You send lots of samples and get nothing back. Write down who gets one, then follow up on every one.

The Sample Policy Builder. The rules we use to approve sample requests, follow up with creators who have a sample and review what each sample returned, for UK brands on TikTok Shop.

Get the policy and Lab

Free. The Lab opens on this page as soon as you submit, with the policy as a PDF to keep.

We use your details to send you this resource and, if you tick the box, the newsletter. Privacy policy.

You're in

The Lab is open below. The policy is yours to keep as a PDF.

Open the Lab

Give every sample a decision, a follow-up and a review

A sample that gets no post usually went wrong in one of three places. It was approved against rules nobody had written down, nobody followed up once it landed, or nobody looked at what the samples returned.

We plan it as three jobs with an owner each. One written card decides who gets a sample, a fixed rhythm follows up with the ones who have it, and a monthly review changes the card.

Approve

One criteria card per product, versioned, with an approver and a backup. Every request is decided with a reason, expiring soonest first.

Request scorer and queue check.

Follow up

A brief on delivery, a personal nudge at day 7 and day 14, and a stop at day 21, so a quiet creator is a decision and not a surprise.

Chase planner.

Review

Each month, orders per sample, the share of sampled creators with no sale and the cost of the samples that returned nothing. Then one change to the card.

Payback review.

Where a sample programme leaks

  • The criteria live in chat, so they drift. In one client's account the GMV floor changed 8 times in 4 weeks.
  • Requests wait. Pending requests build up until the oldest are days old, and outreach carries on while the queue is full.
  • A strong request and a weak one get the same answer, because nobody wrote down what a strong one is.
  • The sample is delivered and nobody follows up until the monthly report.
  • Nobody reviews what the samples returned, so the criteria stay as they were.

In one client shop, requests per 1,000 invites fell from 16.8 in April to 1.7 in June 2026.

Approvals had dropped to 7 to 12% and pending requests had reached 300 to 580. When approvals had flowed, the brand reported almost 100 orders a day. We read it as outreach running ahead of approvals, which is why the plan sets a limit on the queue.

More samples did not mean more sales

In another client shop's data over 51 weeks, samples shipped, invites sent and videos posted barely tracked sales from one week to the next. GMV per sample stayed at about £20 whether sampling was high volume or selective.

So we judge sampling on what each sample returned and on how many sampled creators posted, and we change the criteria from that. The Lab's payback review does the sum for your own month.

One client shop, anonymised. A correlation of 0 means no link and 1 means a perfect one.

£20
GMV per sample, whether sampling was high volume or selective
0.04 to 0.10
correlation between samples shipped and sales, week to week

Six checks before you approve the next sample

Each one closes a gap that turns a sample into a cost with no post at the end of it. If one is open, fix it before the next batch.

  • Criteria cardOne written card per product, with a version, a date, an approver and a backup. Criteria never live only in chat.
  • StockThe product has stock and is not faulty or expiring, and a weekly cap per product keeps hero products and test products in balance.
  • QueueRequests are decided every day, expiring soonest first. The backup clears the queue if pending passes 50 or any request is 4 days old, and outreach pauses at a set level.
  • DeliveryThe community invite goes out, the brief is sent when the sample is delivered and the Spark authorisation is requested with the sample.
  • Follow-upA personal nudge at day 7 and day 14 with the winning format attached, and a stop at day 21.
  • Payback reviewOnce a month, orders per sample, creators with no sale and wasted sample cost, then one logged change to the card.

What happens after a sample is approved

A fixed rhythm so every creator holding a sample gets the same follow-up, and so a quiet creator reaches a decision.

WhenWhat happens
ApprovedCommunity invite sent. The brief goes out automatically on delivery. Spark authorisation is requested with the sample.
Day 7A personal message to anyone who has the product and has not posted, with the winning format attached.
Day 14A second nudge. A creator who has had the sample for 14 days with no post is in the "dropping off" group.
Day 21Stop chasing and log it. A creator whose content is mostly hauls, or whose click-through is under 1% across two posts, gets no more samples and goes on the do-not-sample list.
Targets75% or more of sampled creators posting within 21 days, and a median of under 7 days from delivery to first post.

Samples first does not suit every shop. A new shop that creators do not yet trust, and good creators who will not take a sample, may need paying upfront. Who gets a sample and who gets paid are separate decisions.

What's inside

An operating document and the tools to run it, built from how we approve and follow up samples for UK brands on TikTok Shop.

