Developing a Data-Driven Talent System


Written By: Jim Hunter


In the high stakes competitive world of business, you can no longer afford to get burned by archaic hiring practices. How many more times do you need to go back to the well of reading resumes, setting up a booth at a networking event, and/or relying on background checks and references before you realize that all you’re doing is getting misled, misinformed, and manipulated by the talent pool—all while leaving yourself susceptible to the personal biases of your hiring managers?  

You’re probably feeling the pressure of making the same people mistakes over and over again already. Investing time and energy into finding the wrong people is costly, and not just financially. This carousel of mis-hires and terminations is stealing the joy from your rock star employees, causing your product(s) to be called into question, creating a feeling of angst amongst your customers and stakeholders, and making it impossible for you to reinvest into your business and grow.

Aren’t you tired of it? Wouldn’t you rather find the right people, for the right role, at the right time? You can bet that your competition is fed up and you can be guaranteed that they’re looking for alternative methods to deepen their talent pools to ensure that they have the right people on their team, before they need them. You know your people are your competitive advantage, but what you don’t know is how to find them when your current tactics are failing. 

Imagine if you had a fully customizable hiring, staffing, and talent development system that is built solely around the metrics, principles, character traits, and goals that are most important to your business. Now imagine that this system was completely unfettered by personal bias, insusceptible to misinformation, and could give you the ability hire with certainty—plus enhance your ability to retain your employees. And what if this system could, in turn, learn alongside the leaders of your organization and grow with the business as it scales?

Sounds too good to be true. It isn’t.

What Are You Trying to Accomplish?

It is important to understand the underlying why behind building such a talent system. Because once we understand why we’re investing in the creation of a talent system, we can begin visualizing what the ideal characteristics of fully integrated, end-to-end talent system actually look like and begin building a hypothesis on what it’s going to take to help us build a system that embodies those characteristics. 

There are four primary “people-goals” that companies are trying to achieve:


Talent Acquisition

Hiring new people is the number one priority for the majority of our clients. As the internal demand for new employees becomes critical, the pressure to find new job candidates also mounts. As the pressure gets more intense, it often leads businesses to make questionable hiring decisions.  

It doesn’t take but a few costly mis-hires for a company to realize that they need to refine their hiring practices and look for ways to mitigate poor hiring decisions and streamline the process. This way they can reduce the risk associated with hiring new people, increase their return on investment, and begin realizing their growth potential.   

Unfortunately, protocols such as multi-tiered internal interview hierarchies and intensified HR screening programs often have little to no impact on finding higher performing people and sometimes even have an adverse effect and result in negative hiring outcomes.

Talent Engagement

If talent development and retention are important to your company, then you have to prioritize Talent Engagement over everything else because getting and keeping your employees engaged is how you’re going to unlock development and retention, create a virtuous cycle of employee satisfaction, and ensure that you’re getting the ROI you’re after. 

The fact is, you’re not going to be able to affect retention directly, so that means you’ll have to rely on clearly articulated career paths and robust engagement strategies to ensure that you’re cultivating opportunities and providing a fulfilling work environment for your team.

This means that you’re going to have to come to understand your employees more intimately than ever before. You’ll need to understand what’s motivating them today, in the role that they have. And what’s going to motivate them tomorrow as they look toward the role they desire. Often times, the sheer amount of people and lack of upward mobility options can get in the way but you can’t let it. You’re going to need to develop a balanced and dynamic plan for your people.

However, companies have historically relied on one-size-fits-all training approaches & canned career progression paths to develop talent, and company-wide surveys & human observation to assess engagement. 

These approaches to development & engagement are typically executed on an ad-hoc basis and are insufficient when it comes to making an impact on employee satisfaction, and ultimately retention.

Talent Flexibility

The reality is that today’s marketplace and the way we do business is changing faster than ever and it’s becoming more and more complicated to flex and pivot to meet emerging customer and market demands.
This means that companies are scrambling to develop new business strategies and are having to spin up new products, departments, and teams more frequently than ever.

As your business moves around, you must have a strategy for how your people are going to move with the business. After all, you’ve invested time, energy, and money into your employees, and many of them have become some of your greatest assets, so you’re not going to want to simply lay people off whose roles have morphed or don’t exist anymore. Instead, you’re going to want to leverage your team in a way than not only helps you achieve your strategic business goals but also sets your people up for success.

