CV parsing helps you save time, support recruiter productivity, and reduce bias in your hiring process
A resume parser tool has become an omnipresent feature of the recruiting furniture for growing companies, enterprises and agencies in their efforts to streamline screening and hiring.
Screening resumes is the most time-consuming part of recruiting – around 23 hours for just one hire, according to data from Ideal. So this is an obvious area for recruiters, both in-house and agencies, to make efficiency savings in order to gain an edge over the competition.
More often than not, businesses are using CV parsing to make these efficiency savings. From startups all the way through to industry titans such as PayPal and Nike, hiring teams need to get through an average of 250 resumes per hire. To screen this many CVs using traditional methods is incredibly inefficient. Especially given that between 75% and 88% of them are from unqualified candidates.
But with a resume parser tool, recruiters are able to automatically extract key data from thousands of resumes and CVs to create a database of potential candidates. Resume parsing has the potential to save you a lot of time, free up your recruiters for more impactful activities, and cut out the potential for subjectivity within your process.
That’s not to say CV parsing hasn’t had its teething problems. Its reliance on keywords means it is vulnerable to manipulation, much like SEO in its early stages. And like with SEO, candidates could try to game the CV parser by keyword stuffing in an effort to get ahead of the competition.
However, developments in artificial intelligence and machine learning have made significant improvements to CV parsing. A resume parser tool of today is far more robust when it comes to hacks like keyword stuffing.
In fact, CV parsers have advanced to the stage whereby the benefits overwhelmingly outweigh the drawbacks. As this data from Ideal shows, companies that have adopted AI resume screening software have increased their performance by 20%, decreased staff turnover by 35%, and increased revenue per hire by 4%.
What does a resume parser tool do?
So let’s expand on the previously mentioned functions of a resume parser. Yes, they do extract key data from thousands of resumes and CVs to create a database of potential candidates. But there’s more to it…
A CV parser will analyse and convert unstructured resumes from formats such as PDF, Microsoft Word Documents, Excel and Raw Text files into structured XML or JSON data. In simple terms, that means they take key information from resumes of varying types and turn it into a format that’s compatible with your ATS or CRM.
As you’d expect, this information typically relates to candidates’ hard skills, soft skills, work experience, education history, contact details and achievements.
A resume parser tool can also extract data from job descriptions – job title, function, location, required skills and qualifications, etc. – to enable accurate matching between candidate and role.
As a general rule, a high-performing CV parser will extract more relevant data, and richer data, than a low-performing CV parser. What separates the two are the number of data fields, the spectrum and brevity of keywords and synonyms, how they interpret intended meaning, and general accuracy.
Types of resume parser tool
By and large, there are four different types of resume parser. What defines them is how they read, understand and collect data from CVs.
Keyword resume parsers
A keyword resume parser tool works in the same way as any other keyword-based AI, by identifying predetermined words and phrases. However, they can’t extract information from the CV beyond the keywords they are designed to look for. That’s why they are only around 70% accurate.
Grammatical resume parsers
A grammatical or grammar-based resume parser doesn’t just scan for keywords. This type of resume parser tool reads the entire document, using grammatical rules to understand the context of each word. This enables them to differentiate between meanings when the same words are used but in different contexts, i.e., semantics.
By gaining a deeper understanding of the intended meaning of the entire document, grammar-based CV parsing tends to provide more detail than keyword CV parsing. According to JobsPikr, it can achieve accuracy rates well above 90%.
Achieving accuracy levels of over 90% doesn’t come without a cost, though. Unlike keyword parsers, grammatical parsers require a lot of manual encoding by language engineers. Thorough testing is also required to ensure performance gains in some areas don’t impair performance in others.
Statistical resume parsers
Statistical resume parsing uses probability to determine the most likely meaning of a sentence. This method does understand grammar in a similar way to grammar-based CV parsing. And it can distinguish between contexts of the same word or phrase with a consistent level of accuracy. However, statistical resume parsing requires the input of a huge amount of resumes to read before it reaches that accuracy potential.
Hybrid resume parsers
As you might have already deduced, there’s a fair amount of crossover in terms of the way each type of CV parser tool works, and their capabilities. But some parsers go one further and combine the attributes of two for a best-of-both-worlds solution. For example, a hybrid of grammatical and statistical parsers would deliver highly accurate CV parsing capabilities with the capacity to learn as it goes by applying probability. The result would be more in-depth parsing that only improves over time.
The benefits of using a resume parser tool
Import and screen CVs faster
CV parsing allows you to import resumes to your ATS in minutes. Also, parsing automatically identifies and organises resumes with relevant skills and information, and eliminates those without. One benefit of this, compared with doing it manually, is that you can do hours of work in seconds. Think of how much time you could save on that 23-hour average that it takes to recruit for a single role.
Extract key data accurately
According to Recruit CRM, resume parsing can replicate human accuracy at a rate of 95-97%. Plus, thanks to resume parser tool developments that have made them more refined, there is an increased likelihood of you finding various qualified candidates that match the role. That’s because they now look at a broader and more diverse range of data from each candidate’s unique experience.
Support your recruiters to be more productive
With time-consuming importing and screening taken care of, your recruiters will be able to work smarter and focus on their building a talent community through impactful interactions with candidates. How important this is was highlighted by Bullhorn’s recent GRID Talent Survey, where 70% of candidates said they want human interaction during the search and placement process. It’s a key pillar in delivering a stellar candidate experience, and one that will give you an advantage in an ever more competitive talent marketplace.
To find out more on how you can support your recruiters to achieve this, take a look at our article on Connected Recruiting.
Ensure fairer hiring
Every resume parser tool enables customisation, meaning you can omit specific information and eliminate unintentional biases when looking at a CV. For instance, you can disable fields like age, gender, school or university name, a candidate photo, and date of birth. Dailling down these demographical fields will help you reduce unconscious bias in your hiring process.
Seamless integration with your ATS
Nearly every resume parser tool comes integrated with an applicant tracking system or ATS integration. This means you’ll be able to access everything you need in relation to a candidate in one place.
The challenges of using a resume parser
Even though they continue to become ever more sophisticated, downsides haven’t yet been eliminated altogether. So it’s worth being mindful of the following:
Nuances in context
When you think about the myriad ways there are to write the same thing, you realise that language is a complicated beast. For example, date ranges can be written in many different ways, using many different words – as can skills and achievements. Language is also not a fixed point; what is a generally accepted way of writing something today, may not be in 12 months’ time. Therefore it’s imperative that you use a parser that firstly understands these nuances, and secondly can learn new terms and skills and understand the context in that they are being used.
Cost
Resume parsers are priced either per user or per position. According to Trust Radius, entry-level fees begin at £25 per month per user and £125 per position. Some parsers will also require setup fees. But due to the fact that most are integrated into applicant tracking systems or ATS integrations, specific pricing for parser tools isn’t always readily available.
To get value for money, it’s best to evaluate your requirements in terms of scale and frequency and opt for a CV parser that will deliver that. You can reach out to providers for a quote.
Keyword stuffing
The looming spectre of keyword stuffing is ever-present. As we mentioned before, candidates’ attempts to game the system, by and large, are recognised and filtered by more sophisticated parsers, as they read context as well as keywords. But the risk of the wrong candidates being put forward is never eliminated entirely.
The bottom line
When you look at the numbers, the case for using a resume parser tool is becoming overwhelming. When you can ensure 95-97% accuracy and process in minutes what it would take a human to process in hours, you’re only losing ground on your competitors by not adopting this technology. And with the recession biting as we move deeper into 2023, those small margins are going to be amplified come the end of the year.
Download our E-Book

E-Book
Send download link to: