What a Service's Numbers Really Tell You (and What They Quietly Don't)

I have a bad habit I'm only half-cured of. When I'm checking out something new online — a service, an app, a person I might pay for an hour of their time — my eyes go straight to the numbers. Five years in business. Ten thousand happy customers. 4.8 out of 5. Something in me relaxes the moment I see them, the way you relax when a waiter says a dish is popular. And then, more than once, I've handed over money on the strength of a statistic I never actually thought about.

So I want to talk about those numbers — the ones services put front and center — and what they're really worth. Not because they're lies. Most of them aren't. But a number on a homepage is a headline, and headlines are written to be believed quickly. Slow down on them and they start saying something more honest, and usually more modest, than they first appear.

"We've been doing this for X years"

Longevity is the stat I trust most, and even it has limits. A service that's been running for years has, at minimum, survived. It hasn't vanished overnight, it has presumably dealt with unhappy customers and lived to tell the tale, and there's a track record somewhere to look at. That's real. It's worth something.

But "since 2020" tells you the doors have stayed open, not that they've stayed open for good reasons. Plenty of mediocre things last a long time. And a brand-new service isn't automatically worse — everyone starts at zero, and some of the best people I've worked with were early enough that they had almost no history to point to. Treat age as a floor, not a verdict. It rules out the fly-by-night operation. It doesn't promise you a good experience.

The big round count — "10,000 readings," "50,000 customers"

This is the number I've learned to enjoy the least, because it's the easiest to inflate and the hardest to check. A large total volume mostly tells you a service is busy, which is genuinely useful — busy usually means it's not a one-person operation that'll disappear when life gets in the way. But volume says nothing about whether those thousands of people were satisfied, or whether they came back. A number that only ever goes up, with no sense of how many people returned versus tried it once and left, is a vanity metric wearing a suit.

When I see a count like that now, I mentally add the question the number is hoping I won't ask: and then what? Of those ten thousand, how many would do it again? That second number is the one that matters, and it's almost never the one on display.

The rating, and why a 4.8 deserves a second look

A high average rating is reassuring, and it should be — to a point. But I've come around to trusting a 4.6 with visible, specific, occasionally lukewarm reviews far more than a glossy 4.9 made of one-word praise. Real human experiences are uneven. A service used by thousands of people will, just by the math, let someone down now and then. A rating with a little friction in it looks like life. A rating with none looks managed.

The number itself is also only as good as who was allowed to create it. A 4.8 built from verified clients who actually completed a session means something. A 4.8 from an open box anyone can post in means considerably less. Some services are built so only real, completed transactions can leave a rating — Astroideal is one I've seen do this — and that structure is worth looking for wherever you book, not because it makes the score perfect, but because it makes it harder to fake. The figure matters less than the rules behind it.

So how do you actually read a wall of stats? I do three quick things. I ask what each number leaves out — a total without a return rate, a rating without its rules. I look for the unflattering detail, because a service willing to show a three-star review or admit a limitation is usually the more honest one. And I refuse to let a cluster of impressive figures stand in for the thing I actually care about: what it's like to be a real customer, in real time.

What the numbers can never do

Here's the part worth saying plainly. Statistics describe a crowd. You are not a crowd. Ten thousand satisfied people tell you the odds are decent; they do not tell you that your hour, your question, your particular fit with this service will land well. Aggregate happiness is a useful signal and a terrible guarantee.

And in fields built on reflection rather than fact — tarot, for one — no quantity of glowing numbers makes the thing predictive. A high rating means people felt heard. It doesn't mean the cards forecast the future, and it's no substitute for a doctor, a lawyer, or a financial professional. The numbers measure satisfaction, which is real and worth measuring. They just can't measure certainty, because certainty was never on the table.

A couple of quick ones, since they come up. Do more reviews always mean more trustworthy? Not on their own — a thousand vague raves can be weaker evidence than fifty detailed ones. Should a perfect score reassure me? Slightly less than an almost-perfect one with some honest dissent in it. Is a service with few numbers a red flag? No — just lean harder on what those few numbers actually say.

In the end, I still look at the numbers first. I'm not sure I'll ever stop. But I've stopped letting them do my thinking for me. A figure on a page is the beginning of a question, not the answer to one — and the services worth trusting tend to be the ones whose numbers invite the question rather than try to end it.

Updated: June 2026