The 7 commandments of predictive recruitment
While the predictive approach allows recruitment professionals to radically improve the effectiveness of their selection processes, it nevertheless needs to be applied in a very rigorous way. Here are the 7 commandments that you should follow closely to ensure the successful implementation of your predictive recruitment projects.
#1 Define your objectives
Do you want to recruit better performing workers? More committed ones? Are you seeking to reduce your staff turnover by 10, 20 or even 30%? Every predictive recruitment project absolutely must start with this. In fact, the objectives that you define will shape the data you’ll need to exploit and the type of analysis to be conducted on this data.
If you don’t precisely define your objectives in predictive recruitment, it will be impossible for you – in 12, 24 or 26 months – to judge the relevance of the process.
#2 Identify exactly what data you need
Let’s say you choose to concentrate on recruiting better performing employees for a particular job category (office-based salespeople for example). In this case, you need to know exactly what criteria are used internally to evaluate job performance for each employee (number of calls, number of phone appointments, number of online demonstrations of your solution or products, number of orders signed, turnover generated, margin generated…). Talk with the manager or director of operations if necessary, they will always have a list of criteria that they use to judge the performance of every employee.
If you don’t identify exactly what information you’ll need to collect, you risk becoming unfocused and collecting all the data that comes to hand. Contrary to what is often believed (and heard), when it comes to collecting data, more is not always better. What matters the most is being able to pick out relevant, high quality data.
#3 Target potential!
The main studies conducted on the relevance of predictive recruitment processes have shown evidence that one of the most reliable indicators for the probability of people’s success at work was their potential. When we talk about potential, we mean primarily their skills, their motivations and their personality. In short, focus on “who is this person” instead of the story they’re trying to tell you in their CV or digital profile.
These days, everyone knows that the information presented in candidates’ CVs is “massaged” in the order of 70-85% (for salespeople in particular). Who would rely on data where 70-85% was not completely “clean” to make a decision as important as whether to bring in a new person to their business or not?
#4 Work by stages, in a methodical way.
In basic terms, the introduction of a predictive recruitment process occurs in 3 key stages. One, you collect all the useful data about the performance (or commitment, lifetime etc) of current employees in the role. Two, you collect what we call the “potential predictors”. In this category, we find seniority, the location and personal characteristics of current employees (skills, motivations, personality), as well as any other factors that you think might have an impact on the performance of those employees. Three, you identify the predictive patterns (personality, skills, motivations, seniority, career paths) positively related to the factors that you want to predict (performance, commitment, lifetime…).
Because no other way to approaching it allows you to get better results – that’s all there is!
#5 Identify a model, and stick with this model.
The key thing about setting up a predictive recruitment process is that sooner or later, you’ll be confronted with choices. It’s a good bet that if you focus your study on salespeople, one version of the model will let you identify the best “hunters”, while another will focus more on selecting profiles that are oriented towards loyalty. What is critical is to make a clear decision and not try to maximise every single performance factor at the same time.
The latter option, for example, would usually lead to you averaging out the ROI after one year…
#6 Once you’ve chosen your model, apply it rigorously!
Recruiting in a predictive way means applying a preselection model that allows you to only interview a fraction of the candidates who send in an application. The interest in this approach is that unlike choosing based on CVs, the criteria for your decision is clear: you have a benchmark figure (60% correspondence with the model, for example) and you only choose to interview the candidates who meet that benchmark (or surpass it). If this seems to be a little “barbaric” or “violent”, remember for a minute the way in which most recruiters work today: they receive a pile of CVs, they sort them into 3 piles (Yes / No / Maybe) …but based on what? Based on a 10 or 20 second scan of information that has no proven link with likely performance in the position?
If you start compromising on the benchmark value, it’s the beginning of the end…this candidate at 56% isn’t bad, and after all, we’re not far from 60%. Oh, but this one at 43% also has some interesting elements in their CV…suddenly, you find yourself operating in the same way you always have. From then on, you’ll be back to the same results you were getting before – not worse, but not better, either.
#7 SYSTEMATICALLY measure the results obtained.
This is a key point in the predictive recruitment approach. Once you’ve set up the predictive recruitment process, you must systematically evaluate its relevance in 12, 24 and 36 months’ time. To do this, nothing could be simpler – just collect (for example) the performance or the turnover after 12 months of new recruits (since the introduction of the predictive recruitment process) and compare them to the performance or turnover after 12 months of those employees hired before the introduction of predictive recruitment. Are the results conclusive? Are they meeting the objectives you set yourself in step 1? If yes, then everything is OK. If no, then go back and work on the model.
Because predictive models are living things. You can’t just set them up once and for all, then blindly follow them for the next 3 years. Every new recruit brings you new information. By taking into account their profiles and their performance in an iterative way, you can adopt an approach of continuous improvement. And that, too, is the philosophy of predictive recruitment.