Samples, cuppings and trials for wholesale prospects

Short answer. A sample is not there to prove your coffee is good. It is there to get your coffee made correctly on their machine, by their staff, in front of their customers. Choose the format to match how close the café is to deciding: a drop-in with a recipe card and a named call-back day, a cupping at the roastery for cafés you have already spoken to, and an on-site trial with a written end date for cafés that are ready to switch. Then go back and dial it in, because that is where samples die.

What is a coffee sample actually for?

A sample exists to answer one question for the café: does this coffee work on our bar, with our water, our grinder and the person who opens on Monday. It is not a quality argument. Most cafés already believe your coffee is good, or they would not have taken the bag, and nobody switches roaster because a stranger told them the Ethiopia was floral.

That changes what you put in the bag and what you do afterwards. If the point is to be tasted, 250 g and a business card is enough. If the point is to be made correctly, you need enough coffee for the grinder to be dialled in, a recipe to start from, and somebody in the room who has made this coffee before. Everything on this page follows from that.

This page is the middle of three. How to get wholesale coffee accounts covers the prospect list, first contact and the follow-up schedule. Onboarding a new wholesale café covers the first 30 days after a café says yes. This one is only about the sampling in between.

The three ways to sample, and when each is right

There are three formats worth running, and they are not interchangeable. Match the format to how close the café is to a decision, not to how much you want the account.

The drop-in sample

Two bags, a recipe card per coffee, and a date you will call, handed to a named person at a quiet hour. The prospecting guide has the full checklist of what goes in the bag. Three things decide whether it works:

• Size it to be dialled in. A kilo of the espresso. Two 250 g bags of two different coffees is a tasting, not a trial.
• Time the roast, not the delivery. Roast so the coffee reaches its window while they still have it. If your espresso drinks best from day seven, a bag handed over on day two gives them five days of waiting and three weeks of usable coffee; a bag handed over on day fourteen gets blamed for being flat.
• Book the visit while you are standing there. "I'll come back Thursday at ten and dial it in with you, it takes twenty minutes." A sample with no return visit attached is the version that fails.

The cupping at the roastery

Invite two to four cafés at once, mid-morning or early afternoon on a Tuesday, Wednesday or Thursday. Cafés are calmer then, and inviting several at once turns half a day of your time into four prospects rather than one. Cafés also talk to each other, and a head barista who enjoyed the morning tells the next one.

• Four to six coffees, no more. By the seventh bowl nobody is tasting anything. Include your house espresso, the coffee that fills the gap you saw in their menu, and one that is simply interesting to taste.
• Set up properly: bowls and spoons for everyone, a grinder you purge between coffees, a kettle, a timer, spittoons, water to rinse, and a printed sheet with the origin, process, altitude and price per kilo of each coffee. Grind at the table so they smell the dry aroma.
• Pour it blind and reveal at the end. Nobody has to perform, and a café that puts your €18.20 house espresso above your €24.50 seasonal has told you something useful about what to quote them.
• Invite the owner and the head barista. The owner signs; the head barista decides whether it is a pleasure to work with. If only one can come, take the head barista.
• Finish on the espresso machine. Forty-five minutes at the table, then pull the house espresso and steam some milk. A cupping bowl says nothing about how a coffee behaves in a flat white, and a flat white is what the café sells.
• Show them the roastery. Ten minutes: the green store, the roaster, the packing bench. It is the cheapest credibility you will ever buy.

Budget ninety minutes door to door and say so on the invitation. A café owner who thinks they are losing an afternoon will not come.

The on-site trial

A trial is a week or a month of real service: your coffee on their bar, made by their staff, sold to their customers. It is the only format that answers the question the owner actually has, which is whether the regulars will notice. Give it edges before it starts:

• Which coffee. A filter, a decaf or a guest slot is an easier yes than the house espresso. See how a guest coffee programme works.
• How many kilos a week, and how many weeks. Written down, with the delivery days.
• The price. Free, at cost, or normal price with the first delivery free. Decide before they ask.
• Who dials it in. You, on day one, at a quiet hour, with a recipe card left at the machine.
• The review date. A day in the diary when you both sit down and decide.

Two weeks covers the resting window and one full cycle of their week, including the Saturday. Four weeks is worth it when staff turn over or when the owner wants a month's sales figure to look at.

How much coffee does a real sample take?

