Think Outside the 9 Box

Sick of the performance-by-potential paradigm? So are we. Let’s replace it with something better.

By Allan H. Church, Ph.D.

No matter how hard we try to ditch performance ratings—and many organizations have made valiant attempts—we just can’t quit them. Psychologically, whether we like it or not, companies need ways to segment the best from the rest and distribute rewards, resources, assignments, and promotions accordingly. And performance ratings are still the easiest method for differentiating talent.

One such wildly popular method is the 9 Box talent management model, where one side is performance (the X axis in a matrix from low to high) and the other is potential (the Y axis from low to high). When it comes to determining future leadership potential, however, performance alone isn’t enough. Past behavior may indeed be the best predictor for future performance, but we know from the latest theory and research in the field of industrial-organizational psychology that the concept of leadership potential is far more complex than that.

If predicting potential was so easy, we wouldn’t see highly successful leaders frequently rise to their level of incompetence, or successful executives from one company transplanted into another fail so regularly. So in the context of high potential identification, it should be time to drop the use of performance ratings all together. The problem? We’re not there yet.

In a recent benchmark study of top development companies, my colleagues and I found that 75 percent use these types of indicators. Sure, performance “gets you in the game,” as we’ve been saying around the offices of PepsiCo for years, and it’s an important gatekeeper to potential. But it’s hardly the leading indicator of one’s ability to take on bigger roles in more complex environments.

So why are we still so enamored with current and past performance when predicting potential? Because sometimes it’s all we have, and we’ve been lulled into believing it works. Until it doesn’t.

Performance is one of two core dimensions in the 9 Box framework, which is intended to drive different talent management outcomes. But if potential and performance are essentially synonymous, then isn’t the 9 Box actually a one-dimension approach? We call this the performance-by-potential paradigm, and it’s arguably the biggest reason why so many talent management processes fail to identify and place the right types of leaders for succession.

When senior leaders ask their HR colleagues, “Why do our high potentials fail to deliver?” is it because the definition of potential was based on current performance, not future capabilities? It seems like an obvious answer, yet remains constantly overlooked.

What we need instead is a new way of framing the 9 Box model for differentiating talent, with performance as a gatekeeper to entry into the tool, not a dimension for classification and follow-up actions. If someone isn’t performing well, they aren’t performing—period. And if they’re doing well, then let’s consider their long-term potential using something else that’s more valid.

Great, you say. So what the heck are the other options? The answer depends in part on the goal of your talent segmentation process and the level of rigor you’re willing to adopt.

What follows are three different alternatives to the classic 9 Box model, ranging from the relatively simple to the more comprehensive and targeted. Each has a fundamental purpose in answering the following question: Potential for what? And each relies on some talent classification scheme that’s already present, however muddled it may be.

And remember: The focus here is on throwing out the performance side of the 9 Box—not the potential side. That’s a complex discussion for another time.

Solving for Talent Ownership: Potential by Mobility

This approach to segmenting talent is easy to implement and solves one of the most basic operation issues with respect to talent management process: Who owns the talent and the associated development and movement?

It also helps end another fruitless debate that many organizations continuously have: Should we require someone to be mobile to be a high potential? Given the talent landscape today, with newer generations less willing to move whenever the company taps them on the shoulder, this is an important issue that needs to be solved.

How does it work? It’s simple enough. First, create a basic three- or four-category mobility framework. Here’s one that we’ve used effectively in the past:

  1. Global: Mobile to any location in the world. The organization is supportive of those moves.
  2. Regional: Only mobile to locations within certain areas (U.S. only, Latin American only). Again, the organization is supportive of these types of moves.
  3. Local: Not mobile and needs to remain in his or her current location.

You could easily use an alternate framework—let’s say global, U.S., non-U.S., and local—as long it’s a simple combination that enables better action planning and ownership rights as a result. The point is to figure out what works for your organization.

