Why a sample can win the first order
A first order from an unknown store is a risk the shopper carries alone. They cannot taste the coffee, smell the candle or feel the fabric, and a full-size product that turns out wrong is money lost or a return to arrange. A sample moves that risk. The shopper pays little or nothing to find out, and the store gets the chance to be judged on the product rather than on the page.
That makes sampling an acquisition programme, not a perk. It sits beside the other ways a store lowers the risk of a first purchase: the entry products that the guide to choosing which products to promote defines as lines that bring in first orders and are judged on what follows, and the starter kits that the average order value levers guide treats as a starting point for new customers. A sample is the smallest version of the same idea: the cheapest possible way to find out.
The idea works only where the product is the reason people come back. A sample of something forgettable produces forgettable results, and a sample of something that needs weeks of use to show its value tells the shopper nothing. The first decision is whether the product is the kind a sample can sell.
Kinds of sample and when each fits
Five forms are common, and they suit different products and different stages of the relationship.
| Kind of sample | What the shopper gets | Suits | Watch for |
|---|---|---|---|
| Paid trial size | A small version of one product at a small price | Consumables with one clear question: does it suit me? | A trial price so low it becomes the product people buy |
| Sample pack | Several small variants in one pack | Ranges where choosing the variant is the hard part: shades, roasts, scents | Packs that answer the question so well the shopper buys the full size elsewhere |
| Free sample against delivery cost | The sample free, the shopper pays postage | Low-cost consumables where the delivery charge filters intent | Requests from collectors; follow-up that outruns the consent given |
| Sample in another order | A small item added to a paid order | Introducing a second line to existing buyers | Samples that do not match what the customer bought |
| Starter set | A first full-use kit at a set price | Products with a routine or several parts: skincare steps, brewing kit, grooming | A starter set priced to lose money with no plan for the second order |
The products that suit sampling share a question that only use can answer: taste, scent, texture, skin or hair response, fit or comfort. Coffee, tea, skincare, fragrance, haircare, pet food, fabric swatches and some foods fit. The products that do not suit it are those whose first use is the whole value (a one-off tool, a book, a gift), those that cannot be meaningfully shrunk, and those where a small dose shows nothing. For those, a clear page, honest reviews and an easy returns promise do the job a sample would.
Before building anything, watch a handful of shoppers try to choose and order a sample on the store's current pages. A short session of the kind the usability testing guide describes often shows that the sample is hard to find, or that the credit against a full-size order is not understood.
Pricing a sample so it finds buyers, not collectors
A sample's price is a filter before it is a revenue line. Free attracts the most requests and the weakest intent. A charge, even the delivery cost alone, removes most of the people who would never have bought the full size, and a modest price removes more. The store is trying to reach shoppers who are seriously considering the product and want to remove one doubt; pricing is how it finds them.
Three mechanics help:
- Credit the sample price against a first full-size order. The shopper pays for the sample and gets the same amount back on the full size within a stated period. This keeps the filter and removes the feeling of paying twice. The credit is a mechanic; how large it is comes from the store's own margin.
- Limit samples per customer. One sample pack per household, or one per product, stops the programme becoming a cheap supply of small sizes.
- Keep the trial size from replacing the full size. If the trial size costs less per unit than the full size, some customers will buy trial sizes for ever. Price it so the full size is the better buy once the doubt is gone.
Work out what each sample costs to make, pack and post before setting the price, and how many sample buyers have to go on to a full-size order for the programme to pay. The break-even calculator makes that arithmetic explicit; the inputs are the store's own costs and margin, and the result is a target to read the programme against, not a forecast.
A sample with a deep price cut is also a promotion. The promotion planning guide lists samples among the alternatives to a discount for winning new customers, and its rules on eligibility, honest prices and frequency apply to sample offers too.
From sample to full-size order
A sample has done half its job when it arrives. The other half is the path to the first full-size order, and that path is designed, not hoped for.
Start inside the parcel. A sample that arrives with a card explaining how to use it, what to notice, and how to order the full size with the credit applied gives the shopper everything in one place. Then the messages: the post-purchase email flow tutorial builds the help and next-purchase messages for any first order, and a sample order deserves its own branch, timed to when the sample will have been used rather than when it was dispatched.
What those messages may contain depends on what the shopper agreed to. A sample request is a request for that sample, not consent to marketing. The site's list building and consent guide treats a download as separate from a newsletter subscription, and the back-in-stock tutorial treats a product alert the same way, following the direction of the ICO's direct marketing guidance. The sample request form therefore offers marketing as a separate, unticked choice. Messages the request covers, such as dispatch and how to use the sample, go to everyone; promotional follow-up goes only to those who chose it. Whether a given follow-up counts as marketing is a question to settle with whoever owns compliance; nothing here is legal advice.
Once a sample buyer places a full-size order, they are a customer, and the repeat purchase and loyalty guide takes over: the second full-size order matters more than the first.
