Market Entry

How to Run a Thailand Pilot Launch: Design, Metrics and Decision Rules

By ZhongTai Bridge Editorial  ·  Last reviewed October 2026  ·  5 min read

Photo: Unsplash

What should a pilot prove?

• Demand: will Thai buyers choose your product at a profitable price?
• Conversion: can your listings and content convert at an acceptable cost?
• Operations: can you deliver, handle chats and manage returns reliably?
• Economics: does contribution margin hold after real costs?
• Learning: which messages, channels and creators work best?

How do you scope the pilot?

• Range: three to five products at most, chosen by demand evidence.
• Channels: one marketplace, plus a LINE Official Account and a small content program.
• Budget: a ceiling that you can afford to lose, split between inventory, advertising and content.
• Duration: 90 days, with checkpoints at 30 and 60 days.
• Team: a single owner and a weekly working group.

A 90-day pilot at a glance

Process diagram. 1. Week 0: set: Thresholds, budget ceiling, owner, dashboard. 2. Weeks 1–4: foundations: Listings, tracking, chat, first reviews. 3. Weeks 5–8: test: Hooks, prices, creators, one campaign day. 4. Weeks 9–12: decide: Scale, adjust or stop against thresholds.

Which metrics should you track?

• Funnel: views, add-to-cart, orders and conversion rate.
• Efficiency: cost per order, advertising return and break-even gap.
• Profit: contribution margin per order and per product.
• Quality: rating, review count, cancellations and returns.
• Service: chat response time and resolution.
• Learning: share of orders by message, creator and campaign day.

How do you set thresholds before you start?

Define at least five numbers: minimum conversion rate, maximum cost per order, minimum rating, target contribution margin and a return-rate ceiling. Write the decision rule: for example, scale if at least four of five thresholds are met, adjust if two or three are met, and stop if fewer than two are met. Thresholds should come from your feasibility model, not from hope.

Reading the pilot at day 90

Two-by-two matrix. Horizontal axis: Contribution margin after ads vs target. Vertical axis: Conversion vs threshold. Adjust: Conversion strong, margin thin: change price, cost or pack. Scale: Both pass: expand in steps. Stop: Both fail: end the pilot and document why. Fix listing: Margin fine, conversion weak: improve pages and test again.

What is a good weekly rhythm?

1. Monday: review last week's dashboard and decide on changes.
2. Midweek: launch tests, such as new images, titles or creator videos.
3. Friday: capture learnings and update the test log.

Keep a single test log with hypothesis, change, result and decision. It becomes your playbook for scale.

What happens at day 90?

Hold a formal review with leadership. Present actual versus threshold, the learnings, the cost to date and the proposal for the next phase. If you decide to scale, scale in steps tied to evidence. If you adjust, define the specific changes and a shorter test. If you stop, document why and what you would need to see to restart.

Pilot budget template

• Inventory: enough for the expected 90-day demand plus a buffer.
• Advertising: a ceiling set from break-even ROAS and the learning you need.
• Content: photos, video and Thai copy.
• Creators: samples and fees for a small test group.
• Operations: fulfillment, packaging and customer service.
• Contingency: 10% to 15% for surprises.

Sample thresholds

For illustration only: conversion rate at least 2% on hero listings, cost per order below 60% of contribution before ads, rating of at least 4.7, return rate below 6%, and contribution margin after ads of at least 12%. Set your own numbers from the feasibility model.

What to document during the pilot

• Every change to listing, price and ad settings with date and reason.
• Customer questions and objections, in Thai and in translation.
• Competitor moves and promotions.
• Supply issues and delivery times.
• Learning that should change the plan.

This record is the main asset the pilot creates. It lets a new team member understand why decisions were made and shortens the next launch.

Pre-register your hypotheses and write a one-page readout

A pilot becomes persuasive when hypotheses are written before the data arrives. Write them as statements that can be wrong: for example, that the core product converts at least at the threshold with a rating of 4.7 or better, that cost per order stays below the contribution margin before ads, and that at least a third of orders come from non-campaign days. At the end, produce a one-page readout with five blocks: what we expected, what happened, why, what we will change and the decision requested. The discipline protects leadership from the story that fits the numbers after the fact, and gives the team a template to reuse in the next market.

A worked example: a 90-day pilot for a portable blender

The example is illustrative. The brand sets thresholds before launch: conversion of at least 2% on the hero listing, cost per order below 60% of contribution before ads, a rating of at least 4.7, a return rate under 6% and contribution margin after ads of at least 12%. In weeks one to four it publishes Thai listings, sets up chat scripts and collects first reviews. Weeks five to eight test two hooks, two price points and one creator batch. A campaign day in week nine lifts volume, but the team notes that margin fell on discounted units and sets a tighter discount limit.

At day 90, the hero listing meets four of five thresholds, with conversion just below target. The decision is to adjust: improve the main image and video, run a two-week test and scale only if conversion reaches the threshold. The pilot saved the brand from scaling on a campaign-day spike.

Pilot risks and how to contain them

• Stock-outs on a hero product during campaign days: set safety stock and pause ads when inventory is low.
• Single-person dependency: document processes and train a backup.
• Misleading metrics: separate campaign-day and ordinary-day results.
• Scope creep: agree what is out of scope, such as new channels, until day 90.

Executive questions on pilots

How big should a pilot be? Big enough to give a clear signal and small enough to lose without damage. For many brands that means three to five products and one main channel.

Should we pause during campaign days? No, but analyze them separately. Campaign days show peak potential, and ordinary days show baseline economics.

When can we change the thresholds? Only with a documented reason, and before the data arrive.

What do we deliver at the end? A readout with results versus thresholds, learnings and a recommendation to scale, adjust or stop.

Frequently asked questions