Historically, organizations who have attempted to flex their talent have relied on their leaders to assess and make judgment calls, or simply decided to shoehorn their existing employees into new roles and arenas.
All with little to no success because successfully flexing your talent requires insight and nuance. Not sheer will. Not gut instinct. And certainly not a dictatorial approach. 

Diversity, Equity & Inclusion 

Diversity, equity and inclusion are three of the most important tenets of any self-respecting company and should be taken very seriously. Aside from the moral obligation we have toward humanity, studies have shown that diversity has tons of positive side effects for your business. For example, more diverse companies excel in things like talent acquisition, customer orientation, and decision making—all of which lead to increasing returns. 

The problem is that, despite some of our best efforts, companies still rely too heavily on the human condition to make their talent decisions. Historically, humans aren’t good at embracing diversity and cultivating inclusive environments because we all have our own inherent biases.

Some biases are more insidious than others, not every bias has to do with race, gender, sexual preference, or age. Some might be based on what university a person attended. Or the clothes they wear to the interview. Maybe your potential new hire has tattoos, and the hiring manager doesn’t like tattoos. Who knows? The possible combinations of biases are literally endless—ironically—because humans are so diverse and complex. 

To truly cultivate an environment that embraces diversity, equity and inclusion you must find a way to first identify these biases and then mitigate them from creeping into your people-decisions while preserving your ability to make the best possible hire.

The Fundamental Elements Of A Talent System

When preparing to build a talent system, understanding the four primary “people-goals” that a well-rounded talent system should be able to solve is only half the battle. The other half is knowing the fundamental elements of the system are and what’s getting in the way of standing up those elements.

Every system, talent or otherwise, is powered by some sort of an input. Then, some sort of process happens within the system, and the system delivers an output. 

Let’s use a car’s fuel injection system as an example. Now, I’m no mechanic, so bear with me as I oversimplify the process. But, basically, the way a fuel injection system works is that pressurized fuel (the input) is drawn into one of the engine’s combustion chambers. After being mixed with the right amount of air, it’s ignited, (the process) and you get combustion (the output). 

Like most systems, there are different ways to go about building a fuel injection system, and innovation has improved upon early iterations of these systems. Of course, there are tons of literal moving parts that are hard at work within the system, but at its core—it’s an input, a process, and an output. 

Talent systems are no different. They consist of: 


Data (The Input)

Every talent system is going to rely on some form of data (input) to help you make people-decisions. So, you’re going to have to develop some way of reliably collecting data from and about employees, both old and new. The data should be easy to understand, able to be reliably obtained from a diverse set of people, and should always be presented in a way that is intuitive to anyone who needs to look at it.

In addition, this data can’t be biased in any way, by race, ethnicity, age, gender or even by similarities found in past backgrounds or even where someone went to school. We tend to think of resumes as a data point, but others could be just a simple scorecard of what specifically is needed in a job function and a person can be assessed against an objective criteria.

Ideally, your data collection methods wouldn’t be easily manipulated by savvy interviewees and test takers. And, it would be free of inherent bias and clear enough so that it’s nearly impossible to be misinterpreted by the people responsible for making talent decisions.     

Model (The Process)

Secondly, the data you collect will get leveraged to build a model that will act as a translation mechanism and  help you convert your data (input) into insight (output) once you begin comparing & contrasting your people against the model to make hiring decisions and/or placing existing people into new roles. 

Even though these models are initially based on top performers and ideal job candidates, it’s important that the model doesn’t push those responsible for making talent decisions in the direction of simply looking for carbon copies of the high performing individuals and ideal candidates. 

Rather, your model should point toward the underpinnings of success so that your organization can maintain a healthy diversity among teams and other working groups, and it should be flexible and easy to edit as new information arises to maintain its effectiveness.

Insight (The Output)

Once you have a model and begin comparing people against it, insights (output) will be revealed. Those insights will tell you something about the person in question. Maybe the insight is this person is similar to or different than another person. Maybe the insight is that they have an interesting personality, or that they are going to react in a certain way given a particular set of external stimuli. Your insights may only tell you that the person in question has achieved certain things in the past or possesses a skill that you find desirable. 

