Sample Request Email to a Supplier: Draft It, Then Score What Comes Back

A sample request email to a supplier should say who you are, exactly which item and variant you want, the specification it must meet, how many pieces, whether you will pay, where to send it and by when. Then score what arrives against that same specification, with weights, on a one-page scorecard filled in the day it lands.

Most weak sample requests fail on the second half. People write a polite "could you send a sample?" and then judge the parcel by feel. The email and the scorecard are one document in two halves: every line you ask for in the email becomes a row you mark afterwards. AI is useful for drafting both from a single spec, and for keeping a log of what each supplier said.

Follow me on Instagram@sagnikteaches

Start with the spec, because the email is built from it

Before you write to anyone, write down what a good sample means for you. Put it in a table with three columns: the attribute, the target and the tolerance. If you cannot fill in the tolerance, you do not yet know what you are buying.

Connect on LinkedInSagnik Bhattacharya

Take an illustrative online clothing store sourcing a heavyweight cotton hoodie in three colours. The spec might read:

Subscribe on YouTube@sagnikteaches

That table is the heart of the whole exercise. A furniture maker would swap in moisture content of timber, joint fit and finish; a pest-control firm buying a trial of bait stations would list active ingredient, tamper-resistance and refill time. The categories change but the three-column habit does not.

A sample request email you can copy and fill in

Keep the first message short enough to read on a phone. A supplier receiving ten requests a day will answer the one that is clear and specific. Here is a copy-ready version with the blanks marked.

Subject: Sample request: [product name / your item code], [quantity] pieces

Hello [first name],

I run [business, one line]. We are looking for a regular supplier of [product] and your [catalogue line or item code] looks like a fit.

Could you send [number] sample(s) of:
- Item: [code / description]
- Variant: [size, colour, material]
- Specification we need it to meet: [three to five key lines from your spec]

Please tell me:
1. Is the sample free, paid, or paid and credited against a first order? If paid, the fee and how you take payment.
2. Shipping cost, courier and expected delivery time to [city/postcode field].
3. Is the sample from the same production run and materials as bulk orders?
4. Minimum order quantity, price breaks and lead time for bulk.
5. Any certificates or test reports you can share for this item.

If a sample cannot match the specification above, please say which point and what you would change, rather than send the nearest item.

We aim to decide within [ten] working days of the sample arriving. If this goes well, our first order would be around [quantity].

Thank you,
[your name]
[business]
[phone, if you wish]

A few choices in that email are deliberate. The line "from the same production run and materials as bulk" catches suppliers who send a hand-finished showpiece. The line asking them to say which point cannot be met invites an honest answer instead of a substitute. Naming a decision date and a likely first order quantity tells them whether you are a serious buyer, without promising anything you cannot deliver. Leave the quantity out if you do not know it, rather than inflating it.

Ask AI to draft it, with your spec as the only input

Paste your spec table and a few facts about your business into a chat assistant and ask for the email. The prompt matters, because a model asked to "write a sample request email" gives you something polite and generic.

Write a sample request email to a supplier. Use only the specification and facts below. Keep it under 200 words. Put the spec as a short bullet list. Ask the five questions listed. Do not promise any order size or date I have not given you. Tone: direct and courteous, no flattery.

My business: online store selling hoodies and sweatshirts.
Item: heavyweight cotton hoodie, style H-380, colours charcoal, bone and forest.
Samples wanted: one of size M in charcoal.
Spec: 380 gsm cotton face, brushed back; chest 56 cm +/- 1 cm; shrinkage under 3% at 40C; reinforced neck.
Questions: sample cost and credit, shipping, same run as bulk, MOQ and price breaks, test reports.
Decision date: ten working days after arrival. Likely first order: 300 pieces.

A typical output (illustrative) is a clean message. Read it for three things before it goes: that the numbers match your spec, that it has not added a compliment about the supplier's "excellent reputation" you cannot support, and that it does not invent a deadline. Delete anything that was not in your input.

