Job Recommendation Tools: Lower Your Abandonment Rate

Based on the 2021 report on digital experience benchmarks across industries, almost half of all website visitors abandon the site after seeing only one page, making the bounce rate of an incredible 47%.

While in some cases, this statistic makes perfect sense and requires no need for improvement, if talking about career pages and their success measured in the number of applications, the goal becomes having the abandonment rate as low as possible. More applications means more potential quality candidates and less horror scenarios like the ones described in the beginning of the article. 

Even though bounce rates might be a foreign concept to some recruiters, the abandonment rates, on the other hand, are no stranger to anyone working in the HR industry. However, there is not much difference between the two. Abandon rate stands for an HR metric that shows how many job applicants left the site before submitting the application and a bounce rate indicates how many website visitors left after viewing just one page.

Imagine that your company's career site has a bounce rate (or abandonment rate) of 40% (which is well within the industry benchmark), and by implementing a new strategy, this goes down by only 5%. What effect would decreasing the abandonment rate to 35% have on your average cost per hire and, ultimately, your company's revenue? 

(Source: www.shrm.org)

While all these numbers may sound appealing but hard to achieve, the only logical question becomes - How do we do that? We invite you to keep reading the blog and learn why AI is such a great helper in this mission. 

Job recommendation tools: What do Netflix and Career Pages have in common?

In the past few years, it's become quite common that platforms operate in a way that often feels like they are reading our minds because of the incredibly accurate content recommendations they are offering. From apps that are here to fill in our free time, such as Netflix and Spotify, to job platforms recommending open vacancies that we might be interested in. 

In order to be successful in recommending various content, platforms often use one of the three possible methods (Source: towardsdatascience.com): 

1. Collaborative - "Collaborative methods for recommender systems are methods that are based solely on the past interactions recorded between users and items to produce new recommendations." Meaning that users' preferences towards specific content will ultimately reflect on what the software will offer to the user in the future. 

In recruitment, this means that a Software Developer will only get recommendations for Development jobs if he was previously visiting pages related to software development. The same goes for a marketing professional who would never be recommended a software development position if they never looked into one before. 

2. Content - ".... a model, based on the available "features" that explain the observed user-item interactions… for example, the age, the sex, the job, or any other personal information for users." This model doesn't only consider interactions between a user and the software but also recommends based on the candidates’ characteristics.

3. Hybrid methods - a model that combines collaborative and content-based methods

Regardless of the AI tool's method to recommend specific content, each has an immense benefit for the company. Besides driving the traffic, each technique has the power to deliver the right content to the right audience, which leads to lower abandonment rates and higher conversion rates - the ultimate goal of almost any website.

However, the difference between methods remains essential. Depending on what type of product or service the platform is offering, their AI tool uses different ways to increase the effectiveness. This brings us to the recruitment industry and the practices that lie beneath the seemingly simple process of recommending similar job openings to new candidates or those previously applied for company positions.

Recruitment AI for the best job recommendations 

One of the benefits of using AI lies in leveraging all aspects of job recommending tools. To have a clear understanding of all of the tool's possibilities, it's essential to make an adequate distinction between its tasks. Software's duties can be divided into two categories:

1. Recommending similar jobs to a candidate that is visiting your career site

2. A possibility to send out tailored emails with job recommendations to candidates from the talent pool who previously applied for similar jobs at your company

Career site recommendations 

One of the significant issues that companies face nowadays is a high abandonment rate on their career pages. Besides adjusting the creative part of the page, an extremely effective way of dealing with such issues would be implementing a recommendation tool to the site. Such a mechanism can recommend similar jobs to the candidates that visited your career page and didn't find the perfect fit. This additional content presented on the career page may lead your candidates towards discovering new better positions at the same company.

There are two ways in which software offers accurate recommendations to the candidates: 

1. Job to Job similarity: The software uses this method when it doesn't possess any additional information about the candidate except the title of the position that they are observing. Let's say a candidate is looking into the role of Junior UX designer on your career site. The tool will store this information and display all other open vacancies at the company that might be interesting to this candidate - such as Senior UX designer, Motion designer, Web designer, and so on. In the background, the software takes a job description observed by the candidate and compares it to the other available positions screening for similarities and differences between all available job descriptions. 

2. Job to Candidate similarity: In this case, the software already has some information about the candidate. Whether the candidate applied for your talent pool and filled in some information about his or her employment history or uploaded a CV in a previous job application - the software uses this information to recommend new opportunities to the candidate accurately. In this case, the result for candidates might be the same or more accurate (compared to Job to Job similarity), but the background process is slightly different. This time the software scans a CV looking for relevant information and begins a matching process between the candidate and the job description. Ultimately, based on the scores given after comparing the entire database of job descriptions and the candidate's CV, the candidate is presented with a list of jobs that might spark their interest. The process is similar to the CV parsing discussed in the previous blog post. 

Email tailoring 

Besides recommending various similar job opportunities as part of the career page, recruiters can send out tailored emails to candidates and offer new job openings based explicitly on their competencies thanks to AI. In this process, the software's method is similar to the one used in previous cases (again, the software matches job descriptions and the CVs and ranks jobs based on relevance). However, this time candidates are directly receiving new job openings in their inbox. 

Such a direct way of communicating with the candidates increases their chances of applying for positions and makes sure that your talent pool is aware of all opportunities at your company. This communication is now automated, enabling recruiters to have more time to focus on other tasks while fully trusting the software to do its functions.