Written by Mayke Goethals

It may surprise you, but candidates withdrawing from an employment contract at the last minute is becoming increasingly common. Research by Gartner HR Research from June 2022 shows that 44% of more than 3,600 people who had recently received a new job offer withdrew at the last minute, even though they had already accepted the offer.

But what is the underlying reason why candidates withdraw? The main factor for almost half of the people was that they had received a better offer elsewhere. This probably sounds familiar; it is now rare for candidates to come to the negotiating table with only one offer. Competition for talent remains fierce.

However, the negotiation process during recruitment also requires an investment of valuable time on the employer's side. If that investment is lost, it entails high costs, in addition to the consequences for the quality and planning of HR work. The last-minute loss of a candidate often translates into a major loss of revenue for the organisation. It would therefore be highly beneficial for an organisation to determine which candidate will actually join the company and which candidate will cancel at the last minute. We examine the steps required for this using an example.

Hoewel het onderzoek geen harde cijfers presenteert, laat het ons wel zien hoe bedrijven slim gebruik kunnen maken van HR-gegevens en KPI’s om hun wervingsproces te verbeteren. Het biedt bedrijven als het ware een leidraad voor het opzetten van hun eigen systeem voor het voorspellen van de toetredingsstatus van kandidaten op basis van HR-gegevens.

In an American company, PrimeDataTech Pvt. Ltd, they were facing a similar problem. To provide a solution, they chose to integrate HR analytics into their recruitment process.

Initially, they went through the standard steps in the selection process (screening CVs, technical tests, telephone recruitment and background checks). They then calculated recruitment key performance indicators (KPIs) that estimate and predict whether the candidate will join the company or not. The KPIs in question were ‘offer acceptance rate’, ‘fill rate’, ‘applicants per opening’, ‘time to recruit’, and ‘cost per recruit’. We briefly explain them below:

• ‘Offer acceptance rate’ was defined as the number of candidates who join the company compared to the total number of offers made by the organisation. This KPI should be as high as possible.
• ‘Fill rate’ refers to the percentage of jobs that are filled compared to the total number of vacancies. This KPI should also be as high as possible. The higher it is, the more successful the filling of vacancies.
• ‘Applicants per opening’ looks at the number of applications for a specific vacancy. Through this KPI, the organisation can assess whether enough applications are being received so that the candidate with the right fit can be selected for the vacancy.
• ‘Time to recruit’ is the average number of days the organisation needs to recruit a candidate, from receiving the application to the final onboarding. This KPI should be as low as possible to ensure that vacancies are filled more quickly.
• ‘Cost per recruit’, the final KPI, measures the budget spent on recruiting each candidate. Internal costs include, for example, the recruiter's salary, while external costs consist of advertising costs, agency fees and background check costs. This KPI should be low in order to increase the ROI for

How can you get started with this yourself? Based on the study, we explain it again in 4 steps:

1. Prepare your data. This includes cleaning your dataset by removing outliers and missing values. Outliers are extreme values that differ from most other observations in a dataset.
2. Explore the data. You can do this by analysing your cleaned dataset with a visual tool in order to obtain important insights through, for example, graphs or charts.
3. Develop models. This step focuses on Machine Learning Data Modeling by developing predictive models using algorithms. Machine Learning is a programme that can find patterns or make decisions based on a dataset.
4. Select the best-performing model. From the different models in the previous step, you choose the best-performing one, which can then be used to determine candidates’ joining status in real time. You do this by entering candidate data into the visual interface of the predictive model.

Okay, without a data scientist, you will not develop a Machine Learning Data Model overnight. But even with less advanced analytical techniques, you can already gain important insights in just a few steps into which types of candidates are more likely to sign a contract with you and which factors play a role in this.

With a few data points such as "How long does it take to recruit the candidate?" and "How many alternative candidates do we have for the same job?", HR analytics can help you discover the relationship with your most pressing question: "Will the candidate accept my offer or not?". An interesting way to reduce costly last-minute withdrawals and optimise your recruitment strategies.

Would you like to read the study yourself? You can download it via this site.

Would you like even more inspiration on how to gain these insights from your own organisation through HR analytics? Or are you looking for more concrete solutions to address challenges such as employee turnover and retention? Contact us, or subscribe to our bi-monthly newsletter for tips and trends on HR analytics.