The minute under the tongue is how CBD oil is used
If you already own the bottle, almost everything worth knowing fits in one instruction: put the oil under your tongue and leave it there for a minute before you swallow.
Held sublingually, reported bioavailability is in the region of 13 to 19 %. Swallowed immediately, it drops to around 6 %, because everything that goes down the throat passes through the liver before it reaches circulation.
Two seconds of impatience costs a factor of two or three. That is a bigger difference than moving from a 10 mg product to a 25 mg one, and it is free.
Most pages answering this question hand you a catalogue of formats. You did not ask which format to buy — you have a bottle in your hand. This page is about what to do with it, what not to do with it, and the one thing on that list that is genuinely dangerous.
The technique
Where “under the tongue” actually means
Lift your tongue and place the oil in the hollow beneath it, against the floor of the mouth. That tissue is thin and richly supplied with blood vessels, which is why the route works at all.
Oil sitting on top of the tongue does very little. The taste is the same and the absorption is not.
How long
A minute is the usual advice, two is better. Longer produces diminishing returns and a great deal of tedium.
Set a timer the first few times. Almost everyone underestimates a minute by half.
What to do with your mouth meanwhile
Nothing dramatic. Do not swill it around, do not talk, and try not to swallow reflexively — saliva will accumulate and swallowing that saliva takes the oil with it.
Breathe through your nose and wait. Then swallow whatever is left; the swallowed portion is not wasted, it simply takes the slower route.
Before or after eating
If you are going to swallow a meaningful part of it anyway, and most people do, then fat helps. Oral absorption of cannabidiol has been shown to increase several-fold with a high-fat meal, with figures around four to five times reported.
Taking it after a meal containing real fat is the single largest controllable variable in this whole process, and almost no bottle mentions it.
And do it the same way every time
Same place, same duration, same relationship to food, same time of day. Consistency is not fussiness — it is the only way to remove the variation that would otherwise make your own experience uninterpretable.
What not to do
Never vape an oral oil
This is the one genuinely dangerous item on the page, and it deserves its own line.
Oral cannabidiol oils are made with carrier oils — MCT, hemp seed, olive. Those are lipids. Inhaling lipids into the lungs is associated with lipoid pneumonia, a serious condition, because lungs have no mechanism for clearing oil.
Vape liquids are made with propylene glycol and vegetable glycerine instead, which are not lipids. The two products look similar in a bottle and are not interchangeable. If a bottle does not say it is for inhalation, it is not.
Do not stir it into a hot drink
Cannabidiol is fat-soluble and essentially insoluble in water. Dropped into tea or coffee, most of it will bind to the mug and to whatever fat happens to be present, and the rest will float.
If you want it in a drink, the drink needs fat: whole milk, cream, or a spoon of coconut oil. This is the same chemistry that makes plain CBD tea largely pointless.
Do not cook with it at high heat
Heat degrades cannabinoids and the conversion is not in your favour. If you must add it to food, add it to something already cooked and cooled a little.
Do not put an oral oil on your skin expecting a topical
Topical preparations are formulated to interact with skin. An oral oil applied to skin mostly sits there. It is not harmful, it is simply not what either product was designed for.
Do not judge it on the first day
Cannabidiol is highly lipophilic — it distributes into fat and returns slowly. With repeated daily use, levels build across several days before they settle.
Someone who takes one dropper, feels nothing, and concludes the product is inert has run an experiment that could not have produced any other answer.
Reading your own bottle
Work out the strength
Percentage × 10 = mg/ml. A 5 % oil is 50 mg/ml, a 10 % oil is 100 mg/ml.
Total mg ÷ millilitres = mg/ml. A 30 ml bottle labelled 1,500 mg is 50 mg/ml, which is half the strength of a 10 ml bottle labelled 1,000 mg despite the larger number on the front.
Work out the drop
A pipette delivers roughly 0.05 ml, about 20 drops per millilitre, so mg/ml ÷ 20 ≈ mg per drop. A 10 % oil gives about 5 mg a drop.
