When to use a two-column resume layout

A one-column resume is not the default because single columns look plainer. It is the default because it is the only layout a person and a piece of software agree to read in the same order. Everything below is either an argument about when that agreement is not needed, or a way of getting the look without the disagreement.

Why people want a sidebar

It is worth naming the real reasons, because each one has a single-column answer and the reasons are usually good. People reach for two columns because a narrow rail down the left looks considered, and on a page of plain text it genuinely is the difference between a document and a form. Because the skills list has nowhere else to go once the experience section has filled a page, and a paragraph of comma-separated tools looks like the thing it is. Because right-aligning dates against a wide single column leaves a lot of white space in the middle of the line, which some people read as unfinished.

None of those are silly reasons. Two of them are solved better in one column, and the third is a preference that you are entitled to hold. The problem is not the taste. It is that the layout decision is also a decision about how your file will be read by software, and that part is not a preference at all.

What the second column does to the text

A PDF is a set of instructions for drawing a page. To get anything out of it, something has to decide the order the lines were meant to be read in, and a two-column page removes the cue that decides it. A reader uses position: left before right, top before bottom. That cue still exists, but it now points in two directions at once, and it is the parser's problem rather than yours.

The two vendors that publish failure lists both name it. Greenhouse lists "resumes that have a columned layout" among the formatting that causes a partial parse. Workday's published guidelines say to "avoid tabular and column formats." Greenhouse is also specific that a parse "might be partially successful" — some fields filled, some skipped — and that this needs manual correction. That is the failure mode to plan for: not a rejection, a half-empty profile.

The same page, laid out two ways

Both versions below contain exactly the same information. Only the geometry differs, and that is the whole argument.

Two columns: a rail on the left, the career on the right

One column: the same information, in the order it was meant to be read

Now consider what a text extractor does with the first version. If it reads down the left rail before crossing to the right, the output opens with an email address, a phone number, a link, two languages and four tools, and your name does not appear at all, because in that version it is not on the page. If it reads across the page a line at a time, the output interleaves "English, Portuguese" with "Supply Chain Operations Manager", and the skills and the jobs become one paragraph.

Both outcomes are plausible, both are outside your control, and neither is visible to you in a PDF viewer. In the second version there is no decision to get wrong: the order on the screen is the order in the file. That is the entire argument for one column, and it is not a matter of taste.

The three cases where it holds up

There are three, and all three share one property: you know who is reading.

One: you know a named person is reading it

A referral, a recruiter you have spoken to, a hiring manager you met at something, a contact emailing you directly. In this case the file is not going into a database; it is going into an inbox and being read by somebody who already has a reason to read it. The documented parse risk does not apply, because there is no parse.

Two honest caveats. First, documents get forwarded. The person you send it to may paste the text into their own system, at which point a column that read well on screen becomes a column in a database. Second, the conversion of this situation is low, which is exactly why the design is safe: you are optimising for a conversation you are already having, not for volume.

Two: it is not a resume

Some documents that share a resume's vocabulary do not share its constraints. An academic or research CV with a rail of publications, teaching, and supervision is a genre with its own long-standing convention, and the sidebar is part of that convention rather than an innovation in it. The same is true of a creative one-pager sent alongside a portfolio, where the page exists to direct attention to the work rather than to be parsed into fields.

The test is whether the document's purpose is to be scanned into a form. If it is not, the usual reason for one column does not apply, and you can design for the reader in front of you.

Three: the second column holds no qualification

This is the narrowest version and the one most people could adopt safely. A true margin note carries contact details, a link, a language pair, and a one-line tools list. Nothing that qualifies you exists only there. The main column stands up as a complete resume on its own, and the rail is a convenience that can be ignored or lost without any consequence.

Notice what has to be true for this to work: the main column must be complete. If the only copy of a skill is in the rail, then the rail is load bearing, and a partial parse loses it. The rail is a margin note or it is not.

The decision rule

Three questions. All three have to be yes. The honest answer to question one is usually no, and that is the intended result.

