CV parsing / CV parser: why and how?

CV parsing definition

CV parsing is a technology that extracts important information from a CV (or candidate profile) and organizes it in a structured format. This format is then used for automated data processing in tools such as :

• ATS (application management)
• CRM (talent relationship management)
• Career sites
• Job boards
• Etc.

The aim is to extract data quickly and automatically to speed up the CV screening and analysis stage.

Which data can be scanned by a CV parser?

When we talk about a CV parser, we are talking about the tool that will be used to parse the CV. Depending on the reference system used and the settings of this tool, the data scanned and extracted can be very large, classified in fairly broad categories. Let's look at a few examples.

Candidate datas

• Last name
• First name
• Place of residence
• Age / seniority
• Telephone number / email address
• Studies / diploma

Candidate's experience data

• Start date / end date of experience
• Company name
• Job title
• Mission description
• Management or not

Candidate’s skills data

• Technical skills
• Tools used
• Foreign languages mastered
• Soft skills

These are just examples, but they show the wide range of data that can then be used in different ways, either by AI or by humans (or both).

How does the CV parser work?

A CV parser scans the content of a document (PDF, Word, raw text, LinkedIn profile, etc.) to extract structured data (identity, skills, experience, education, etc.).

CV parsing technologies analyse letters and words and are based on identified keywords, often grouped by family. It's also possible to identify some of the text's subtleties. Some parsing technologies incorporate grammatical rules specific to each language.

For example, depending on the language detected (in this case English), a good parsing tool will differentiate between :

• communication = as a skill
• communication = as a professional sector / profession

CV parsing and data conversion

Depending on how the tool is configured, the data is converted into different formats, such as JSON or XML, which can be easily used by analysis software. For example, a recruiter can set up his parser to automatically highlight applicants with a particular criterion in their CVs.

The potential limitations of CV parsing

You should be aware that not all parsing technologies are equivalent. And some, more than others, will reach some limits due to artificial intelligence, which can never replace human intelligence 100%.

For some algorithms, reading and extracting data can be more complex on creative, highly designed CVs, particularly because parsing technologies can also use page layout as a way of analyzing data.

For example, on a CV, it's common for the first name and surname to go together. So if the algorithm identifies a first name, it can conclude that the next word is the last name. Some layouts may ignore the surname, which can confuse the tool. As we have seen, not all parsing technologies are on the same level of quality.

And the quality of the algorithm will determine the quality of the information extracted and the subsequent analysis possibilities.

The variations of certain terms in different languages and cultures can also be complex. Take the example of dates, which are not written in the same order in French and English, for example. Once again, the importance of having parsing technology available in several languages is a real asset, especially for international companies.

CV parsing applications in recruitment

Enrich ATS with applications without any manual action from candidates or recruiters: a CV parser must-have

Filling in an ATS with applications can be time-consuming:

• either for the candidate, who is asked to fill in information on extensive application forms,
• or for the recruiter, who is sometimes obliged to do so in order to spare the candidate this task.

CV parsing allows CV elements to be identified and automatically imported into a company's application management system (ATS).

The main advantage for the candidate? A faster application with no complex forms to fill in - everything is pre-filled by the parsing technology. All they have to do is submit their CV and the CV parser will handle the rest! This is what we call the "1-click application".

The added benefit for recruiters is that they get more applications, because the application process is easier and almost automatic.

The first step towards CV matching

As we have seen, analyzing the phrases in a CV allows you to quickly ‘read’ the main information in a profile. But CV parsing is also done on job offers.

Once this technology has extracted the data, another technology takes over to draw parallels between the candidate's and the job offer's data: this is matching (or CV matching).

The matching algorithm defines a compatibility score between a vacancy and a candidate, based on the recruiter's criteria. It guides the recruiter to identify the most relevant profiles.

A good matching tool will give the recruiter the ability to adjust the search criteria and identify at a glance these matching criteria directly in a CV.

Matching can also be a technology applied to an existing database. In this case, the algorithm will look in a talent pool for the most relevant talent for a particular job posting. A real time-saver!

CV parser: a quick and easy way to boost your talent pool

If matching can be used in a talent pool to identify the most promising talents... CV parsing can also enable a recruitment team to complete its talent database automatically and simply.

Instead of copying data from identified profiles line by line, information by information, with the risk of making mistakes, parsing extracts information to avoid manual entries. Recruiters save time and are sure to transcribe all the important information from a profile.

Warning: this technology must obviously meet data management & security challenges by being compatible with the RGPD (General Data Protection Regulation).

👉 The easiest way to benefit from parsing and matching within a structured talent pool is to use a CRM (Candidate Relationship Management) system.

Candidates' use of CV parsing

Candidates don't use CV parsing directly, but parsing technology also exists for job offers. What can come out of a CV parser thanks to matching is also useful to candidates.

For example, on a job board or a company career site, a candidate can upload a CV in just 1 click, and the job offers will be matched thanks to the combination of parsing and matching. This saves time and increases the relevance of the search for a new job.

CV parsing, CV matching: what are the differences?

These are two different goals, even if they are complementary:

• Parsing (step 1): reading, extracting and sorting data (CV & job offer), and potentially copying this information into a database
• Matching (step 2): analyzing the data collected (CV & job offer) to identify consistent information and classify it according to the probability of a correspondence.

Beyond the stage and the objective, there is a real difference in the way the data is processed.

CV parsing can extract absolutely all the data depending on the settings of its algorithm. Fortunately, not all of this data will necessarily be used in the matching step. Because there are biases associated with some data, and matching technologies cannot be based on all of them. This is the case with age, gender, origin, etc., which can lead to discrimination.

While this specific data can be extracted, for example to automatically fill in forms or candidate files, it will not be used by matching repositories to assure recruiters and candidates that there is no discrimination of any kind.