About this site
Margaret Ellison
Writes and tests everything here. Grand Rapids, Michigan. Former market research analyst. Not a clinician.
I started this site because of a bottle of melatonin I had been taking wrong for six months.
I was 47 when my sleep came apart. Not trouble falling asleep, which I had never had, but waking at three in the morning wide awake and staying that way until the alarm. My doctor ruled out the things worth ruling out, thyroid and sleep apnea among them, and what was left was perimenopause and a shrug. There was no prescription for it. So I did what everyone does and went to the supplement aisle, where I bought the melatonin that was on the shelf at the dose that was on the shelf, which was 10 mg.
I felt awful every morning for half a year before I read a single study on it. When I finally did, I found something that genuinely made me angry: an MIT trial had found that 0.3 mg worked better than 3 mg for exactly my kind of age-related insomnia. I had been taking more than thirty times the dose the research supported, and nobody, at any point, in any aisle or on any label, had told me that.
That gap between what the studies say and what the shelf sells is the entire reason this site exists.
Why I am qualified to read a study, and why that is all I claim
I spent eighteen years as a market research analyst. The job was reading other people's research and figuring out how much of it was real: sample sizes, who funded it, whether the conclusion actually followed from the data, and what the press release quietly left out. It is not a medical training. It is a professional habit of not believing a number until I have seen how it was produced.
That is the only expertise I claim here. I am not a doctor, a nutritionist, a sleep coach, or a certified anything. I do not diagnose and I do not prescribe. When a question needs a clinician, I say so, and there is a version of that sentence in nearly every guide on this site.
How I actually test things
I run protocols on myself, one variable at a time, and I write down what happens before I decide how I feel about it. Sleep memory is unreliable, and mine is no better than anyone's.
- One change at a time, for at least fourteen days. The trials that found effects ran seven to fourteen days. Anything shorter tells me nothing and anything with two variables tells me less.
- Three numbers a night, in a spreadsheet. Time in bed, times I remember waking, and a one-to-five score for how I felt at mid-morning rather than on waking. Mid-morning is harder to fool yourself about.
- I report the failures. Tart cherry juice did very little for me at the two-week mark. The concentrated version did more, and I still do not know whether that was the format or the season.
A note on the cherry, since it comes up: I grew up in Michigan, which grows most of the Montmorency crop in this country. It was a fruit from the farm down the road long before it was a capsule with a marketing budget. That does not make it work better. It does mean I was skeptical of the price of it before I was skeptical of the science.
How I read the evidence
Sleep research is full of small studies with mixed results, and the honest position is usually somewhere short of a recommendation. Three rules I hold to:
- Negative trials get the same billing as positive ones. A 2024 trial found that a commonly sold dose of tart cherry powder did nothing for sleep or inflammation. It is in the guides. It should be.
- I check the citation, not the claim. Sales pages routinely cite real studies for conclusions those studies did not reach, and sometimes get the journal wrong entirely. I have caught that more than once.
- Where the evidence is thin, the sentence says so. "There is no good evidence either way" is a legitimate answer and I would rather publish it than a confident guess.
How this site makes money
Some links here are affiliate links. If you buy a product after clicking one, I may earn a commission, and it costs you nothing extra.
What that does not buy: a good review. No company has paid me to write anything, nobody sees a guide before it goes up, and no product is described as better than the evidence supports because there is a link attached to it. The clearest proof I can offer is that the guides here criticize the sales presentation of the product they link to, tell you which claims to disregard, and point out where a cheaper option in a grocery store does the same job.
If that ever stops being true, the site is not worth reading and I would rather you knew how to check.
Getting in touch
Corrections are welcome and I would rather hear about an error than not. If a study is misread or a number is wrong on any page here, write to Galaxy-trade2000@gmail.com.