iCIMS resume parsing: an explainer, not a specification

This is the shortest and least satisfying of these pages, and it is short on purpose. iCIMS publishes a clear explanation of what parsing is for. It does not publish how its parser behaves, and pretending otherwise would be the least useful thing this guide could do.

What iCIMS actually publishes

iCIMS's own explanation of the feature defines the job it does and gives the reason it exists:

"CV or resume parsing is the technology that extracts data, usually from a Word or PDF document, and converts it in a structured way, often self-completing a form. It automatically identifies important, relevant terms in a CV such as a candidate's contact details, language level, years of experience, etc."

"It's simpler: the candidate drag & drops or submits his or her CV and CV parsing does the rest of the job. It's faster: the candidate saves time and applies in one click."

iCIMS, What is CV/Resume parsing?

Two things in that text are worth pulling out. First, Word or PDF is the documented input. Second, the stated goal is the candidate's convenience — fewer clicks, no retyping — not the employer's. The article frames parsing as a form-filling feature, which is a fairer description of iCIMS than "screening engine".

Note also what it is not claiming: there is no mention of ranking, scoring, matching to a requisition, or rejecting candidates. On iCIMS, parsing is an intake convenience. Any evaluation of your resume happens after a human has read the form.

The part that makes iCIMS different from the other five

iCIMS does not document parser behaviour, and it does not have to, because the behaviour is set by the customer rather than by the product. What varies from one iCIMS career site to the next is:

• Which fields exist in the application form. A government contractor's iCIMS form has clearance, citizenship and security-investigation fields. A retailer's has store number and availability. The parser has a different job on each.
• Which fields are prefilled and which are not. Recruiters configure this per job.
• Whether a candidate can correct a prefilled value before submitting, and how easy it is to find the field.
• Whether the employer has parsing switched on at all for a given career site.

This is the same structural problem as Oracle Taleo, with a different cause: Taleo's parsing is a partner service, iCIMS's is a configured feature. In both cases the honest conclusion is that there is no single "iCIMS format" to optimise for, because there is no single iCIMS behaviour.

The published facts, for what they are worth

Where iCIMS pages come from

iCIMS maintains two public documentation surfaces and it is worth knowing which one you are reading, because they differ enormously in specificity.

Neither surface is a parsing specification for a job applicant. That is not an oversight to be fixed by a better search; it is a consequence of the feature being customer-configured.

What to do when there is no spec to design against

This is the situation the other five pages do not put you in, so the advice here is deliberately general. It is the union of what the four vendors that do publish have in common, which is the closest available proxy:

1. Export a text-based PDF or a DOCX. iCIMS names both, and they are the two formats the other vendors document most precisely.
2. One column. Greenhouse names columned layouts as a failure mode and Workday's guidelines say to avoid column formats. There is no evidence iCIMS is different, and every reason to think the same text-extraction problem applies.
3. No tables, text boxes, images or logos. Same reasoning.
4. Standard upper-case section headings, so that whatever fields the employer's form maps to, your content is where they expect it.
5. Name at the top, in initial capitals, with your email and phone on the lines immediately below. "Contact details" is one of the three things iCIMS says it identifies, so make them impossible to miss.
6. Spell out employers and job titles, and write your years of experience in numerals with context ("8 years", "since 2018"), since iCIMS explicitly names "years of experience" as something it identifies.
7. Review every field on the iCIMS form before you submit, especially any that arrived pre-filled. On a configurable form, a wrong pre-fill is invisible unless you look.

The print-safe template satisfies all seven and adds no decoration for a parser to trip over. The compact template also satisfies them and is worth a look if you are applying to a large employer with a long, form-heavy iCIMS process and want the resume to be as short as possible.

A word on the folk claims

You will see online that iCIMS scores resumes, that it rejects on keywords, or that it has a known list of parse failures. None of that is in anything iCIMS publishes. iCIMS's own description of parsing is an intake feature that saves candidates typing. If you want to reason about iCIMS, reason from the form in front of you: that is the part of the process that is real, configured, and in your control.

• iCIMS sources What is CV/Resume parsing? — iCIMS's own definition of the feature, quoted in full above
  iCIMS Community knowledge base — the administrator-facing documentation surface
• Cross-references on this site Greenhouse for the published failure-mode list · Workday for the published formatting guidelines · Oracle Taleo for the other customer-configured system