By Nicole · product development, Dongguan Mingmei · updated November 2026 · 8 min read

Two coffee scales can look identical, both say 0.1 g, both cost 0.1 g of nothing to photograph — and one sells for $12 while the other sells for $45. The difference is not the shell. It is nine decisions inside, and almost none of them appear on a product page.

This is the list we walk through with buyers who are specifying their own model, written in the order that actually moves the price.

1. The load cell is the whole argument

The load cell is the metal bar that bends when you put weight on it. It is the single largest quality decision in the product, and it is invisible.

• Class. Load cells are graded by accuracy class. Two cells that both read 0.1 g can differ in how much they drift at 40 °C or after 5,000 cycles.
• Off-center error. Put 500 g in the middle of the platform, then in the corner. A good scale repeats within one division; a cheap one moves two or three. Coffee scales get loaded off-center constantly because the carafe sits wherever it sits.
• Creep. Leave 500 g on for ten minutes and watch the number. If it walks, your customer's brew log is wrong and they will blame the app, the coffee, then the scale.

Ask any factory one question: which load cell do you use at 3 kg, and what is the off-center error at half capacity? A factory that builds scales answers in one sentence. A trader forwards the question to someone else.

2. Settle time beats resolution

0.1 g is a resolution, not an accuracy. What the user feels is how fast the number stops moving. Pour-over brewing is a race: you are watching the weight climb toward 250 g while the kettle is still pouring.

• A scale that settles in under a second feels expensive.
• A scale that takes two to three seconds feels broken, even if it is perfectly accurate when it finally stops.
• Refresh rate matters as much as filtering: 5 readings a second is a different product from 10.

This is where firmware filtering gets traded against responsiveness. Cheap designs over-filter to make the number look stable — and create the lag everyone complains about in reviews. It costs nothing to fix in code and a lot to fix in reputation.

3. Repeatability is what your customer actually feels

Put the same 200 g weight on the scale ten times, taring between each. Note the spread. That spread, not the printed resolution, is the product.

For a kitchen scale used for flour, a two-division spread is invisible. For a coffee scale used to dose 18.0 g, it is the difference between a good shot and a sour one — and it is the number one reason a scale gets returned as "inaccurate" when a lab would call it in spec.

4. Drift: the first ten minutes

Most scales are temperature-compensated, few are compensated well. Turn the scale on from cold, leave it on the counter, and watch zero for ten minutes while the electronics warm up.

If zero moves, the user tares constantly and trusts the product less. This is also why a scale that reads perfectly in a 22 °C office can behave differently in a 30 °C kitchen — worth testing if your market is warm.

5. Platform size and where the cup lands

A 58 mm portafilter and a 600 ml server are not the same footprint. Too small a platform pushes weight to the edge, where off-center error lives.

• Around 100 × 100 mm is the practical minimum for espresso work.
• For pour-over with a server, 130 mm and up stops the "it wobbles" complaint.
• Bigger platform, bigger load cell, bigger housing, higher cost — it is a real trade, not free real estate.

6. Water protection: what the silicone mat really does

Almost no coffee scale in this price range carries a genuine IP rating, and any supplier who says "waterproof" without a test report is using the word loosely. What actually protects the product:

• A removable silicone mat that lifts the platform edge above spills (and can be washed separately).
• Gaskets or sealed seams at the housing joint, not just a tight fit.
• Keypad membrane rather than separate metal domes that let liquid under the overlay.
• Drain geometry: where does 30 ml of spilled water actually go?

A mesh or grid silicone mat that can be lifted out and rinsed is one of the cheapest upgrades you can put on a coffee scale, and one of the easiest to show in a listing photo.

7. Power: AAA or USB-C

8. The timer is half the product

On a coffee scale, the timer is not a feature — it is the workflow. Three details decide whether it feels designed or bolted on:

• Where the button is. If starting the timer means covering the display with your hand, it will not be used.
• Auto-start on pour. Nice when it works, maddening when a wobble triggers it. If you ship it, make it switchable.
• Dual display. Timer and weight side by side on one screen — the alternative is a mode toggle mid-brew.

9. Display and the low-light test

Cafés are dim. Take the sample into a room with the lights low and read it at arm's length. Contrast, viewing angle, and whether the backlight washes out the last digit matter more than the LCD size on the drawing.

Where to spend first

If you are building to a target retail price, spend in this order. Every line above it is money you cannot recover later.

A spec sheet you can send to any factory

• Capacity and division: 3 kg / 0.1 g, 5 kg / 0.1 g
• Units required: g, oz, ml (plus ml milk if you sell into EU kitchens)
• Platform size and whether the mat is removable
• Power: AAA count or USB-C, or both
• Timer: separate button / auto-start / dual display
• Required certifications and destination market
• Target quantity, target landed cost, target retail price
• Packaging: colour box, manual, barcode, language set

With those eight lines, any real factory can quote in a day. Without them, you will get seven prices that are not comparable.

Related: our coffee scale range · how to vet a factory · private mold vs public mold