A coffee bag is two documents printed on the same surface. One of them is advertising, written by someone whose job is to make you pick it up. The other is a set of constraints the seller accepted because the coffee physically required them, or because a buyer somewhere in the chain insisted. Those two documents are printed in very different sizes, and the useful one is almost always the smaller.
Read a bag in this order. Construction first, because the valve and the seal are manufacturing decisions rather than copy. Then the roast date, because it is the only claim on the bag that expires. Then how narrowly the coffee is identified. Then process and variety, which tell you what kind of coffee to expect rather than how good it is. Everything set in the largest type is the cheapest thing on the bag to say, and that ordering is not a coincidence.
The valve, and the thing most advice gets wrong about it
Roasting drives carbon dioxide into the bean, and it comes back out over the following days and weeks. A properly sealed bag with no way to vent that gas will swell and eventually split. The small plastic disc on the side or the base is a one-way degassing valve, and it solves the problem in one direction: gas leaves, oxygen does not enter.
Here is the part that gets repeated wrongly across a great deal of consumer coffee writing, including at least one published book in our own source set. A bag cannot meaningfully be both vacuum-packed and fitted with a one-way valve. They are competing solutions to the same problem. A genuine vacuum brick is hard-packed and rigid, with no valve at all, and it is rigid precisely because the coffee inside had already finished degassing before it was packed. That means it sat somewhere first. A valve bag is soft and unvacuumed, and it is soft because the coffee inside is still working. Any bag advertising both is describing something that does not exist, and whoever wrote the label was not the person who packed it.
Hoffmann sets out the packaging tiers plainly, worst to best: unsealed kraft paper, which holds condition for days; foil with a degassing valve, which holds it for weeks; and nitrogen or inert-gas flushed foil or aluminium, which holds it for months. The threshold that matters is residual oxygen. Below roughly one per cent, shelf life extends dramatically, which is why gas flushing is worth the equipment it requires.
One practical consequence nobody prints on the bag: the moment you open a valve bag, the valve has done its entire job and stops contributing. From then on the bag is a paper sack with a fold in it. Push the air out before you roll and clip it, and if you buy in quantities you cannot finish inside a month, freezing is a genuinely good answer and not the compromise it is usually treated as. What you should not do is put the open bag in the fridge.
The roast date is a number you count from, not a number you obey
A best-before date sitting twelve to twenty-four months out is not a lie. It is a statement that the seller does not consider freshness the thing being sold, which is a legitimate position and a completely different product promise. Anette Moldvaer's position in Coffee Obsession is the useful one: the roast date is the number that carries information, and its absence is itself information.
What almost no bag tells you is that the date is a starting point for a count, not an instruction to brew immediately. Fresh coffee is still full of carbon dioxide, and that gas disrupts the water's contact with the grounds. The reference sources disagree about how long to wait, and the size of the disagreement is worth knowing. Hoffmann suggests three to four days at home. Colonna-Dashwood writes that the peak can arrive three to six weeks after roasting. Moldvaer's guidance for home-roasted coffee is one to two days for filter and about a week for espresso.
That is not a small gap. It is close to an order of magnitude, from two experienced authorities and a third somewhere between them, which tells you the honest answer is that it depends on the coffee and the method you brew it with. Coffee that is too fresh is a real and common problem, and a week is the useful default. Treat the printed date as the beginning of your own record rather than as a verdict.
How narrowly is this coffee identified
Kenneth Davids sets out the naming layers the green trade uses: country, then market name, then grade, then estate or cooperative, then occasionally botanical variety. He notes drily that a fully specified label is theoretically possible and almost never seen, because a customer confronted with country, region, farm, process and grade all at once tends to lose interest before buying.
For reading a retail bag, treat it as a question of narrowness. A country on its own is the floor. A region narrows it. A washing station or a cooperative narrows it considerably, because that is a physical building where a defined group of growers delivered fruit on defined days. A farm name with a harvest year is about as far as consumer labelling goes, and it is the point at which a false claim can be contradicted by someone with records.
Grade names sit oddly here because they read as quality and mostly are not. Kenyan AA and Colombian Supremo are screen sizes. Guatemalan Strictly Hard Bean is a proxy for growing altitude. Davids is explicit that grading tends to measure externals such as defect counts and bean size rather than taste, because those can be assessed at a mill by someone who is not a cupper.
Variety tells you the ceiling, not the cup
If the bag names a variety, that is a signal about the roaster more than about the coffee. Davids observes that the fine-coffee trade is only beginning to market by variety at all, and that for many lots nobody in the chain could say which arabica varieties produced the fruit. A roaster who prints Bourbon, SL28 or Caturra has a supply chain detailed enough to know.
What the name does not do is predict the cup. Jessica Easto puts the mechanism cleanly in How to Taste Coffee: genetics set the potential, and everything after that can only preserve or destroy it. If the quality is not encoded in the seed, no amount of careful processing, roasting, brewing or storage will add it. If it is, four subsequent stages have the opportunity to lose it. A variety name is a statement about the raw material, and the raw material is the first of five things that had to go right.
Process is the most predictive word on most bags
Washed coffee has all the fruit removed before drying, and it tends to read cleaner and brighter. Natural process dries the whole cherry around the bean, and it tends to read fruitier and heavier in the mouth. Easto attributes the body difference to the effect processing has on polysaccharides rather than to any fruit flavour soaking in. Honey and pulped natural sit between the two, with the skin removed and a controlled proportion of the sticky mucilage left on.
Davids goes a layer deeper in a way that is rarely covered outside the trade. Within washed processing there is wet fermentation, where clean water is added while the fruit pulp breaks down, and dry fermentation, which skips the water and lets the beans break down in the pulp still on them. Kenya and Ethiopia ferment wet. Most Guatemalan coffees are dry fermented. It is a small structural difference upstream of anything you can control, and it is part of why two coffees both labelled washed can taste so unlike each other.
What the bag cannot tell you
It cannot tell you how the coffee was stored between the roastery and your hand. A bag roasted ten days ago that spent nine of them in a hot shop window is in worse condition than one roasted a month ago and kept cool and dark. Every date on a bag is a proxy for condition rather than a measurement of it.
It cannot tell you whether the cherries were picked ripe, which is the single largest quality decision made on any farm, and one that no label anywhere records.
It cannot tell you whether the tasting notes will arrive. Those notes describe what the roaster found, at their roast level, on their equipment. Easto is honest about how unreliable that transfer is, and she notes that even professional cuppers are vulnerable to finding the attributes they expect, which is why the Specialty Coffee Association recommends that a cupper know as little as practical about the coffee in front of them.
And it cannot tell you whether you will like it. Nothing printed on a bag has ever been able to do that, and a bag with more information on it is not closer to being able to.
The four-step read
Turn the bag over and look for the valve or the gas-flush claim. Find the roast date and work out how many days old the coffee is. Read how narrowly the origin is stated, and take the specificity rather than the adjectives. Then read the process word, because it is the one that most reliably predicts what you are about to taste.
If a bag passes all four and the coffee still disappoints, the bag was probably not the problem. The grinder sits above the coffee in the order of things worth fixing, and a bag chosen well and ground badly tastes exactly like a badly chosen bag.
We sell the objects coffee is brewed in and drunk from. We do not sell the coffee, so nothing above is written to move a bag off our own shelf, and you can weigh it accordingly.