More than roasters expect, because dialling in is expensive before anyone has tasted anything. On a single grinder a café has to purge the old coffee, then find the setting. Ten to twenty test shots at an 18 g dose is 180 to 360 g, plus whatever the grinder retains between adjustments. Only then does anyone drink it in milk, which is the test that matters, and that takes a few more.

Then there is service. A café selling 100 espresso-based drinks a day at 18 g uses 1.8 kg in a day. So a kilo is roughly one morning on bar: enough to dial in and serve, not enough to see whether customers say anything. That is the honest ceiling of a drop-in sample, and the reason a trial is a different thing rather than a bigger sample.

If a café has a second grinder free, a sample is far cheaper for them to try, because nothing has to be purged and nothing stops being sold. Ask. It is also the argument for sampling a decaf or a filter first: those usually live on their own grinder or brewer.

What does sampling cost, per prospect and per account won?

Work the cost from what the coffee costs you to produce, not from your price list. Giving away a kilo you would sell at €18.20 does not cost you €18.20; it costs the €10.91 it cost to make, plus the margin on a sale you were never going to make anyway. Use your own figure from the wholesale coffee price calculator and change it in the table below.

Coffee at €10.91 a roasted kilo, the cost per bag from the price list worked example, and your time at €30 an hour including the van. Both are stand-ins. Put your own numbers in the calculator.

The line that matters is the last column against the win rate, not the coffee on its own. Most samples come to nothing: a café that already has a roaster it likes, a head barista who moves on, an owner who never got round to it. That is normal and it is not a sign you are doing it wrong. What it does mean is that cost per sample is the wrong number to manage and cost per account won is the right one. A €90 drop-in that wins one account in ten costs €900 an account. A €530 four-week trial that wins one in two costs €1,060, and takes a month of deliveries with it.

Two things this calculation changes in practice. First, spending more per prospect on fewer, better-chosen cafés usually beats spreading bags thinly, because the time is the expensive half and the time is the same either way. Second, a café that will never reach the volume in the last input does not justify a trial, whatever it says about wanting one. Check the arithmetic in how many wholesale accounts you actually need before you roast 32 kg for a café taking 3 kg a week.

The dial-in is where most samples die

A sample fails at the grinder far more often than at the palate. The café pulls three fast, sour shots on a Monday morning with a queue building, decides the coffee is not for them, and puts the bag in the cupboard. Nothing you write on the bag prevents that; a visit does.

The recipe card

Every bag goes out with a starting point in writing: dose, yield, time, water temperature and a grind note, plus the window you roasted it for. It is a starting point on your equipment, not an instruction, and saying that on the card stops a head barista feeling contradicted when they have to go two steps finer. The recipe card template on the onboarding page is the one to use; copy it and fill it in per coffee. Add the mobile number of whoever at your roastery can answer a dial-in question in the next ten minutes.

The water question

Ask about the water before you blame the coffee. Water is most of what is in the cup and it decides how your coffee tastes more than the roast level does. Two questions get you most of the way: is there a filter on the machine, and when was the cartridge last changed. Hard, heavily buffered water flattens acidity and sweetness and leaves a bright coffee tasting of nothing in particular; very soft water makes espresso thin and sour; and untreated hard water scales the boiler, which is the café's problem long after your sample is finished. The SCA publishes a Water Quality Handbook built around total hardness and alkalinity if you want targets to work to. Carry a test strip in the van; it takes a minute and it is a better conversation opener than a tasting note.

When the grinder or the water is the real problem

Sometimes the sample cannot work, and the honest thing is to say so rather than roast around it. Worn burrs produce fines that stop a shot settling at any setting, so ask how many kilos have gone through them and compare that with the grinder maker's guidance. A machine running hot, a cracked group gasket or a blocked shower screen will make every coffee taste the same kind of bad.

When you find one of these, three things work better than a second sample:

• Show them, do not tell them. Pull a shot on their machine next to one made properly. Nobody argues with two cups on a counter.
• Bring a name. An engineer you trust, or the number of the machine's distributor. You are not selling the repair; you are being the supplier who noticed.
• Price the fix into the deal, or walk. A filter cartridge with the first order costs you little. New burrs on a café taking 4 kg a week do not pay back, and a loan espresso machine on that volume pays back even less. The espresso machine loan agreement guide has the arithmetic.

What does not work is choosing a darker or more forgiving coffee to hide a water problem. It hides it until the day the café tastes your coffee somewhere else.

Should you charge for samples?

Give the small drop free and charge for anything that looks like supply. The first two bags are a cost of selling, and asking for €25 at that stage buys you a reputation for meanness in exchange for the cost of a kilo. The line moves once you are delivering: a month on bar at 8 kg a week is 32 kg and four delivery runs, which is an account you are servicing for nothing.