The point is that no matter what your insights are, they’re going to tell you something about the individual. The key is understanding what the insights mean to your organization. Are they academically interesting? Are they just another data point for your interviewers to use when making their judgment calls? Or is the insight meaningful enough to use as a filter when making your next round of people-decisions?

Understanding What You’re Measuring

Now that we’ve established the fundamental elements of a talent system, we must understand the people side in what it is we’re going to measure. What type of insights we choose to understand people will ultimately affect what type of model we’re able to build and impact the insights that our system can reveal to us. There are three main focus areas that organizations tend to look for when building their talent systems. 

The big issue here is that not all tools are effective and this area is often the most confusing of all talent systems, specifically due to a lack of effective insights inside who someone really is and a haphazard ability to track or measure, if indeed, any of your past people efforts are effective now or in the future.



A person’s skills are often tactical and matter of fact. You might look at a resume and see items such as the following: proficiency in Microsoft Office, multi-lingual, expert in programming languages, SEO/SEM marketing specialist, database management experience, Adobe software suite certified, etc. These would all fall into the category of hard skills. On the other hand, you may see some organizations use things like mechanical reasoning, writing, or math tests to look things like the ability to learn, problem solve, or plan.

Skills, both hard and soft, are dynamic and can be developed over time. Additionally, the skills that people currently possess were often obtained in a particular context which may or may not be suitable for your organization.

For example, someone may know how to use a particular tool within the context of their old organization, but that doesn’t mean that you use the tool in the same way. Maybe you’re using it for a different purpose, or you may have it set up differently, and so a person’s experience with the tool may not translate immediately.  

For these reasons, skills have very little bearing on whether or not a person is going to thrive within your particular organization. It’s also going to be incredibly difficult to build a talent system around this type of data because the models will either be too restrictive or too loose, and at the end of the day, the insights skills give you are subjective, at best.


Personality tests are designed to reveal aspects of a person’s character. This type of data is wildly popular, and more and more companies regrettably are beginning to build their talent systems around it. In fact, recent statistics suggest that 60% of workers are asked to complete a personality test as part of the hiring process. 

Personality is also fixed. Depending on who you talk to, a person’s personality is either something they’re born with that cannot be changed, or at the very best it takes years for someone’s personality to change even a little. So, either way, from an employer’s perspective, personality isn’t something you’re going to be able to influence. Despite being a fixed entity, this type of data still seems like it would be sufficient to build a talent system around, in theory. 

There are a ton of personality tests out on the market that provide a reliable way of collecting personality data—the Myers Briggs Type Indicator (MBTI), Color Code, and DiSC are a few of the most well-known options. 

Each personality test will give you a model, in this case they call them personality types. And it’s true that these models will allow you to compare people against them and provide insight that will enable you to make talent decisions.

The question you must ask yourself though is, “Are the insights I’m getting the right insights, and will they enable me to make good talent decisions?” 

Despite their popularity, personality tests have their drawbacks. For example, many people when put under pressure to take one of these assessments, often for the purpose of getting hired, will respond with answers that they think the employer will want to hear. This can lead to skewed results that will not be indicative of the individual’s true nature. 

Additionally, personality tests have the potential to screen out otherwise qualified candidates. For many jobs, especially in the new digital world of work, there isn’t a main personality type that fits the job role. Many organizations are creating new positions, people are shifting toward remote work, and personality tests haven’t caught up to the times yet. 

Lastly, some personality tests have come under legal scrutiny over the last several years because of claims that they violated the American’s with Disabilities Act. Plaintiffs in these cases accused the personality tests of being able to identify people who may be living with bipolar disorder, mania, depression, and other mental illnesses.


Behavioral assessments are designed to determine what the unique characteristics of a person are, how they respond to outside stimuli, what their interpersonal interactions look like, and how they will handle specific circumstances. Will they recover from negative interactions quickly? Are they capable of displaying empathy & understanding? Are they quick to anger and lash out? Or do they tend to shut down emotionally? 

This type of data is different than personality because personality tells us who a person is, and behavior tells us what a person is likely to go do—who you are and what you do aren’t mutually exclusive. Also, behaviors—like skills—are dynamic. This means that a person can learn new ways of responding to the world around them through the influence of targeted coaching and/or training.

