How to Tell If a Supplement Is Actually Working
Most people take supplements on faith. You read something, you buy the bottle, you take it for months, and at no point does anything tell you whether it did anything at all.
There is a whole category of apps that looks like it solves this. It does not, and the gap is worth understanding before you download any of them.
What the supplement apps actually do
I looked at the main ones. Here is what each is genuinely for.
Prove It gives you evidence-based reviews of what the research says about a supplement's efficacy and safety, including recommended dosages from clinical studies. It is a good literature shortcut.
SuppCo analyses your stack against protocols for common goals and suggests improvements. Useful for spotting redundancy and gaps.
Supplement Snoop tells you whether a product passed third-party testing for contaminants and whether it actually contains what the label claims.
NSF Certified for Sport does the same job for a narrower, sport-focused list, and it is the one to use if you are drug tested.
CareClinic and Human Health are trackers. They log brand, dose and timing, and let you record how you felt afterwards.
Now notice what that list has in common.
The gap nobody in that category fills
Every one of those apps answers a question about the bottle. Is the ingredient supported by research. Is the product contaminated. Did you remember to take it this morning.
Not one of them answers the question about you. Did this specific thing, at this dose, do anything measurable to your body.
That is not an oversight. It is hard, and it is commercially unattractive. A tracker that logs your dose has satisfied its promise. An app that measures the effect has to be willing to come back and tell you the effect was zero, in a category that is largely funded by people continuing to buy things.
Why self-assessment does not work here
The obvious move is to take the thing and notice how you feel. This fails for reasons that are well understood and hard to overcome by trying harder.
Expectation is powerful. If you paid for something and read that it works, you will tend to notice evidence that it does.
Almost nothing acts alone. You started the supplement in January, when you were also sleeping more, drinking less and training differently. Any of those could be the cause.
Recall is unreliable. Asked in March how you felt in January, you will reconstruct rather than remember, and you will reconstruct in the direction of the story you already hold.
Regression to the mean fools everyone. People start supplements when they feel worst. Feeling worst is usually followed by feeling better, whether or not you took anything.
This is exactly why clinical trials exist and why they are expensive. You cannot get around it with a notes app.
What a real answer would need
Four things, and if an app does not have all four it is not measuring effect.
Continuous objective data, not how you remember feeling. Sleep, HRV, resting heart rate, temperature, activity, and where available, bloodwork.
Your own baseline before you started, so there is a genuine before to compare against. A population average cannot serve as your before.
Enough observations to separate signal from noise, and an honest statement of how many there were. An effect claimed on six days is not an effect.
A willingness to return no result. Any system that always finds an effect is not testing anything.
How to do this yourself, imperfectly but honestly
You can approximate it without any app at all, and it is worth knowing how.
Change one thing at a time and give it long enough to show. Write down, before you start, exactly which two or three numbers you expect to move and by roughly how much. Then stop taking it for a period and see whether the numbers move back. That last step is the one almost nobody does, and it is the one that carries most of the information.
It is laborious and most people will not sustain it. But it is the honest version, and if you only ever do it for the two most expensive things in your cabinet, you will still be ahead.
How NUVARD approaches it
I built NUVARD, so read this as the interested party's account.
NUVARD tests what you take against your own record and reports the result with the effect size and the number of observations behind it. It reads more than 300 devices and apps into one model across 15 body systems, with direct integrations to Oura, WHOOP, Garmin, Withings, Fitbit, Ultrahuman, Polar, Amazfit and Hume, so the before and after are drawn from continuous data rather than from memory.
The Learning Record has three states. Testing, confirmed, and no signal.
No signal means we spent weeks checking whether something you take does anything for you, and the answer was no. It is reported as a finding, not hidden as a gap.
I know what that costs. An app that tells you half your cabinet is doing nothing is arguing against the industry that funds most of this category, and the safe version of this product simply does not ship that screen. Anyone can build the feature. Very few can afford to show you the result.
The honest limitations. This needs time and data. A supplement started yesterday will read as testing for a while, and it should. Effects smaller than your day-to-day variation may never separate from noise, and NUVARD will say testing rather than invent a number. And none of it is medical advice.
The one question to ask any app in this category
Has it ever told a user that something was not working.
If the answer is no, it is reviewing bottles or logging doses. Both are useful. Neither is what you were actually asking.
NUVARD provides wellness intelligence. It does not diagnose, treat, cure or prevent any disease, and it does not replace medical care. Talk to a clinician before starting or stopping anything.