The Sample Policy

The criteria card, the queue rules, the follow-up rhythm and the monthly review. Five pages, PDF.

Request scorer

Check one sample request against the criteria card and get approve, hold or decline, with the reason code.

Queue check

See whether your pending requests need the backup approver or an outreach pause, and how long the queue takes to clear.

Chase planner

List the creators holding a sample and get who needs a nudge today, who is due next and who to stop chasing.

Payback review

Orders per sample, creators with no sale, wasted sample cost and what the samples paid back against your own margin.

8 templates and a prompt

The criteria card, the daily shortlist, the day 7 and day 14 nudges and more, plus a prompt that turns the policy into your own.

Get the free policy and Lab

The Sample Policy Lab

Write your criteria card first. Then use the scorer and the queue check daily, the chase planner weekly and the payback review monthly.

Download the policy (PDF)

Request scorer

For one sample request. Every criterion on the card has to pass. Enter the creator's figures and tick what is true. The example values are ours, with one criterion failing so you can see the output.

Leave blank if they have had no sample from you.

    Queue check

    For one product or one shop. It shows whether the pending requests need the backup approver today and when the queue would reach the level where outreach should pause. The example values are ours.

    Your own number. 150 is only an example.

    Chase planner

    One line per creator holding a sample: handle, delivery date and whether they have posted (yes or no). Dates can be 2026-11-10 or 10/11/2026. The example lines are filled in relative to today.

    Payback review

    For one product and one month. Use your own figures from the Affiliate Centre and your sample log. The example values are ours, not a client's, and the margin is an example.

    Templates

    Fill these in once and every template updates. Anything left in [brackets] is yours to complete.

    Make it yours

    Paste this into Claude or ChatGPT with the policy PDF attached. Fill in the brackets first. The more real numbers you give it, the better the policy.

    Prompt
    You are my sample policy designer for a UK brand selling on TikTok Shop. Use the attached Pear Sample Policy as your method. Do not invent figures. Where you need a number I have not given you, ask for it.
    
    My brand [brand, what we sell, price range]
    Products we sample [product, stock, cost per sample including postage]
    Our affiliate position [how many creators sell for us, how many sampled creators posted in the last 30 days]
    Sample requests [requests a day, requests we decide a day, pending now, oldest pending in days]
    Who approves, and the backup [names]
    What the samples returned last month [samples shipped, creators who posted, creators with a sale, orders, GMV, our margin on GMV]
    What we already require of creators [GMV floor, niche, content rules, anything else]
    
    Do this.
    1. Write a criteria card for each product, with every criterion pass or fail, a daily quota, a weekly cap, and decline reason codes.
    2. Tell me whether my queue needs the backup approver today and at what pending level I should pause new outreach.
    3. Set out the follow-up rhythm for creators holding a sample, with a nudge at day 7 and day 14 and a stop at day 21, and write both nudges in plain UK English.
    4. Work out sample-to-post rate, orders per sample, the share of sampled creators with no sale and wasted sample cost from last month's figures, and tell me which one change to make to the criteria card.
    5. Tell me where samples first may not suit my shop, and what I should decide about paying creators upfront instead.

    The policy is how we work. Pear runs it for UK brands.

    Pear is a UK TikTok Shop agency. We run sample approval, creator follow-up and the monthly payback review for brands on TikTok Shop, with a named approver and a backup on every account.

    If you would rather have it run for you, book 30 minutes with Tom, our founder.

    Book a call with Tom

    Before you download it

    Who is it for?

    Brands that send affiliate samples on TikTok Shop and see few posts or sales from them, and the people who approve sample requests or manage creators.

    Should we approve fewer requests?

    Approve against written criteria and decide every request daily. In one client shop, requests per 1,000 invites fell from 16.8 to 1.7 as approvals dropped to 7 to 12% and the queue grew past 300. The Lab's scorer applies your criteria and the queue check shows when to bring in the backup.

    What if creators take a sample and do not post?

    Message them at day 7 and day 14 with the winning format attached and stop at day 21. The chase planner shows who is due, and a creator who keeps taking samples and not posting counts against their post rate on the next request.

    What post rate should we expect?

    We aim for 75% or more of sampled creators posting within 21 days, and a median of under 7 days from delivery to first post. Your own numbers in the payback review show where you are.

    Give every sample a decision, a follow-up and a review.

    The policy, the Lab and 8 templates. Free.

    Get the policy and Lab