Charging works when it is framed as a commitment rather than a fee, and when the money comes back:

• Charge cost, not price. "The two bags are on us. If you want a full month on bar, I'll do it at our cost, €10.91 a kilo, so we're both in it."
• Credit it against the first order. The café pays nothing if it signs. This is the version that almost never loses a prospect, because a café that means it does not mind, and a café that does not mean it says no now instead of in five weeks.
• Charge normally and give the first delivery free. Best for a café that is already close. You are selling from the first week, which is also a cleaner test of whether they will actually pay on time.
• Charge nothing, but cap the kilos. If you would rather not talk about money at all, say how many kilos the trial covers. Anything beyond it is invoiced.

Whichever you choose, put the figure in writing before the first delivery. A trial whose price is discussed afterwards is an argument about an invoice, which is the worst possible start. If you are going to invoice at all, agree terms up front the same way you would for a real account: see the wholesale terms template.

The trial that becomes a free supply

The most expensive mistake in sampling is the trial nobody ended. It happens quietly. Week two goes well, week three you are busy, by week six the café is ordering the trial coffee as if it were an account, and by month four you are delivering 8 kg a week to a business that has never been sent a price. Asking for money at that point feels, to them, like a price increase.

The fix is one sentence written down before the first delivery: what, how much, for how long, at what price, and when you will both decide. Send it as a note the café can keep, not a remark at the door.

Hi [first name],

Confirming what we agreed so nobody has to remember it.

Trial: [House espresso], 8 kg a week, delivered [Tuesday].
Starts: [Tuesday 6 October]. Ends: [Tuesday 3 November], four deliveries.
Price during the trial: [our cost, €10.91/kg ex VAT], invoiced at the end and
credited in full against your first normal order if you carry on with us.
I'll dial in with you on [6 October] at [09:30] and come back on [14 October].
We'll sit down on [3 November] and you tell me yes or no.

After [3 November] the price is our normal list price of [€18.20/kg ex VAT].
No hard feelings either way, and I'll take the equipment and the bags back if
the answer is no.

[Name], [Roastery], [mobile]

Put the end date in your own calendar too, a week before it lands, so you are the one who opens the conversation. A roaster who turns up on the agreed day and asks for a decision is easier to say yes to than one who has been quietly hoping for a month.

Samples for hotels, restaurants and offices

In a café the person who tastes your coffee usually makes it too. Outside cafés they rarely do, and that single fact changes how you sample. A general manager tastes a beautiful cup you made and signs; six weeks later the coffee is being made by breakfast staff at 06:30 on a machine nobody dialled in, and the account is lost before you knew it was in trouble. See wholesale coffee for restaurants, hotels and offices for how these buyers differ more generally.

• Sample where it will be made, at the hour it will be made. Breakfast service, not a 2pm meeting room. Ask to be there for one morning shift.
• Get the maker in the room. If the decision-maker is the GM or the chef, ask them to bring whoever is on the machine. If that is impossible, run the tasting twice: once for the decision, once for the bar.
• Restaurants: taste it after food. Coffee at the end of a meal follows dessert and wine, and a delicate washed Ethiopian that sings on its own can disappear after a chocolate pudding. Sample it the way the guest will drink it, including with the sugar most guests add.
• Offices and bean-to-cup: send it to the machine. A super-automatic grinds and brews on its own settings and is unforgiving of a coffee roasted for a manual bar. Sample the coffee in their machine, on their milk system, and be honest if it needs a different roast profile than your café blend.
• Hotels: ask what happens at the other outlets. Breakfast buffet, bar and room service can be three different machines and three different standards. One sample per outlet, not one per hotel.
• Groups and tenders: find out who is on the panel. A purchasing team may taste blind against two other roasters, and the sample sizes and the paperwork are specified. Ask what format they want before you roast anything, and ask when the current contract ends.

Volumes outside cafés are usually smaller, so keep the sampling cost in proportion. A restaurant taking 3 kg a week does not justify a four-week trial, but it very often justifies one morning of your time at service.

Following up on a sample without nagging

Phone, on the day you said you would, with one specific question about the coffee rather than a general request for a decision: "did you get the espresso through the grinder, and how did it dial in?" That is a question a busy owner can answer in twenty seconds, and the answer tells you whether the sample worked or never got opened. Fresh Cup's review of a year of its RoasterLink wholesale matchmaking programme found that roasters who prioritised email underperformed by a wide margin against roasters who consistently called their leads, and that roasters who kept up several touches saw response rates as high as 75%. The full rhythm, day by day, is on the follow-up schedule. Stop when a café says no plainly; silence usually means they are busy.