If the supplier is found through a marketplace and you want to check them before spending on a sample, the Alibaba vetting tutorial covers that earlier step.

Free samples, paid samples and credited samples

Suppliers handle samples in three common ways, and each should change what you write and how much you read into the reply.

• Free. Normal for small standard items where the supplier carries stock. Expect to pay shipping. A free sample of a standard line tells you about the standard line; it does not tell you whether they can match your custom colour or size.
• Paid. Normal for custom, made-to-order or high-value pieces. Ask for a written quote of the fee, and ask whether it is credited if you place a first order of a stated size. Be wary of high fees with no credit and no description of what the fee covers.
• Paid and credited. The fair middle. You carry the cost of proving the product, and the supplier gets a reason to take you seriously. Write the credit terms into your email reply so there is no argument later.

Whichever applies, pay through a method that leaves a record and keep the receipt in your log. Do not send money to a personal account or to a new payment route proposed halfway through a thread, and treat a supplier who asks you to move the chat to a different app with caution until you have verified them another way.

Score the sample the day it arrives

Judging by memory three weeks later is how a charming supplier wins on personality. A scorecard forces you to separate the product from the relationship, and it lets you compare two or three suppliers on the same footing.

Weight the rows by how much they matter to your customers. For a clothing store, fit and fabric matter more than packaging; for a furniture maker, joint strength matters more than box design. Give the weights before the sample arrives, not after, so you do not bend them to favour the one you liked.

The weighted column is score divided by five, times the weight. The figures above (illustrative) total 81 out of 100, and a simple rule of thumb is that anything under 70 does not get a bulk order without a second sample, 70 to 85 gets a small trial order, and above 85 goes forward. Choose your own cut-offs, write them on the sheet and stick to them.

Because weights multiplied by scores is arithmetic, let a spreadsheet total it. If you want AI to help, it can build the sheet from your spec, but do not ask it to do the sums in conversation. The tutorial on AI and maths explains why. Hand-totalling has its own trap: score the packaging row as 6 instead of 3 (forgetting to divide by five) and this card reads 84, close enough to a cut-off of 85 to change the decision. A formula column cannot make that slip.

A second illustration: comparing two furniture suppliers

A small furniture maker wants a supplier of oak drawer fronts. Two suppliers send one sample each. The maker's spec lists thickness 18 mm plus or minus 0.5, moisture content 8 to 10 per cent, finish smooth to 180 grit, and no open knots on the face.

Supplier A's sample measures 18.2 mm, moisture 9 per cent, but has one small knot; supplier B's is 17.4 mm with moisture 12 per cent and a perfect face. On a scorecard, A loses a few points for the knot but passes the critical rows; B fails the thickness and moisture rows, which would cause warping and fit trouble, so its beautiful face counts for little. Without the scorecard the maker might have picked B for looks. This is why the weights are set before opening the parcels.

Using AI to tally and to draft the follow-up

Once you have marks, there are two jobs where AI saves time: turning scores into a plain summary for a partner or colleague, and writing the reply to the supplier. Paste your completed scorecard rows (not any confidential pricing) and ask:

Here is a supplier sample scorecard with notes. Write (1) a three-line summary of strengths and weaknesses, and (2) a reply email to the supplier. In the reply: thank them, list the two points that missed the spec with the measured figures, ask whether they can correct them and send a second sample, and say we will decide within seven working days of receiving it. Do not agree to any order. Keep it under 150 words.

A realistic result (illustrative) would be a firm, civil email that quotes "chest 55.5 cm against a spec of 56 cm plus or minus 1 cm" as within tolerance and picks out the loose cuff threads and the missing size sticker as the issues. Check that it has quoted your numbers rather than inventing others, and that it commits you to nothing. If you want to understand how to compare the resulting price quotes, the quote comparison tutorial picks up from here.