Once you have this model in place, gather preferences from employees about their mobility (if you don’t have them already) and classify talent accordingly (potential by mobility). Next, segment the boxes according to ownership.

Typically, the global high potentials will be the assets owned by the enterprise talent function, while regional high potentials might be controlled by country presidents and their HR leaders. The one additional consideration for discussion in talent reviews is whether or not the organization is supportive of the mobility preference.

So, for example, just because someone is deemed a high potential and he indicates he’s mobile anywhere in the world, you might find in a talent review with senior leaders that some wouldn’t be willing to take the risk in certain cultures, functions, or roles. Thus, that individual becomes regional talent or perhaps not a high potential at all if the discussion forces a deeper review. Then you’re having this discussion: Would you really take them since they’re ready to go anywhere? But that’s the point.

The other nice thing about this method is that you don’t lose sight of the local high potentials in your organization, as is often the case when overemphasizing mobility. The list of high potentials is always maintained, even if their mobility is limited and their development and career progression is owned at the local level. This way potential is pure and careers can be managed locally if appropriate and desirable.

Equally important: If one day a local high potential’s circumstances and preferences change, he or she won’t struggle to get back on the list. This is perhaps an even greater benefit to the talent pools who may have taken time off for specific periods in their careers but want to ramp up later, like working mothers and fathers. This way, they won’t have to prove themselves all over again.

What are the downsides? First, this approach relies on mobility information gathered from employees. Collecting that data, keeping it up to date, and ensuring its accuracy can all pose challenges at times. After all, people may tell you what they think you want to hear.

Second, while the framework clarifies ownership and keeps the local high potentials top of mind, it doesn’t add new information to the future potential identification process.

For that we need more data—and the next approach.

Solving for Future Success: Potential by Behavioral Science Insights

This framework can take many forms, but it really gets to predicting future success—i.e., the true nature of leadership potential. Depending on the robustness of the assessment tools used (measurement reliability and validity), the process can range from basic to extremely sophisticated and powerful in its approach.

As you might expect, the more design work and rigor we place in collecting data and insights on leaders, the better the predictive power and development insights we can obtain. Finding the right balance between quality data and complexity is the key to a successful implementation.

But once a behavioral insights program is established and existing ratings of potential (as defined by leaders) are compared with assessment insights data, the results can be eye-opening. This can be very helpful in highlighting potential sycophants in the mix, as well as identifying exceptional talent that might be suppressed or hidden for a variety of cultural, political, or situational reasons.

This is what the traditional 9 Box was meant to do: It’s essentially organizational perception of potential (politics) versus the reality of potential as measured (data).

So how does it work? First, you need a set of leadership competencies and behaviors that measure future success. Having a core set of values is useful for ensuring the right culture is in place, but they’re not always enough.

While developing the leadership competencies of the future can take some time, once you have a new model or framework ready, you select the tools (like surveys and tests) that measure these competencies in the most valid and reliable ways. The best approach is to use more than one tool to avoid the biases and challenges associated with any single measure.

While 360 feedback, for example, is perhaps the most popular survey process, sometimes it can be gamed if individuals are allowed to select only people who will provide favorable ratings, like their friends.

This can be offset by other tools like business simulations, personality assessments, structured interviews, situational judgement, or even cognitive tests, depending on the level of sophistication and rigor being employed.

It’s important to note that the quality of the insights generated is directly proportional to the rigor of the tools being used, and the capability of those talent management professionals who are interpreting the data.

The bottom line: You need to make sure whatever tools you select have the right properties to be legally defensible if you’re making decisions based on the data generated. This is where I-O psychologists can play a particularly helpful role in designing your systems.

As for the downsides, this approach require some knowledge and upfront design work to get the system right. Simply using any tools without establishing their validity adds risk to an organization’s decision-making process. Plus, these tools aren’t always the shortest or easiest to complete for participations. After all, you’re collecting data that actually predicts some powerful results, so the robustness does matter.