You can certainly build a talent system around this type of data. There are many behavioral assessments out in the market today that can reliably gather behavioral data. Companies like Omnia, Caliper, and Hogan are some of the big players in this space. From these assessments you’ll yield a behavioral profile that you can use as your model. And the model will provide insights that you can use to make talent decisions.

For all the reasons above, behavioral data is widely regarded as the optimal data set to use when building a talent system. But that doesn’t mean that this type of data doesn’t have its own set of drawbacks.

The problem with behavioral data doesn’t lie with the data itself. It’s how it’s obtained that makes it hard to use. Historically, behavioral data requires the one-off expertise of an I/O psychologist to come in an do an assessment of your organization. They go back and crunch the numbers and in 12 weeks you’ll have a behavioral profile. And it’ll cost you $20K-$30K dollars to do it.

That’s a huge investment. Perhaps too big if you can talk yourself into believing that the MBTI, which is free, is a viable alternative. And let’s say you do decide that you want to invest the money. You’re going to want that profile to have legs, so you’re going to hold on to it for a while. But if your business is moving, then that profile isn’t going to be relevant a year from now, let alone 5 years from now.

Validating Your Data-Driven Value System

So far, everything we talked about isn’t unique to any one particular company. We’ve merely been exploring the truths behind building a talent system. We’ve established that when companies set out to create a talent system, they are hoping to achieve one or more of the four primary people-goals: Talent Acquisition, Talent Engagement, Talent Flexibility, and Diversity, equity and inclusion.We also discussed the three fundamental elements that a talent system should have: Data, a Model, and Insight. Then we discussed the different types of data that you potentially could build your talent system around: Skills, Personality, and Behavior. But how will we know when we’ve built the right system? What does it look like when you’ve got all the pieces in place and it’s working effectively? The way you’ll know you’ve gotten it right is by assessing the system you’ve built against the following criteria: Is my talent system: 



An effective talent system is going to be able to predict the likelihood of an individual’s, team’s, or department’s success by whatever criteria you’re measuring success against. If Talent Acquisition is what you’re after, then your system should be able to predict the likelihood of a new hire’s ability to be successful within your company. 

If Talent Engagement is what you want to achieve, then the system should be able to predict the likelihood of your engagement strategies and development plans to make an impact on the people in your organization.
If Talent Flexibility is your top priority, then your system should be able to predict whether or not people in new roles will succeed and how the change is going to affect the people within your organization.

If Diversity, equity and inclusion is your goal, then your system should be able to get diversity in the door and predict whether or not your people will be inclusive enough to leverage that diversity. 

As you might imagine, in a data-driven talent system, the data is grease that keeps this engine turning, so getting predictive starts with the data. Your data has to look toward the future. So, lagging indicators of success, things like achievement awards, certifications, job history, and experience don’t necessarily mean that someone is going to be primed and ready for success within the walls of your organization. 

Instead, your data should be pointing toward leading indicators of success. When I say leading indicators, I mean attributes and character traits that will give you an idea of how someone might behave an a given situation. Things like internal drive, ability to adapt, appetite for continued learning, and humility are often exponentially more indicative of future success than the things you might see on a resume or discover in an interview.


Looking back on the fundamental elements of a talent system, we know that your data should allow you to build a model that will enable you to compare and contrast your people so that you can make the best possible people-decisions. But a one-size-fits-all or subjective ad hoc approaches to data mean that your models won’t be inherently useful to your organization.

A one-size-fits-all approach will give you a canned model that’s been predetermined and will lack the context of your organization. This type of model doesn’t take into consideration your company’s values, its unique goals, its particular constraints, the exact specifications and duties of the roles your looking to have your people fill, or how you define success.

On the other hand, allowing your models to be fueled by the gut instinct of your talent managers and HR people isn’t going to help you create a suitable model either. When you add the human element and allow it to become the driving force of the talent system, you don’t get models that are customized to your business. Instead, you get models that are customized to the life and experiences of the individual(s) who’s making the people-decision. In some cases, this might work, but typically, the success of talent systems that revolve around these kinds of models is spotty at best and certainly isn’t scalable. 

So, to achieve a model that’s customized, you have to be selective about the data you choose to base it off of. You’re going to need dynamic data that’s meaningful to your organization so that you can create a model that’s aware of the current landscape of your organization and the particular people-goals you’re trying to achieve. 