Keep a supplier reply log

A sample request generates a thread: the first reply, a price list, a question about lead times, a courier tracking number, a second quote after you push. Keep it in one log, one row per supplier per request, or the details scatter across inboxes.

Columns that earn their place:

• Supplier and contact name.
• Date request sent and date of each reply.
• Sample status: asked, quoted, paid, shipped, received, scored.
• Sample cost, shipping cost and whether credited.
• Stated MOQ, price at your order size, lead time.
• Claims made in writing (same run as bulk, test report available).
• Scorecard total and decision.
• Link to the email thread or a saved copy.

Written claims matter most. If a supplier says in an email that bulk will match the sample, copy that sentence into the log. Months later, when a bulk order arrives with a different feel, you want the quote and the date. AI can pull these rows out of a pasted thread if you ask it to return only a table with the columns above and to write "not stated" where the thread says nothing. Check the table against the thread before you trust it, because a model may fill a gap with a plausible guess. The supplier management tutorial shows what to track once you have chosen.

Where sample requests go wrong

Picture a pest-control firm that buys a trial of bait stations after a good sample and then finds the bulk delivery has a different lid fitting. In a case like that, the cause is almost always one of four things.

1. The sample was not from bulk production. Fix: ask the question in the email, and keep the reply.
2. The spec had no tolerance. "About 56 cm" lets a supplier ship 54 and call it close. Fix: give the number and the plus or minus.
3. Only one sample, from one run. Fix: for batch-variable goods, ask for two, or place a small first order and score that too.
4. The scoring moved with the mood. Fix: set weights first and fill the card the day it arrives.

A fifth cause is not about the sample at all: a sample posted by courier to an address with no one in, or held at customs, loses a week. Ask for tracking, give a delivery contact and mention any import paperwork the supplier needs from you.

A short run of other requests that need slightly different emails

The same template flexes. Here are four variations, each with the one line you change.

• Packaging sample for a food product. Add "please confirm the material is suitable for direct food contact and share the declaration". Ask for the weight the box is rated to carry.
• Software or service trial. Replace "sample" with "trial account" and ask for the usage limits and whether trial data can be exported. Score it on how many of your real tasks it did.
• Several variants. Put the variants in a table, ask whether samples can go in one parcel, and score each variant on its own row.
• Re-sample after a fail. Quote the failed measurement, ask what changed, and request the same spec on the corrected piece. Keep both scorecards.

If a faulty supplier part is feeding rejects in your own production, the scrap tracking tutorial shows how to see whether the problem sits upstream before you change anything.

A filled-in log row and what it let the buyer say

To make the log concrete, here is one row as it might look for the clothing store after three weeks (all figures illustrative).

Supplier: [supplier A], contact [first name]
Request sent: 3 Nov | First reply: 4 Nov | Sample paid: 5 Nov ($38, credited against first order of 200+)
Shipped: 8 Nov, tracking noted | Received: 14 Nov | Scored: 14 Nov, 81/100
Written claims: "same fabric and cut as bulk" (email 4 Nov); MOQ 150 per colour; lead time 5 weeks
Issues: two loose cuff threads; no size sticker
Decision: second sample of the cuff only; trial order of 150 if it passes

Five weeks later the first bulk delivery arrives and the cuffs are loose again. Because the row holds the date of the written claim and the scorecard note, the buyer's email is one paragraph: "Your sample was scored on 14 November with loose cuff threads, which you agreed to correct; this delivery has the same fault on about one piece in six." That is a very different conversation from "the quality seems lower than we expected".

Testing samples that are not physical goods

Not every sample is a parcel. A freight forwarder offers a trial shipment, a printer sends a proof, a software vendor gives a two-week trial. The same two halves apply: write what the sample must prove, then mark it against that. For a printer's proof, the rows might be colour match to your brand reference, edge sharpness on small type, paper weight and trim accuracy to a millimetre. For a trial shipment, they might be pickup on the promised day, tracking updates, condition on arrival and how fast the forwarder answered a deliberately awkward question about customs paperwork.