When done well, however, this approach has the potential to significantly improve an organization’s overall talent management processes and really pinpoint where future potential sits, and where the organization is placing barriers to deploying great talent. It’s the best method for uncovering hidden gems that would probably otherwise be overlooked.

Solving for Succession and Placement: Potential by Future Destination

This is the most targeted approach to classifying future leaders that an organization can apply. Here, the emphasis is truly being placed on “potential for what?” in comparison to the general high potential classification. The framework is most useful at more senior levels in the organization, where succession needs are prioritized over long-term planning horizons.

It’s also particularly helpful in organizations that have adopted a more generalist leadership framework within their functions and are looking to make a determination about each individual’s destination role.

In the finance function, for example, it can help classify potential based on those leaders who can go all the way to CFO, versus those who might have specialist paths to controller or treasurer as their senior-most positions.

For marketing, it might be used to segment classic marketers destined for CMO roles versus those with GM potential and interest to take on broader line positions. While an analysis of candidate slates can provide some of the same insights, by classifying talent this way, it’s easily apparent whether you have the right mix of specialists and generalists in any given function and which destinations are best covered by what types of talent.

Since this model requires a deep understanding of the talent needs for each function across the organization, the first step is ensuring you have a strong handle on the functional capabilities needed for the future of the business. Unlike the first two methods, the focus here is on segmentation by knowledge, skills, ability, and career interests, so you’ll need detailed information at both the job level and from the individuals in the talent pool.

Again, validated assessment tools are the best method for measuring these capabilities. But even simply understanding where people want to go and which types of destinations the organization has in mind for them (specialist or generalist) can be helpful.

Once the framework has been created and the right types of data collected, it’s a simple enough task to classify R&D or IT talent, for example, based on expertise and interest versus general leadership trajectory. The key is making sure you know which capabilities you’ll need in the future and how to measure them among those currently in the functions.

The results can be powerful when you see that you have a group of marketers, for example, who all want to be GMs, but none of whom have the right balance of skills to take on these types of roles. For those with considerable potential, it would indicate that significant (and perhaps urgent) development steps are needed to get them ready if they’re ever going to ascend to the GM roles desired in the future. Classifying talent this way can also really help with an organization’s short-term buy-versus-build strategy.

As with any model that relies on data, the tools used to measure capability need to be valid and accurate. But the real challenge is figuring out what the capabilities of the future are going to be. It takes courage and thought leadership to create a future state vision that provides the framework for defining what success will look like as you build this type of approach.

But get it right and you’ll take a major step toward ensuring the next generation of senior leaders are being placed in the right development experiences to close the gaps needed for them to reach their full potential.

How to Put It All Together

When it comes to talent management, we’re all looking for the silver bullet that will deliver the single answer of who is and isn’t a high potential. It’s never that simple. Having a high-quality performance management system is critical for driving performance today, but it isn’t particularly helpful in identifying future skills, capabilities, or potential. This is where other types of data and frameworks have more value to talent management efforts than annual performance ratings used in traditional 9 Box models.

These three approaches can help organizations get significantly better at classifying their talent and taking action (development or otherwise) from their talent management reviews. While some require more effort than others to establish, the most important commonality is that data is key to providing insights, which in turn drive outcomes.

And while the three alternatives have been presented as distinctly different models, you can easily combine them if you want to truly apply an integrated approach. The key determinants of success are the level of sophistication of the organization, the quality of the data being collected, the capability of those using the data to generate the insights, and the most pressing underlying strategic business and talent issues facing your firm.

No matter which new approach you choose, the old performance-by-potential paradigm is dead. Welcome to the future.

Allan H. Church, Ph.D., is the senior vice president of talent assessment and development at PepsiCo. Prior to his role at PepsiCo, he served as an external OD consultant working for W. Warner Burke Associates and several years at IBM in the Communications Measurement and Research and Corporate Personnel Research departments.