Percentile benchmarking compares your value on a health marker against a real reference population, showing where you actually stand relative to others rather than simply confirming whether you fall inside a generic clinical "normal" range. Longevity Coach IQ's benchmarking is built specifically around real, user-supplied reference data, resolving each marker into one of a small number of defined tiers, Top 1%, Top 5%, Top 10%, or, honestly, "below Top 10%" when the source data doesn't support claiming anything more precise.
- Built from real reference data, not derived or invented estimates.
- Resolves each marker to Top 1%, Top 5%, Top 10%, or "below Top 10%."
- Doesn't claim finer precision below Top 10% than the source data actually supports.
- Distinct from a clinical "normal" range, which only indicates absence of disease.
Why only three tiers
Not because finer gradations don't exist in the literature. For several markers they plainly do: the FRIEND registry, for instance, publishes the 5th, 10th, 25th, 50th, 75th, 90th and 95th percentiles for VO2 Max by age and sex.
The reason is that tier boundaries do not all come from the same kind of source. Some are lifted from a published guideline. Some reflect clinical consensus without a single authoritative table behind them. A few are this app's own estimate, arrived at by interpolation.
Reporting a single continuous percentile across markers of unequal provenance would imply a uniformity that does not exist. It would present a number derived from a landmark paper and a number we worked out ourselves in identical typography, and the reader would have no way to tell them apart.
Three tiers is the resolution the scoring is designed around, and it is deliberately coarse enough not to overstate what the weakest source can support.
This is a deliberate choice to stay within the actual precision of the underlying evidence rather than manufacturing false confidence, a more honest approach than generating a specific-sounding percentile number that isn't actually backed by the reference data it claims to come from.
Where each threshold comes from
Every marker the app scores carries a provenance label, and the benchmark sources page lists all of them with the reasoning for each.
There are three:
- Published guideline. The threshold is taken directly from a named guideline or reference dataset. ApoB against the ESC/EAS targets is one of these.
- Clinical consensus. Widely used cut-offs that are not traceable to one authoritative table, but are consistent across sources.
- App estimate. The underlying quality is well studied; the specific tier boundaries are ours. Four markers sit here, and the label exists so you are not left guessing which.
Some markers are mixed cases, where the headline threshold is a guideline value but the sub-splits between tiers are our interpolation. Those carry a note saying so.
The point of the labels is the point of this whole page. The precision claimed should match the precision available, and the only way a reader can check that is if we say where each number came from.
Percentile vs. clinical "normal"
A clinical "normal" range is built to flag the presence or absence of disease. That is a different question from where a value stands relative to other healthy people. A marker can sit comfortably inside a clinical normal range and still be far from the Top 10% of the reference population. Both facts can be true at once, and both are useful, for different purposes, one for ruling out pathology, the other for understanding where there's still room to optimize.
From a raw number to a score: a worked example
The clearest way to explain percentile benchmarking is to follow one measurement through the whole process. Take a 45-year-old man with an ApoB of 88 mg/dL.
ApoB is the right marker to demonstrate this with, because its thresholds come from a published guideline rather than from our own interpolation. Not every marker does, which is the subject of the section below.
Step 1: find the right reference band
The raw number means little on its own. The 2019 ESC/EAS dyslipidaemia guideline sets ApoB targets by risk category: below 100 mg/dL at moderate risk, below 80 at high risk, below 65 at very high risk.
This is also where the honest limitations live. Reference tables are built from finite samples, and their composition affects every number derived from them.
Step 2: locate the value within the band
At 88 mg/dL he sits inside the moderate-risk target of 100 but above the high-risk target of 80. Eight points would move him past the next threshold.
That gap is the actionable part, and it is why the app reports the next threshold rather than only the current position. A number with nowhere to go is a worse number.
Step 3: convert the tier to a score
Tiers map to fixed scores rather than to a continuous scale: Top 10 percent scores 90, Top 5 percent scores 95, Top 1 percent scores 99. Three tiers rather than a precise percentile is a deliberate concession to the precision the underlying data can support.
Step 4: weight it inside its domain
The marker score is combined with the other markers in its domain, weighted by how strong the published evidence linking each one to outcomes is. ApoB carries substantial weight inside Cardiovascular Health for that reason.
Step 5: combine domains into the overall score
Domain scores combine into the final 0-100 figure, with domain weights that always sum to 100 and that shift depending on how many domains your tracking mode covers.
Why percentile and clinical normal disagree
A clinical reference range answers whether you are ill. A percentile answers where you stand among people like you. The two frequently disagree, and fasting insulin is the clearest case. Values well inside a standard laboratory range can sit outside the optimised band.
The reason is that a laboratory range describes the spread of results in the population that was tested, and metabolic dysfunction is common enough in that population to widen it. Normal, in that context, describes what is common rather than what is good.
Percentile benchmark vs. clinical range
| Clinical "Normal" Range | Percentile Benchmark | |
|---|---|---|
| Answers | Is disease present? | Where do I stand vs. others? |
| Typical width | Broad | Narrower, tiered (Top 1%/5%/10%) |
| Useful for | Ruling out pathology | Identifying room to optimize |
When a percentile is honest
Consumer health content presents an exact percentile for every marker regardless of whether the reference data supports that precision.
Where good population data exists, a percentile is the most useful way to read a number. Where it doesn’t, an invented percentile is worse than a broad tier.
That is why some markers here get three tiers rather than a running number. The precision claimed should match the precision available.
In the app, the benchmarks come from real reference data, honestly capped where the evidence stops.
Sources
Key references for the claims on this page. Where a figure is attributed to a specific study or body, it is named here.
- This entry describes Longevity Coach IQ's own scoring method rather than an external finding. The underlying percentile reference tables are cited in the individual marker entries, and the full weighting is set out in the methodology.
Frequently asked
What does percentile benchmarking mean?
Comparing your value against a real reference population to see where you stand, distinct from a clinical normal range.
Why does it sometimes just say "below Top 10%"?
The source data only defines Top 1%/5%/10% tiers, claiming more precision below that would be false precision.
Is percentile the same as clinical "normal"?
No, a normal range indicates absence of disease, percentile shows how you compare to others.
Which population are the percentiles drawn from?
Published reference datasets for your age band and sex, cited in the relevant entries. This matters because those datasets carry the characteristics of whoever was studied. Where the underlying population skews toward one country or one health profile, the percentile inherits that skew, and the methodology page says so plainly.
Does Top 10% mean the top 10% of everyone, or of healthy people?
Of the reference population, which includes people across the health spectrum. That is worth sitting with: being in the top decile of a population that is largely unfit is a lower bar than it sounds. It is a relative position, not a certificate of health.
Why not show an exact percentile number?
Because a figure like 73rd percentile implies a precision the underlying data can’t support. Reference tables are built from finite samples with real measurement error. Three tiers communicate roughly where you sit without dressing up an estimate as a reading.
Can I be Top 10% and still not be healthy?
Yes. Percentile is relative to other people, not to a physiological ideal. You can sit in the top decile on one domain while a different marker is genuinely poor, which is precisely why the score reports six domains separately rather than collapsing everything into one number.
Do my percentiles change as I get older?
Yes, because the benchmarks are age-banded. Holding a raw value steady while moving into an older band will generally improve your percentile, since the comparison group has declined. Your absolute capacity hasn’t improved, and the app treats those two situations as different.