That is an estimate rather than a measurement. Drop size varies with pipette bore, viscosity and temperature, and 20 % variation between drops is ordinary.
Work out the value
Price ÷ total mg = cost per milligram. It is the only comparison between products that packaging cannot distort.
Check what else is in it
If the label lists melatonin, CBN, valerian, ashwagandha or L-theanine alongside the cannabidiol, whatever happens next tells you nothing about any single ingredient. For learning what cannabidiol does to you, buy the plainest bottle available.
Storing it properly
The three enemies
Heat, light and oxygen. Cannabinoids degrade and carrier oils eventually go rancid, and both are accelerated by a warm, sunny shelf.
Why the bottle is dark
Amber or cobalt glass is not styling. It blocks the wavelengths that do the damage, which is why a clear bottle is a small warning sign in itself.
Where to keep it
Closed, cool, dark. A cupboard is enough. A fridge is unnecessary and causes condensation every time the bottle comes out, and cold oil is more viscous, which changes your drop size.
What a degraded oil does
Nothing dangerous. It becomes weaker and tastes worse — which is enough to ruin any conclusion you were trying to draw about whether it works for you.
The timeline, so you know what you are waiting for
A large share of the disappointment with this product comes from expecting the wrong thing at the wrong moment.
The first fifteen minutes
Almost nothing, other than taste. Sublingual absorption is faster than swallowing, but it is not inhalation, and the tissue under your tongue is not a shortcut to instant anything.
If you feel something dramatic in the first two minutes, it is expectation. That is not an insult — expectation is a real effect with real consequences — but it is not pharmacology.
Fifteen to forty-five minutes
This is where the sublingual portion arrives. Reported onset for the route sits in that window, and whatever the held portion is going to do begins here.
One to two hours
The swallowed portion arrives now, having gone through the liver. It is the larger share by volume and the smaller share by effect, and it is slower and flatter.
This is also why a dropper taken at bedtime peaks somewhere in the middle of the night rather than at the moment you wanted to fall asleep.
The rest of the day
Cannabidiol is highly lipophilic. It distributes into fatty tissue and comes back out slowly, which is why estimates of how long it persists vary so widely — from a few hours after a single administration to two to five days with sustained daily use.
The practical consequence is a long, low tail rather than a clean stop.
The first week
Because of that tail, daily use does not reach a stable level on day one. Levels accumulate over several days before they plateau.
So the first week is the loading period, not the trial. Whatever you conclude during it is a conclusion about a moving target.
The second week is the actual test
By then, if you have kept the route, the timing and the food constant, you are measuring something steady. That is why a fortnight is the minimum honest window, and why anyone reporting a verdict after two days has not run an experiment.
When the oil is the wrong tool
Being useful here means saying when the thing in your hand is not the right thing, and a page that only ever recommends is not giving advice.
If you cannot manage the minute
Some people gag on the taste, some find the wait impossible, and some simply will not do it consistently. All three are fine and all three make an oil a bad choice.
A capsule delivers a fixed, tasteless quantity that you cannot get wrong. It absorbs worse — around 6 % — but a capsule you actually take beats an oil you take badly.
If you need precision
Drop size varies with pipette bore, viscosity and temperature. If you want the same quantity every single time, for a personal trial or because you are tracking something carefully, a capsule is the format that guarantees it and an oil is the format that does not.
If the problem is in one place
A knee, an elbow, a patch of skin. A topical is formulated to interact with skin and stay roughly where it is put; an oral oil is formulated to be absorbed through mucous membrane and gut.
Neither substitutes for the other, and rubbing an oral oil into a joint is the most common wasted purchase in this category.
If you are screened at work
Full-spectrum oils contain trace THC by design, and it accumulates in fatty tissue with daily use. A workplace panel looks for THC metabolites, not for cannabidiol.
If your job depends on that test, the only defensible version is broad-spectrum or isolate with a batch certificate showing non-detectable THC — and even then the certificate is a document you did not write.
