About

Why we measure baby names by sound, not spelling

Most charts rank spellings. Namesonder ranks names as you'd hear them — so Aiden, Ayden, and Aaden count as one popular name, not three small ones.

The test: yell it across a crowded park — who turns around? Everyone who'd turn for "AY-den" is one group, however it's spelled. That's the unit the whole site is built on.

Split apart on a spelling chart, each version of a name looks rarer than it really is. Grouped by sound, a name shows its real popularity.

What counts as the same name?

That's partly your call. “Same name” isn't one fixed line — Aiden and Ayden are obviously one name, but Ana and Anna? A four-way control on the chart sets how loosely sound-alikes merge; each step to the right combines additional sounds:

Spelling

No merging; every spelling stands on its own. (Jackson ≠ Jaxon)

Same Sound

Spellings said identically. (Aiden = Ayden, Catherine = Kathryn) — no one would hear a difference.

Sounds Similar (the default)

Sound the same in ordinary speech; you'd only hear a difference if someone said them deliberately. (Cody = Coty, Don = Dawn)

Sound Family

Distinct but close, treated as one name. (Ana = Anna, Tamara = Tamera) — you can hear the difference; it's still one name, loosely.

The default is the middle setting, Sounds Similar — close enough that most people wouldn't notice the difference in everyday speech. However loose you go, genuinely different names stay apart — Maria ≠ Mariah, Charles ≠ Carlos. It's an automatic rule, not a hand-picked list, so a few odd pairings slip through; hover any name on the chart to see exactly which spellings it includes.

Where does the data come from?

U.S. Social Security Administration national baby-name records, 1880–2025. Namesonder groups those names by sound one spelling at a time — each pronunciation is looked up individually — so the chart doesn't yet cover every name in that database. It currently covers the most common names, and more are added as the work continues.

How is popularity measured?

The default is share — births per 100,000 that year — because raw counts mislead: far more births are recorded today than in 1900, so a raw-count line mostly tracks population growth, not a name's popularity. Share asks the real question: of all babies born that year, how many got this name? You can switch to raw count or rank anytime.

Which pronunciation is the main one?

These names come from the U.S. Social Security database, so Namesonder reads each one in American English, and the American English pronunciation is the primary one on record for every spelling.

This is a consistency choice for the data, not a ranking of languages. Measuring every name against one yardstick is what lets spellings from every origin — Spanish, Irish, Hebrew, Vietnamese — sit on the same chart and be compared fairly; mixing pronunciation systems would make the groups inconsistent and the counts unreliable. Other pronunciations are no less valid: where a name is also commonly said another way, that reading is recorded too, as an alternate that feeds the shaded band (below), so it still shapes the group's range.

What does the shaded band mean?

Some spellings could belong to more than one sound group — is Louis said LOO-is or LOO-ee? The source records spelling and count but never pronunciation, so a group's size is shown honestly as a range:

• The solid line is the floor — only the spellings that unambiguously belong.
• The shaded band above it, up to the dashed edge, is the ceiling — the most the group could reach if every ambiguous spelling counted here.

A thick band means a lot of the size is genuinely uncertain; no band means every spelling is clear. (Shown for share and count, not rank.)

The looser the grouping you pick, the thinner this band — coarser tiers fold most ambiguous spellings into one sound, so there are fewer spellings left to contribute to the band.

How does each spelling get its sound?

Spelling is an unreliable guide to sound — the letters in Phoebe barely hint that it's said FEE-bee. So rather than guess from the spelling, each name's real pronunciation is looked up and recorded as a short respelling, and every group is built from those sounds.

When a spelling is genuinely said more than one way, we don't try to guess how many babies use each version — the records don't tell us. Rather than invent a split, we keep both readings: one anchors the group, the other feeds the band. Because the band is that one spelling's own uncertainty, it shows up on the other group — the one it might also belong to — not on its home group.