A sample-drop note to copy

One page that goes in the bag with the coffee and the recipe cards. It is not a sell sheet and it is not a letter: it tells whoever opens the bag what to do with it and when you are coming back.

For [café name] — from [Your Roastery]

In the bag
1 kg [House espresso], roasted [Tuesday 22 September]. Best from around
[29 September]. Recipe card in the bag: [18 g in, 38 g out, 28–32 s, 93 °C].
250 g [Seasonal filter], roasted [22 September]. Recipe card in the bag.

How to give it a fair go
Dial it in on a quiet morning, not during the rush. Expect to go [finer] than
your current coffee to start with. Taste it in milk before you decide — that is
what your customers will drink.

If it fights you
Call or text me on [mobile] while you are on the machine. I would rather spend
five minutes on the phone than have you decide it is sour.

What happens next
I will call you on [Thursday 1 October] in the afternoon. If you would like me
to come and dial it in with you instead, say so and I will bring the coffee.
It takes about twenty minutes.

Prices, minimum order, delivery days and payment terms are on the sheet behind
this one. No obligation on any of it.

[Name], [Roastery], [mobile], [email]

The recipe figures are an example of how to fill it in, not a recommendation. Use the recipes and resting windows you roast each coffee for.

Questions roasters ask

How much coffee should a roaster give a café as a sample?

Enough to dial in and serve, which means about a kilo of the espresso, not a 250 g bag. Ten to twenty test shots at an 18 g dose is 180 to 360 g before anyone has tasted it in milk, and a café selling 100 drinks a day gets through 1.8 kg in a morning. A 250 g bag is enough to taste and not enough to judge.

Should a roastery charge cafés for coffee samples?

Give the first drop free and charge for anything that looks like supply. Two bags and a visit are a cost of selling. A month of real service at 8 kg a week is 32 kg, which is a delivery you are making for nothing. Charge cost for a long trial and credit it against the first order, so the café pays nothing if it signs.

How many coffees should be on a cupping for a wholesale prospect?

Four to six. Any more and the last bowls are tasted by people whose palates have gone. Include the house espresso, the coffee that fills the gap you spotted in their menu, and one coffee that is simply interesting. Finish on the espresso machine, because a cupping bowl tells a café nothing about how the coffee behaves in a flat white.

Why does my coffee taste worse at the café than at the roastery?

Usually the water, the burrs or the recipe, in that order. Hard or heavily buffered water flattens acidity and sweetness; worn burrs make fines that will not let a shot settle at any grind setting; and a recipe copied from your own bar rarely survives a different grinder. Pull a shot on their machine before you decide the coffee is the problem.

How long should a wholesale coffee trial last?

Two to four weeks, with the end date written down before it starts. Two weeks covers the resting window and one full cycle of the café's week. Four weeks is worth it if staff turn over or the owner wants to see a monthly sales figure. Longer than that is not a trial, it is a free supply arrangement nobody has ended.

Do free coffee samples actually win wholesale accounts?

Samples do not win accounts on their own; the call after the sample does. Fresh Cup's review of the first year of its RoasterLink matchmaking programme found that roasters who prioritised email underperformed by a wide margin against roasters who consistently phoned their leads. A bag left on a counter with nobody following it up is coffee you gave to a stockroom.

Where Order Book fits

Order Book keeps a café in the same book before it is a customer. Add it as a prospect the day you drop the first bag and move it through Lead, Sample sent, Tasting and Negotiating, so the visits, the cupping and the trial are logged against the account rather than in a notebook in the van. Send a sample from its page and it goes on the picking list and the roast plan like any order, labelled Sample, at no charge and out of every figure. Once the café is signed in it can rate the sample — Loved it, It was fine or Not for us. Prospects do not count toward your plan.

When the sample turns into an order, the café already has its own account, signed in with the address you hold for it and no password, showing your catalog at the prices you set for it, in its own language. Its first real order makes it a client by itself. That makes the first order a few minutes on its own page rather than a price-list email and a phone call, which is the point at which a lot of trials lose their momentum. If you charge for a trial, the payment is recorded against the account with a running balance; the invoice itself still comes from your own invoicing software.

Start free for 30 days, no card, or see what it costs.