An effective talent system, one that’s based off of predictive data and centered around customized models, should also give you actionable insights. When you have insights that are truly actionable, they will enable you to make a wider array of people-decisions than ever before. 

Historically, the way many companies have approached building talent systems have typically left them with little to no insight. Worst case scenario is your talent system leaves it up to human judgment, which we know is susceptible to biases and misinformation. In the best of cases, your talent system is enabling you to make the most basic of decisions and that’s to say yes or no to pulling the trigger on hiring, promoting, terminating, etc. But, when you have predictive data, customized models, and actionable insights you can begin to do more. You’ll discover that you have the ability to leverage your system to make staffing decisions. You can build better, more cohesive teams. 

You’ll be able to leverage the system to create robust engagement plans for your people because you’ll know what motivates them and how to communicate in a way they’ll receive.
Your career progression and development plans will improve because you’ll know what skills gaps your people have and can close them faster, plus you’ll know what roles are most suitable for any given person.    

To go one step further, the insights you get from the type of talent system that I’m describing will allow you to get out in front of potential problem areas in your company.
You’ll be able to leverage the system and its insights to identify areas of your organization where there is potential for low morale, low engagement, and low tolerance for Diversity, equity and inclusion.


The presence of predictive data, customized models, and actionable insights isn’t quite enough to call your data-driven talent system effective because you also need a talent system that will be able to leverage these three elements in a way that’s sustainable.

So, what are the main characteristics of a sustainable, data-driven talent system? 

•  It has to turn data over quickly and with pinpoint accuracy
•  It must be cost effective to allow for frequent use.

One thing in the business world is certain, and that’s the fact that change is coming. Change is inevitable and when it happens it’s going to come fast. So, you’re going to need a talent system that can keep up with you as your business matures and new needs arise. Nowadays, you know more about your business and your market than ever before, and you need a talent system that can learn as fast as you do.

Think about it. If you need to hire someone next week so you can sell your next high, six-figure deal and it takes three months to get a profile—is that going to help you sell that deal? Probably not. If developing a model is going to cost $20K each time you need one and can only be done as a one-off by an expert I/O psychologist—are you going to want to create a new one monthly? Annually? Probably not. If it costs $20K and the profile sits on the shelf for 5 years—do you think it’s going to help you when you need it? Probably not. 

If you had a cost-effective system that could pump out custom profiles in minutes, the possibilities are endless. Not only could you hire someone when you needed them and use an up to date profile to do it, but you could also use it to tighten up hiring practices in general, and save money on training & onboarding.

Imagine a world in which you could create new models over and over again as new data becomes available. That new data being employees whom you’ve hired and employees who’ve left the company. 

You could then tweak the models ever so slightly on a whim, getting the customization tighter and tighter. The system would—in effect—get smarter the more you used it. It would learn new things about your company and the people who work there as fast as you wanted it to.

Parting Thoughts

One of the things that comes up often when I’m speaking to people about developing a data-driven talent system is that even if they’re willing to concede that what I’m saying is the truth, they don’t know how to execute on it. So, let me break it down.

You want to build a data-driven talent system to help you achieve one or more of the people-goals I mentioned earlier, and you know that the driving force is the data, so you have to choose which type of data you’re going to measure wisely to ensure that you can create custom models and get actionable insights. 

You want to choose behavioral data because you know that it’s the optimal data set because it’s most likely to be predictive and it will tell you what people will go and do, not who they are or what they’ve done. But this type of data is expensive and hard to replicate without one-off expertise. 

And you also know that somehow, you’re going to have to pull this off in a way that removes human bias so that you can foster a diverse and inclusive working environment.

There’s just one way to do it all. 

You have to use machine learning/artificial intelligence (ML/AI).

AI that’s been programmed to use a blend of behavioral science and psychology to get the data you need to build your models, and ultimately get the insights that are going to unleash your organization’s people-decision-making excellence.  Think about it. ML/AI is the only way mitigate human biases. It’s the only way to circumvent the need and expense of hiring an I/O psychologist, so that you can consistently and frequently gather new data, and can tweak your models and improve your insights. This type of talent system, one that’s powered by ML/AI that can extract behavioral data is the only option that checks all the boxes. It’s going to be the type of system you need to get your desired outcomes and achieve any of your people-goals.


Connect with us!