Samples in other orders and through partners
Samples do not have to be requested. A small item added to a paid order introduces an existing customer to a line they have not tried, at the cost of the sample alone. The repeat purchase and loyalty guide names a free sample with orders among the low-cost perks a store can offer. The rule is relevance: a sample chosen from what the customer bought (a different roast for a coffee buyer, a matching step for a skincare buyer) reads as care; a random leftover reads as clearing stock. The card that travels with it is an order insert, and the packaging and order inserts guide covers what an insert should say and how to track what it brings in.
Samples can also travel through partners whose customers overlap with the store's: a complementary brand's orders, a subscription box, an event bag. The store gives up control of the moment, so the sample needs everything it needs to work on its own: what it is, how to use it, and a way to order the full size with an identifier the store can track. Treat the partner as a separate source when reading results, because its customers will behave differently from the store's own visitors.
Samples for reviewers and creators
Sending samples to reviewers and creators is one of the oldest ways to get a new product talked about, and the rules on it are clear. Free product is a material connection between the brand and the person talking about it, and the FTC's Endorsement Guides Q&A expects such a connection to be disclosed clearly. The disclosure appears in the review or post itself, not in a profile or a footnote.
For a store, three practices follow. Creator agreements carry the disclosure requirement as a term, as the influencer contracts and disclosure guide sets out, and the brand checks what is published. Reviews on the store's own site from people who received free product are labelled as such, which the reviews and social proof guide covers in its section on incentives and fake reviews. And a free sample is never conditional on a positive opinion: it is sent whatever the recipient goes on to say. For a new product, the tutorial on launching a new product in an online store places this seeding before launch day, so honest early opinions exist when paid traffic arrives. Gifting to creators as a standing programme, with an ask before each send and a gifting note that requests a clear label, is covered in the guide to running a product seeding programme. When creators write to the store asking for samples instead of waiting to be chosen, the guide to answering creator requests for free product and paid work covers screening them and replying.
Judging a sampling programme
The number of samples sent says nothing about whether the programme works. The question is what the sample buyers went on to do, compared with similar shoppers who did not take a sample.
Build sample buyers into a CohortA group of customers who share a starting event and period, tracked together over time. by the month they took the sample, as the cohort analysis tutorial shows, and read at fixed ages: how many placed a full-size order, how many placed a second, and the Contribution MarginRevenue left after every cost that varies with each sale. per customer after the cost of the samples. Compare them with a group that is as similar as the store can make it, such as shoppers who bought the full size first in the same months from the same channels. Before a trial size becomes permanent, the trial size product test is a design that replaces this matched group with a random one, offering the trial to half of first-time visitors. The guide to new and returning customers explains how to record customer status so a sample buyer's first full-size order is counted as a new customer's order, not a returning one.
Three readings keep the judgement honest:
- Cost per full-size customer, not per sample. Divide everything the programme cost by the number of full-size customers it produced; that is the programme's real CACCost to acquire one new customer..
- Value after the first order. Sample buyers who go on to buy only once may be worth less than they look. Read repeat purchaseShare of customers who placed more than one order in the period. and lifetime value (LTVTotal value a customer brings over time.) at the same ages as the comparison group.
- Where the samples went. Read by source: requested on the site, added to orders, through partners, to reviewers. Each has a different cost and a different audience.
How sample stock is accounted for, as cost of sales or as marketing expense, is a policy decision. The marketing P&L guide lists free samples and promotional stock as an item whose treatment varies and says to ask finance; agree it before the programme is judged, so the cost is counted once.
A skincare brand prices a trial set (illustrative)
Pumicelle Skincare replaces free samples with a paid trial set
Pumicelle Skincare is an invented online skincare brand. For two years it sent free samples of its cleanser to anyone who asked. Over the last twelve months 3,200 samples went out and 96 of those requesters, 3%, placed a full-size order. Each sample cost £2.10 to make, pack and post, so the programme spent £6,720 to win 96 customers: £70 each.
It replaces the free sample with a paid trial set of three small sizes (cleanser, serum and moisturiser) at £8, including delivery, with the £8 credited against a first full-size order within 60 days. The set costs £5.40 to make, pack and post. Requests fall to 640 in the following six months, and 154 trial buyers, 24%, place a full-size order. The sets bring in £1,664 more than they cost, which more than pays for the £1,232 of credits redeemed, so the programme now wins full-size customers at no net cost before the margin on their orders.
Pumicelle reads trial buyers as a monthly cohort against shoppers who bought a full-size product first from the same channels. After six months, 38% of trial buyers have placed a second full-size order against 29% of the comparison group, and contribution per customer is £4 higher. The groups are small, so the team treats the gap as a direction rather than a precise figure and keeps the set for another two quarters before deciding on its final price.
Common mistakes
- Counting samples sent as success. The measure is full-size customers and their margin, against a comparison group.
- Sampling products a sample cannot sell. If the first use is the whole value, a sample answers nothing.
- Giving samples away free by default. Free attracts collectors; a price or delivery charge filters for intent.
- Treating a sample request as marketing consent. Offer marketing separately; follow up within what the request covers.
- Pricing the trial size below the full size per unit. Some customers will never trade up.
- Leaving disclosure to chance. Free product to reviewers and creators is a material connection to be disclosed clearly.
- Sending random samples in orders. A sample chosen from what the customer bought works; a leftover does not.