A good habit is to plant one small test in every trial: ask something you already know the answer to, or request a small change. How a supplier handles a minor awkward request during the sample stage is the best forecast you will get of how they handle a major one in bulk.

How much time and money to spend on samples

A sensible limit stops sampling becoming a hobby. For low-value goods, one sample each from two or three suppliers, with a paid and credited fee below the price of one bulk unit, is plenty. For items where a bad batch would cost real money, such as a first order of several thousand dollars, spend more: two samples from separate runs, an independent measurement or test, and a small trial order before the big one. Put a number on it: if the first order is $6,000, spending $150 on samples and tests is a sensible 2.5 per cent, while spending $600 would need a reason.

The time cost is smaller than people fear. Drafting the email takes ten minutes with AI, scoring a sample takes twenty, and updating the log takes five. The delay is in the courier, so send requests to all shortlisted suppliers on the same day and compare them when the last one lands.

Checking the process worked

After your first round, ask four questions. Did every supplier answer the five questions in the email, or did you chase? Did the scorecard separate the suppliers, or did everyone land within a few points (in which case your weights are too soft)? Did the winner's bulk delivery match the sample when scored on the same card? And did your log hold every written claim you needed?

The third question is the real test. Score the first bulk delivery with the same sheet and compare totals. If the bulk scores ten points below the sample, you have evidence, in writing and numbers, for the conversation with the supplier, and a reason to hold back the next order until it is fixed.

A request email and a scorecard are small things, but together they turn supplier choice from a feeling into a record. That record is what lets you say, calmly, "your sample measured 56 and this batch is 54", and have the conversation go somewhere.

Sample request questions, answered

Should I pay for samples?

Pay when the sample is custom, made to your spec or costly to ship, and ask for the fee to be credited against a first order. Ask for free stock samples only when the item is a standard catalogue line. A supplier who charges nothing and ignores your spec has not tested anything useful.

How many samples should I ask for?

One of each variant you might sell is the minimum. For anything with batch variation, such as fabric, food or moulded parts, ask for two from different runs if the supplier will allow it. A single perfect sample shows what the supplier can do once, not what arrives in a pallet.

What if the supplier will not send a sample?

Ask why. Reasonable answers include a custom item that costs real money to tool, or a minimum order. If a supplier refuses every kind of sample and every kind of paid pre-production piece, treat that as a risk score of its own and consider a very small first order instead.

How long should I wait for a reply?

Allow two or three working days for a first reply and diarise a polite nudge on the fourth. Note the response time in your log: a slow answer to a simple question tends to predict slow answers when something goes wrong.

Can AI score the samples for me?

AI can build the scorecard, tally your marks and draft the follow-up email. The marks themselves must come from handling the sample, measuring it and testing it against your spec. A photo scored by a chatbot misses weight, smell, finish and tolerances.

Further reads

• Can AI Update Your Plumbing Price Book From Supplier Invoices? — Keep supplier prices current once you buy regularly.
• How to Match Supplier Invoices to Purchase Orders Automatically — Check that invoices match what you ordered.
• How to Automate Purchase Orders and Supplier Emails With AI — Turn the winning sample into a repeatable order.
• AI Quality Control for Small Manufacturers: Where to Start — Carry your scorecard into incoming-goods checks.
• Daily Production Report: Summarise Counts, Scrap and Stops With AI — See supplier quality show up in daily counts.
• How Much Does AI Visual Inspection Cost a Small Manufacturer? — What automated inspection of incoming goods costs.
• Supplier Audit Checklist: Plan and Write Up Visits With AI — A small maker's audit of a powder-coating supplier: pre-visit questions from the NCR log, a 3-hour agenda, an on-site checklist and a findings report.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: general sourcing practice; no vendor pricing is quoted in this tutorial.