Two further habits make the rule sharper. Have a one-column version ready, because you will sometimes apply to a large organisation and sometimes to a person, and deciding at send time is much better than deciding at design time. And run the copy-and-paste test from the ATS formatting guide on any file you are genuinely unsure about, which takes two minutes and settles it.

The single-column fix for each problem

Every reason people reach for a sidebar has a better single-column answer. This is the table worth keeping.

The two most common reasons, taken together, produce the layout almost everybody is actually copying. Here is the same person's page with the skills list moved, which is a fifteen-minute change and loses nothing:

Nothing was added, and the reader now meets the skills before they meet the career rather than having to look sideways for them. The rail version is not dishonest. It is simply the version where the tools have no employer, no dates, and no verbs attached, which is the block most likely to come loose when a page is read by something that is not a person.

Single-column alternatives to common reasons for a two-column layout

What you want	The single-column fix
My skills visible at all times	Two or three grouped lines under the header, before the experience: what you use, what you have shipped, what you hold. The reader meets them in the first two seconds, which is more visible than a rail they have to look sideways at
My contact details always on the page	Put them in the header. They are the first thing anyone looks for and the last thing anybody calls you about, and a running footer is a documented parse problem in its own right
Long date ranges without crowding	Right-align dates to a consistent edge. A single column at 34rem of text with a four-character date block is not cramped, it is aligned
The page to look designed	A larger name, one rule under the header, consistent heading positions, generous top margin, and one type family at three sizes. Hierarchy is what reads as design
To fit more on one page	Cut a bullet, not a column. Two columns saves a little vertical space and costs you measure; removing the weakest bullet from your weakest entry saves more and costs nothing
To separate sections visually	Space and a consistent heading style. A rule between sections is enough, and it survives monochrome printing where a colour wash does not
To list a lot of tools	Group them and attach each group to something. Four tools that appear in a bullet you wrote are more credible than fourteen that appear only in a list

Print, paper, and narrow measures

Most applications are now read on a screen, which is the argument people use for the sidebar. Paper is where the compromise becomes impossible to hide.

A rail spends ink on every page to repeat a name and an email address the reader already has from page one. The narrow measure it leaves for the bullets is worse than it sounds: a 34rem single column gives roughly 70 characters a line, which is comfortable, while a sidebar layout leaves a main column closer to 50, which is enough to force a second line in almost every bullet and turn a tidy list into a ragged one. And a resume is genuinely still printed in some sectors, and interviews are frequently conducted on a printed copy, where every colour separator between the columns has already vanished.

None of that is a disqualifier on its own. It is a list of costs that should be weighed against an aesthetic gain, rather than a reason the question cannot arise.

If you are going to do it anyway

Some people will read this and use two columns regardless, and that is a legitimate choice. Here is how to build one that degrades safely rather than one that fails silently.

• Keep every qualification in the main column. Contact, links, a language pair, a short tools line. Nothing else. If a recruiter or parser loses the rail entirely, the resume must still be complete.
• Keep all dates and employers in the main column. The rail is the worst possible place for employment history, because the two most fragile fields are the ones that need to stay attached to each other.
• Make the rail narrow and the main column wide. If a bad extraction order happens, a wide main column survives as a coherent block. A design with two equal columns does not.
• Put the name at the top of the main column, not in the rail. It is the first thing a person looks for and, per Workday's published guidance, the first thing a parser is looking for. Do not put it in the position that is hardest to reach.
• No text boxes, no tables, no background images. All three are named independently as parse failures by the two vendors that publish failure lists, and a two-column layout is a good moment to introduce none of them.
• Keep a one-column version. Different version for a large organisation, same file for a person you know. This is the single most useful thing on the list, because it removes the need to predict.
• Test it before you send it. Open, select all, copy, paste into a plain text editor, read it in order. Two minutes, and it converts the whole question from opinion into evidence.

With those seven in place, a two-column file is not reckless. It is a considered trade: you are accepting a documented parse risk in exchange for a design you prefer, having made sure the consequences of losing the rail are nil. That is a defensible position. Sending a resume with the skills list in the rail, to an organisation you have never spoken to anyone at, is not.

Questions about two columns