How long to give a product before you decide

The most expensive mistake in skincare is not buying the wrong product. It is buying the right one, using it for nine days, deciding it does nothing, and buying something else.

The second most expensive mistake is the mirror image: keeping something for eight months because it cost a lot and you have not yet admitted that it does not suit you.

Both are failures of timing rather than judgement. So here is the timing, as plainly as we can put it.

A routine is not one experiment. It is five, on five clocks — and a sixth with no clock at all.

Almost all of the disappointment in this category comes from reading one of those clocks against another one's scale.

• How something feels — minutes. Whether a cream sits under sunscreen, whether a cleanser leaves you tight, whether a serum rolls into little beads under everything else. You know this on day one, and you are allowed to act on day one. A product you dislike is a product you will stop using, and there is no prize for persevering.
• Hydration — days. Skin that is short of water can feel different inside a week of getting some. This is the fastest honest change available in skincare, and it is also the one most often mistaken for something larger.
• Irritation — days to two weeks. If something is going to sting, redden or itch, it usually starts early. Later reactions exist. Early is the pattern.
• Congestion and texture — four to eight weeks. Skin turnover runs roughly four weeks. Anything that depends on cells that have not surfaced yet cannot be read before they have.
• Marks and evenness — twelve weeks, often more. This is the slow end of what a cosmetic can be expected to influence at all, and it is where people give up earliest and most predictably.
• Sunscreen — never. Its entire value is a comparison you will never get to run, because you cannot see the version of your face that did not wear it. Judging a sunscreen on visible results is a category error. Judge it on whether you will actually put it on.

Twelve weeks is the number worth remembering

Not because twelve weeks is magic, but because it is the point past which your opinion starts to be worth something for the slow categories. Most published work on ingredients aimed at texture and evenness runs for eight to twelve weeks, and it runs that long because shorter studies were not able to see anything.

You are not running a study. You have no control group, no standard lighting and no second face to compare against. That is an argument for giving it longer than a laboratory would, not less.

The practical version: for anything you bought to change how your skin looks rather than how it feels, decide in advance that you will not form a verdict before week eight, and write the date you opened it on the bottle in biro. Memory is generous to things we paid for and harsh to things we did not.

How to make the verdict worth anything

Four rules, and the first one does most of the work.

Change one thing at a time. If you start a new serum and a new cleanser in the same week, you have not run one experiment. You have run none. When something improves you will credit the wrong bottle and rebuy it forever; when something goes wrong you will have two suspects and no way to choose between them.

Give it a fixed frequency and keep it. "Most nights" is not a frequency. Twice a week, or every night, decided at the start. A product used erratically for six weeks has not been tested for six weeks.

Write down what else changed. Skin moves for reasons that have nothing to do with what you bought: a fortnight of bad sleep, a heated office, a new job, a change in medication, a holiday, the season turning. If you are keeping any kind of note, keep that in it too. Half the products people swear by were bought in the same month something else in their life got easier.

If you photograph, photograph badly on purpose. Same window, same time of day, no filter, no flattering angle. Consistency matters far more than quality. This is also, incidentally, why we will never show you a before-and-after: change the light and the angle and you can produce one for a product that did nothing at all.

When to stop early, and not feel bad about it

Patience is a virtue in a routine and a trap in a purchase. Stop before the twelve weeks if any of this is true:

• It stings every single time. Once can be your skin having a bad week. Every time is information.
• You are breaking out somewhere you do not normally break out. That pattern is worth reading properly — we have written about it separately.
• You dread the step. You will stop eventually. Stopping now is cheaper.
• It will not live with the rest of the routine — it pills under sunscreen, it never sinks in, it leaves a cast you dislike. No amount of time fixes a texture problem.

And the honest one: if you have stopped being able to say what you are waiting to see, you are not running an experiment any more. You are just spending.

Some products have no result to wait for

A cleanser's job is to be gone. A good one leaves your face comfortable and unremarkable, which is not a result you can photograph. A toner that is mostly water and humectant has done its job when the next step spreads better. Neither is going to reward you at week ten with something visible, and neither is failing.

Before you start any clock, it is worth asking what you would actually accept as an answer. If you cannot finish the sentence "I will know this is working when…", the problem is not the product.

What we do with this

Every product page here carries its step and its frequency, because a product used at the wrong frequency will fail a test it was never given fairly. Where the published evidence for an ingredient is thin, we say so on the page rather than letting the brand's confidence stand in for ours.

And if you are between things and not sure what to hold on to, the routine builder will give you three or four steps and tell you what it left out. It is a better use of twelve weeks than a shelf of half-finished bottles, which is the thing this shop exists to argue you out of.