In this guide 7 sections
“The ATS rejected me” can feel like the only explanation when an application disappears into a system. But it combines several different processes into one story. A file can be hard to read. An answer can trigger a screening rule. A tool can help a recruiter compare applications. Those are not the same event.
Understanding the distinction gives you something practical to check. It does not let you see an employer's private process or establish why a particular application was unsuccessful.
Parsing: getting information out of the document
Parsing means extracting information from a CV into fields. In Greenhouse's documentation, an unsuccessful parse can leave the CV attached without automatically filling the candidate details. The vendor identifies images, complex tables and headers or footers among possible causes of parsing problems. That is evidence about a particular product, not proof that every two-column CV is rejected everywhere. [1]
For your next application, use the file format the employer requests. Check that the exported file contains selectable text. Copying a paragraph into a plain-text editor can reveal missing characters or a strange reading order, but it is only a basic check, not an ATS simulator.
When the form shows information extracted from your CV, review it. Look at dates, job titles, contact details and qualifications rather than assuming a successful upload means every field is correct. Greenhouse explicitly advises candidates to check and edit parsed information before submitting. [2]
Screening rules: answers can matter separately from the CV
Greenhouse also documents an auto-rejection feature based on answers to configured application questions. The employer sets the rule; it is a different mechanism from a failed parse. The documentation demonstrates that automated rejection can exist, so “an ATS never rejects anyone” is not a safe general statement. [3]
The practical response is to read the questions carefully and answer honestly. A CV redesign will not change whether you hold a required qualification. If the wording is ambiguous, use any available clarification or accessibility route rather than changing an answer merely to avoid a filter.
A fast rejection does not, by itself, reveal which rule or person made the decision. Treat the timing as an observation, not a diagnosis.
Matching: a score may exist, but context matters
It would also be misleading to claim that no ATS ever assigns a match score. Greenhouse's candidate FAQ describes an optional Talent Matching feature that compares applications with job requirements and assigns a score. It says that this assistive feature does not itself automatically advance or reject candidates. Employers choose whether to use it for a job. That description should not be generalised to every vendor or configuration. [2]
The useful conclusion is narrower than either popular story: one score from an outside CV-checking service does not prove how a specific employer will assess you. Ask what that service measures before treating its percentage as a hiring forecast.
Make truthful relevance easy to find
If a vacancy asks for supplier coordination and you have done that work, describe it plainly and show an example. If your organisation used an unfamiliar label for the same activity, explain the connection. Do not claim a qualification, software skill or job title you do not have.
For example, a fictional applicant could replace “supported business operations” with “coordinated delivery updates between the warehouse and two suppliers.” This is an edit for clarity, not a promise to beat a ranking system.
Avoid hiding extra words in the file or pasting the entire vacancy into it. A useful application should remain understandable when another person reads it normally.
Diagnose what you can actually observe
Consider three fictional applicants. The first uploads a CV and sees an old employer name in the current-role field. That is a visible data problem. Their next step is to correct the field where the form permits it and check the rest of the extracted information. Buying a keyword score would not fix the incorrect field they can already see.
The second sees accurate fields but is asked whether they hold a qualification they do not have. That is a requirement question, not a formatting problem. They should check the wording and answer accurately. If the question appears inconsistent with the vacancy, a published contact route may help clarify it. Changing the CV's font does not resolve the underlying mismatch.
The third sees a successful submission message and later receives a rejection with no explanation. They cannot infer from that sequence alone whether a person, a configured rule, a matching tool or another part of the process influenced the result. Their useful actions are narrower: keep the submitted version, compare its relevance with the vacancy, and request feedback if an appropriate route exists. None establishes the hidden reason for the decision.
These examples illustrate a practical distinction: fix a visible error, clarify an ambiguous requirement, and treat an unexplained outcome as unexplained. That is less dramatic than a promise to defeat an algorithm, but it directs effort towards something you can check.
Evaluate a CV-checking service before trusting its score
Ask what the service is comparing. Is it checking text extraction, comparing words with a vacancy, assessing writing clarity or attempting to predict employer behaviour? Those are different tasks. A percentage without a defined measure is not a useful basis for deciding whether your experience is competitive.
Look at the explanation beside the score. “The exported document did not expose readable text” points to a problem you can investigate. “Add leadership five more times” needs a much better justification. A suggestion should help you describe truthful experience clearly, not encourage repetition or imply qualifications you do not hold.
Finally, consider what you are sharing. A CV can contain contact details and information about previous employers. Read the service's privacy information and use only a service you are comfortable entrusting with that document. You should not need to disclose confidential customer names or internal project details to obtain writing feedback.
Your pre-submission check
Open the final exported file, not only the editable version. Check its text, headings and dates. Review any automatically filled fields. Read screening questions separately. Confirm that your examples match the actual job and remain true.
Then save a private copy of what you submitted and the vacancy. If you later receive feedback, you will have something concrete to compare. If you receive no explanation, resist turning an unknown outcome into certainty about a hidden algorithm.
Product features can change. The vendor documentation above was checked on 5 September 2026; it illustrates specific capabilities rather than certifying every employer's process.
Sources and further reading
Optional source checks. External sources open in a new tab.