Let's say you want to track someone through the 19th century or early 20th century in the United States. The Federal Census will give you, once a decade, a snapshot. They are available for free (we're not going to talk about the 1890 census because I don't want to cry) online through familysearch.org or for pay at ancestry.com, which many public libraries subscribe to. Both services include a transcription and an index built on top of that transcription. Ideally, the person you are tracking has a somewhat unusual but easy to spell last name. If the name is too unusual or difficult to spell, the census taker may have gotten it wrong or the transcriber or whatever, thus causing it to fail to appear on a search and forcing you to read through the census line by line. Which can be awful or impossible unless _you already know where they live_.
Now, if you have good enough family history to know where they lived in, say, 1920 or whatever, well, good for you! But if you don't, you _might_ do a little better with the City Directory for their town. It will be alphabetical AND there is often a street listing. Best of all, these have few to no errors in the spelling of names. Downside: they won't list the whole family. Further, during the early 20th century, names weren't necessarily all that stable, as relatively recent immigrants experimented with how best to present their name as they assimilated. Thus: my great uncle Hein became Harry and my grandfather Simon became Sam.
At least Dutch and English are pretty closely related, with only a few phonemes unrepresented in each other's language. Not so with, say, central European names. On top of unstable first names (Orlam, Harlam, Harlan -- oh, and I left out "Karp", which seems to be the short form of the man's middle name which was a patronymic, Karpov), there can also be phonemes in the original language that transcribe as multiple letters in English. And that transcription didn't stabilize for a while. So there are individual letters in the original name that wind up transcribed as dje or dze or she or z, you get the idea.
Between the WW1 draft (I love the WW1 draft, because damn near every adult male in the country was tracked down and had a card filled out for him during that draft; WW2 was almost as comprehensive) and the city directory, I am quite certain that Orlam (Harlam, Harlan) is somehow related to Vasia (Wasilia, Wasilisa, Wasilena -- hey, I know about Russian names; you can't surprise me with the variety there), because he says he is (WW1 draft card) and lists her as a closest relative (ditto) and the address he gives on the card is the same as her husband's on the 1918 city directory.
But beats the hell out of me _how_ they are related. I've got a findagrave entry in the same town (with a picture of the headstone) that says he died in 1960, and I'm so far not finding him in any city directory or any census. I'm sure he's there. I just gotta keep looking and hope at some point, I find him under whatever name variation he is listed.
ETA: He shows up as Harlam in two Ansonia city directories (1931 and 1932). But weirdly, NOT in the street part of the directory, only in the name listing. I sure hope there's a story here. Maybe he moved out of town and then was buried back in Ansonia after a long life elsewhere.