice to marketers is to diversify to other channels, naming YouTube, Instagram, and podcasting.

For a creator, the practical translation is uncomfortable but simple. Your historical performance data described a system that no longer exists in the same form. A video that would have found 400,000 people in November might find 40,000 in August, and the reason may have nothing to do with the video.

The commercial side is moving too

The advertiser story has been reassuringly upbeat, which is its own signal.

At its NewFronts pitch, covered by Marketing Dive, TikTok told buyers the platform now operates with greater independence while preserving the user experience, and introduced formats including Logo Takeovers, a Prime Time placement that sequences spots across roughly fifteen minutes of scrolling, and Pulse Tastemakers for aligning brands with creators. TopView and TopFeed were merged into a single media buy.

If you sell TikTok work to brands, that list is a price sheet in disguise. Formats built around creator alignment mean brand budgets flowing toward creators with defined audiences rather than toward whoever went viral last week. Know the formats by name before a client asks.

What to do this quarter

Rebaseline instead of panicking. Pull your median views per post for the three months before January 2026 and the three most recent months. Compare medians, not your best day. If the median moved, that is a system change. If only your peaks moved, that is variance.

Stop renting your entire audience. The oldest advice in the creator economy is still the only one that survives a platform restructure. Move the relationship somewhere you control, which in practice means an email list or a newsletter, and treat social as discovery. How that connects to actual income is covered in how creators get paid.

Republish rather than rewrite. The same vertical video runs on Shorts and Reels with an hour of extra work. Doing that consistently for a quarter tells you which platform actually rewards your format now, with your data instead of somebody's prediction.

Be careful about the global feed assumption. If your audience is international, watch whether US and non-US performance start to diverge. Forrester flags exactly this question for creators and brands working across markets, and it has no answer yet.

Reprice with a floor. If you sell TikTok content to brands, sell deliverables and rights, not promised views. Anyone quoting guaranteed reach in 2026 is underwriting a system they cannot see inside.

How to tell a system change from a bad month

Creators are terrible at this, and the reason is that the platform gives you exactly the data that encourages a wrong conclusion.

Use three habits instead.

Work in medians, over at least twelve posts. One video carries so much variance that a single flop or a single hit tells you nothing. A median across a dozen posts moves only when the underlying distribution moves.

Separate reach from retention. If views fell but average watch time and completion held steady, distribution changed and your content did not. If retention fell too, the content is the problem and no algorithm story will fix it. This is the single most useful split you can make, and most people never make it.

Hold format constant while you measure. Changing your hook style, your length, and your posting time in the same month as a platform restructure guarantees you learn nothing. Pick one variable.

If medians dropped, retention held, and format stayed constant, you are looking at the system rather than at yourself. That conclusion has a specific implication: producing more of the same content harder is the wrong response. Testing different formats, and posting the same work where a different system can find an audience for it, is the right one.

If brands are your clients, not your audience

For the freelancers who make TikTok content for other people, the practical risk sits in the proposal rather than in the feed.

Three changes are worth making now.

Sell outputs and usage rights. Nine videos, two rounds of revisions, six months of paid usage, priced as production. Performance stays an outcome you influence rather than a number you guarantee.

Rewrite any bonus clause tied to views. Bonuses keyed to a metric the platform is actively retraining are a coin flip you agreed to in advance. If a client wants upside sharing, key it to something you both can audit, like leads or promo-code redemptions.

Quote cross-posting as standard. The same shoot, delivered cut for TikTok, Shorts, and Reels, is a better product in 2026 than a TikTok-only package, and it is the version a nervous marketing director wants to buy. Forrester is telling their bosses to diversify. Sell them the thing they are being told to buy.

There is a reporting change worth making too. If a campaign spans the January transition, split the results at that date and say why in one line. Clients who see a single blended number will read a decline as your work getting worse. Clients who see the split, with a note that the platform moved its recommendation system into a new company and retrained it, understand what they are looking at. Being the person who flagged that first is worth more than any individual video, because it is the thing that makes you sound like a partner rather than a supplier.

The same discipline applies to your own rate. If reach across the category has genuinely compressed, hourly-equivalent income falls unless the price of a deliverable rises. Have that conversation at renewal with the data in front of you, rather than absorbing the difference quietly for a year and calling it a slow patch.

Delivvo gives creators who take brand work one branded portal for the contract, the deliverables, and the invoice, so a sponsorship that spans three platforms still has one place where scope, approvals, and payment live. See how it works

What not to overreact to

Three things do not warrant a strategy change.

US users are not being asked to migrate to a new app, so the account-loss scenario people spent 2025 worrying about did not happen. Moderation authority now sits with the US entity, which means appeals and policy questions have a clearer owner than before. And the audience itself has not moved: Pew Research Center still finds TikTok used by 63 percent of US adults aged 18 to 29 and 44 percent of those aged 30 to 49.

One more thing not to overreact to: the ownership chart itself. Which investor holds 15 percent has no bearing on whether your next video finds an audience. Spend your attention on the retraining, because that is the only part of this that touches your work.

The platform is not dying. It is being rebuilt underneath you, quietly, while the interface stays the same. That is a reason to hold a second channel and a real audience list, not a reason to delete the app. If you are weighing where that second channel should be, the audience platform comparison is a reasonable starting point.