If you are already taking something
Cannabidiol inhibits CYP3A4 and CYP2C19. If you take a regular prescription, the right first step is not choosing a format, it is a two-minute conversation with a pharmacist about whether you should be choosing one at all.
If nothing has changed after a fair trial
Two weeks, consistent route, consistent timing, a number written down, and no movement. That is a result, and it is worth as much as a positive one.
The failure mode this page most wants to head off is not buying the wrong bottle. It is spending six months and several hundred euros escalating through products because no one ever defined what “it did not work” would look like.
Running an honest trial on yourself
Give it a defined stretch
Two weeks minimum, at the same time each day, changing nothing else. Two nights prove nothing; you have two good nights most fortnights regardless.
Write down a number
A number, not an impression. A sleep time, a pain score out of ten, a validated questionnaire — the instrument matters less than the fact that it produces a figure you can compare.
Impressions on this subject are shaped by what you paid, which is precisely the effect an honest trial is trying to see past.
Decide in advance what “no” looks like
Pick the threshold before you start, and the point at which you stop. Without a defined failure condition, a personal trial can only ever confirm, and the money keeps going out.
One variable at a time
If you change the product, the route, the timing and the meal in the same week, you have learned nothing about any of them.
What the evidence supports, plainly
There is no authorised claim
Not one authorised health claim for cannabidiol exists in the EU register, for any purpose. Whatever a bottle implies is doing legal work rather than scientific work.
The quantities do not line up
EFSA’s provisional safe intake for cannabidiol as a food, published 9 February 2026, is about 2 mg a day for a 70 kg adult, and does not apply to anyone under 25, pregnant, breastfeeding or taking medication. A 2023 University of Sydney review found benefits became clearly evident at 300 mg a day and above.
Consumer products sit between the two, which is worth knowing before you form expectations.
What is actually in the bottle
A 2024 review in The Journal of Pain found products sold in North America and Europe with cannabidiol content ranging from none to much more than advertised. A batch certificate matching your bottle is the only check available.
Interactions, and who to ask
The enzymes
Cannabidiol inhibits CYP3A4 and CYP2C19, which between them metabolise a large share of common medicines. Concentrations of those medicines can shift without anyone deciding they should.
The grapefruit shorthand
If a medication leaflet warns against grapefruit, the same pathway is involved and the same caution applies.
Who should ask before starting
Anyone on a regular prescription, anyone pregnant or breastfeeding, anyone with a liver condition, anyone under 25 given EFSA’s own exclusion, and anyone screened at work.
A pharmacist answers the interaction question in two minutes with your medication list in front of them, and it costs nothing. That is the correct destination for it, and no page can substitute.
The short version
The one instruction
Under the tongue, one minute, then swallow. Everything else on this page is secondary to that.
The second one
Take it with fat, and take it the same way every day. Those two remove most of the variation you would otherwise be fighting.
The one that matters most
Never inhale an oil made for swallowing. Carrier oils in the lungs are a genuine medical risk, and the two bottles look alike.
And the honest framing
None of this makes a product work. It makes your experiment fair — which is the most any page can give you, and more than a dosage chart pretending otherwise.
Why a page about using it says so little about what it does
You will have noticed that this page explains technique in detail and effects hardly at all. That is deliberate, and it is worth saying why rather than leaving it as an odd absence.
There is no authorised health claim for cannabidiol anywhere in the EU register, for any purpose, so a specific promise here would be unlawful as well as unearned. But the deeper reason is that the evidence does not support one. For pain, a 2024 review found fifteen of sixteen randomised trials showed no benefit over placebo. For sleep, the best systematic review found only two of thirty-four studies were in people with insomnia.
What we can give you honestly is the mechanics: how much of it reaches your bloodstream by which route, what multiplies that figure, what destroys it, and how to run a trial whose result you can believe. That is a smaller offer than most pages make, and it is the part that is actually true.
If the answer at the end of your fortnight is that nothing happened, this page will have done its job — because you will know that cleanly, in two weeks, instead of suspecting it expensively for a year.