DEV Community

Flowup Insights Team
Flowup Insights Team

Posted on

Run TikTok tests for a new product in Thailand in rounds, changing only one thing at a time

A practitioner note from the Flowup Insights Team, for brand and shop teams who are about to spend the first money on TikTok for a new product in Thailand. We also keep a longer, step-by-step guide on our site; this is the version you can use to design the rounds, not a condensed copy of that page.

A new product does not need a four-week content calendar. It needs rounds in which only one thing changes, so that when a clip "does not work" you know whether you tested the product, the angle, or the audience.

Write the question of the round before you brief anyone

If the room cannot finish the sentence "this round is for learning ___," you are not ready to hire creators or to boost a clip. Typical first questions are more modest than people like to admit:

  • Will a Thai viewer pause and understand what the product is?
  • Once they understand it, which reason to try it is even slightly clearer than the others?
  • Once one reason is clearer, can a slightly broader audience still understand it?

Those are three different rounds. If you brief five creators, three hooks, and two audiences in the same week, you will get activity and you will not get an answer.

Round one is about the product, not the talent list

In the first round, keep the audience tight and the brief almost blunt. The clip has to show the product so that it can be recognised later, and it has to say what it is in language a first-time viewer can repeat. Famous creators are a later tool. A small number of clear clips is enough to see whether people understand the object at all.

Do not judge this round only on orders. If nobody understands the product, adding budget will not help. If people understand it and still do not act, you have a different problem — the offer, the shop path, or the fact that the product is not yet easy to buy — and that problem does not belong to the hook.

Later rounds change one variable

In the second round, keep the same product presentation and change the angle: one reason to try it at a time. In the third, fund the angle that was actually clearer, then widen the audience. If you change the angle and the audience together, you will not know what you just bought.

The same rule applies to the shop. If the product name, the pack shot, and the purchase path are still inconsistent, pause the creator work and fix the shop first. A clean test on a broken product page only tells you that the page is broken.

We wrote the round-by-round sequence, including what to hold constant in every round, in Testing a new product on TikTok in Thailand, round by round. Open that guide when you are writing the brief and the stop rules for each round; use this note to keep the meeting from collapsing all three rounds into a single content burst.

Define "we have seen enough" before you spend

Before money leaves, write what would make you stop, continue, or widen. "The clips feel good" is not a stop rule. "Viewers cannot name the product after watching" is a stop rule for round one. "One angle is clearly easier to understand, and the shop path still works" is a reason to fund a third round, not a reason to book a month of mixed activity.

If the product is also meant to be found in 7-Eleven, add one more constraint: do not spend the test budget on areas or stories that point to a purchase path the viewer cannot complete. A test that generates interest in a store that does not stock the SKU is not a product test. It is a leak.

The point of rounds is not to look careful. It is to stop paying for a second and third change before you have read the first one.

Flowup Insights Team · a test-design note for new products on TikTok in Thailand

Top comments (1)

Collapse
 
clickpopai profile image
ClickPop Team •

"This round is for learning ___" is a great forcing function. One practical snag in round two: changing only the angle is hard when each angle is a separate creator shoot, because the person, the lighting and the pacing change too, so you are back to testing three things at once. What works better is holding the presentation constant and making the angle variants from the same base footage or product shot, so only the reason-to-try line and the first frame move. We do that in ClickPop when a fresh shoot is not practical, but even one creator recording three openings in a single session gets you most of the way. The 7-Eleven point is a good reminder that the purchase path is part of the test.