Reading 105,000 popularity curves: what the data quietly shows
If you look at one name’s popularity over time, you see a noisy line. If you look at a hundred thousand of them at once, you see shapes.
Sit with the US Social Security Administration name data — every baby name registered in the country, by year, from 1880 to the most recent published year — long enough and the recurring curve shapes start jumping out. There are four of them. They show up over and over, across origins, across genders, across centuries. Almost every name in the catalog fits one of the four, and the shape of a name’s curve is more predictive of its near-future trajectory than almost anything else you could ask about it.
Here are the four, what they look like, and what they imply for parents choosing today.
The stable classic
The first curve shape is essentially flat. The name appears in the top several hundred ranks in 1900, and it appears in the top several hundred ranks in the most recent year, and it has done so continuously in between, with small bumps and dips but no peak or trough that defines an era.
James, Michael, William, John, Anna, Sarah, Elizabeth, Emma — these are the canonical examples. They are not the most popular names in any given year, but they are never far outside the top hundred. Their curves are gentle waves on a long horizontal line.
For parents choosing today, a stable classic is the lowest-risk pick. It will not feel of-the-moment, but it also will not feel dated. It is about as fashion-resistant as a name can be. The cost is that it is, by definition, common. Your child will not be the only one in the class. They will also not be the only one in any class for the rest of their life.
The slow burn
The second curve looks like a steady ramp. The name was rare in 1900, has climbed gradually for decades, and is now meaningfully popular but still climbing. The slope is shallow enough that no single decade owns the name, and the climb is unbroken enough that it doesn’t feel like a fashion peak.
Olivia is the cleanest contemporary example. It was rare in the 1970s, climbing through the 1990s, in the top 20 through the 2000s, and at the top of the list now. It has not peaked. There is no sharp inflection in the curve. It will eventually plateau or decline, but the descent — if our reading of comparable curves holds — will be slow.
A slow burn is the most reliably "modern but durable" pick available. It is a name that signals the current era without being defined by it. The risk is that some slow burns, like Madison in the early 2000s, eventually inflect upward into a hard peak that does become time-stamped. The tell is whether the slope is accelerating in the most recent five years. If it is, the name is approaching the boom-and-bust profile rather than continuing the slow burn.
The boom and bust
The third curve is the comet. The name was nowhere — outside the top 500, often outside the top 1,000 — until a sharp inflection point at which it climbed steeply into the top 50 or top 20 within five to ten years. It peaked for a brief plateau, and then started coming down at roughly the same slope it climbed.
The boom-and-bust pattern is the most strongly era-coded shape in the data. Jennifer in the 1970s. Brittany and Ashley in the 1980s. Madison in the early 2000s. Aiden in the late 2000s. Each of these names is, in a single glance, readable as the year cohort that named it.
If a name has hit boom-and-bust, the question for parents is whether they want the era association. Some bearers love being unambiguously of their cohort. Some don’t. The data is silent on the right answer; it only tells you which question you’re actually being asked when you look at the curve.
The vintage revival
The fourth curve has two humps. The name was popular in a historical era — usually the early 20th century — declined to near-zero through the middle of the century, and is now climbing back. The two humps may be of different sizes; the second hump is usually still in progress.
Eleanor, Hazel, Ezra, Theodore, Beatrice — these are the contemporary canonical examples. They were popular when our great-grandparents were named, then fell out of fashion for two or three generations, and are now rising sharply among parents who find them charming precisely because they sound un-current.
Vintage revivals are interesting because they sit in a structurally specific cultural slot. They feel like a deliberate aesthetic choice rather than a fashion-following one. The risk is that vintage revival names cluster. If you pick a name from this group, your child will share a cohort signature with the other vintage-revival children of the same year, which is itself a kind of era-coding. The cost is lower than boom-and-bust era-coding, but it isn’t zero.
What the curves don’t tell you
The data is excellent at describing where a name has been. It is mediocre at predicting where a name is going. Slope and curvature are useful, but the inflections that turn a slow burn into a boom-and-bust, or that begin a vintage revival, are driven by cultural events — a popular TV character, a celebrity child, a viral moment — that are not in the data.
So the curves are a baseline, not a forecast. They tell you the shape your name has had so far. They suggest, statistically, what the most likely next several years will look like. They cannot guarantee that your child’s name will not be propelled into a fashion peak by something none of us see coming.
A practical reading
If you’re looking at a shortlist of candidate names and you’ve plotted each one’s curve — which you can do on NameMatch’s individual name pages — here is roughly what to look for.
A stable classic curve says: low risk, low novelty, durable.
A slow burn says: modern, currently durable, watch the most recent slope.
A boom and bust says: era-defining if peaked recently, look at where the curve sits now.
A vintage revival says: distinctive in the moment, era-coded as part of a cohort.
None of these is the right answer. They are different answers to different questions. The reason it’s worth looking at the curve at all is that the question your name is implicitly answering — "what kind of era signal do I want this to send?" — is one you should be answering deliberately rather than by accident.
The data is patient. It will wait while you decide.