Steven Forth is a Co-Founder of TeamFit. See his Skill Profile.

 

TeamFit’s goal is to help individuals, teams and organizations to understand and then reach their full potential. What do we mean by ‘potential’ and can it be measured. This is the theme on an important new article in MIT Sloan Management Review by Reb Rebele.

Can we really test people for potential?

The article is organized around two themes: “People Metrics Are Hard to Get Right” and “What Would It Take to Go Granular?” These two themes are central to TeamFit and it is worth thinking through each of his points (please read his article as well).

The headings below follow those of the Rebele article.

People Metrics Are Hard to Get Right

Not all assessments pass the sniff test

I have a difficult relationship with the common personality tests. I don’t really think Myers Briggs or its descendants like Colours or DISC measure anything meaningful. The recent book The Personality Brokers: The Strange History of Myers-Briggs and the Birth of Personality Testing by Merve Emre provides some useful insights into the origins and deceptions that are part of these tests. Despite these views, I am still an enthusiastic participant in everything from Belbin and StrengthsFinder to DISC. I even put my Belbin and StrengthsFinder strengths in many of my profiles. They can be a tool for self reflection and can frame conversations with other people on my teams. Do they measure potential? I don’t think so. The only thing that reliably seems to measure potential is IQ tests, and in the work context, where diversity of approach is often important, raw intellectual horsepower is not always the most important thing.

People often differ more from themselves than they do from one another

My performance on any task is highly contingent. It depends on my level of stress and fatigue, what I have just been working on, and, most importantly, on who I am working with. It is this internal variation that is the well of our potential. Just as variation is the critical input into natural selection, variation is the key to uncovering and developing potential. This is one reason we have so many different skills in TeamFit skill profiles. Most people have more than 100 different skills at all different levels.

“Variation indicates potential.”

People change — and not always when you expect them to

Change over time matters. Our skills and potential are not static. All of us are developing and adding new skills everyday, or we should be, and we should be actively seeking out opportunities to do so. At the same time, the skills we are not using are decaying. Skills have a half life. Some take a long time to decay (once you have learned to ride a bike, or to swim, you are unlikely to forget the basics).

The other thing that can dramatically change our skills is who one works with. This is why TeamFit tracks who you use your skills with, and in doing skill interviews, one sequence of questions we often as is as follows:

  1. Who do you work with in this role?
  2. What skills do you use with this person?
  3. What skills do they use when working with you?

The nature of the task can matter more than the nature of the person

TeamFit is still working out how to model ‘tasks’ within a skill management system. We are able to use our skill inference engine to map from tasks to skills. And we can connect skills to a role. It is an open debate on our team whether we need to be able to add tasks to our data model and if so where. That said, there can be no doubt that …

  • Different tasks require different skills
  • The same task can be accomplished with different skills

Systems that are too rigid in mapping roles, tasks and skills are likely to shut down the variation needed to discover and realize potential.

What Would It Take to Go Granular?

Rebele goes on to look at three themes for how we could improve people analytics, or as he might say, ‘person analytics.’ These are ‘context,’ ‘variability,’ and ‘data ownership.’

For starters, we would consider the context

Skills have meaning in context. A single measure of any kind is of little use without an understanding of the context. This is especially true of something as essentially human as a skill. Skills have meaning in the context of their relationship to other skills, to jobs, roles and projects, and especially as a way to connect people with complementary and connecting skills.  See A skill name is just a word. Isn’t it?

“Skills can connect people.”

Next, we would design new measures with variability in mind

Here Rebele is referring back to his second and third points, “People often differ more from themselves than they do from one another” and “People change.”

The variability Rebele has in mind would seem to be two-fold, the natural variability in all of the factors that shape performance and the way these change over time. To push on the evolutionary metaphor, one could say we are interested in the variability within the genome of skills and how that variability gets expressed in different relationships, roles and projects, and in how both the skill genome and its expression evolve over time.

At TeamFit, one way we represent variability is through uncertainty. Our measurements are all Bayesian and include both a value and a measure of confidence (in the UI, the measure of confidence is signaled by the degree of saturation in the Skill Map. I had been thinking of the degree of confidence as a measure of the reliability of our own measurements but I am realizing it has two components, a measure of the uncertainty of measurement and a measure of basic underlying variability. Attempts to eliminate variability would likely lead to a less accurate system.

 

 

The temporal question is even more interesting as it is over time that we can see how skills and relationships are evolving. This is an area we are working on in TeamFit, and we hope to be able to put some of our insights into the user experience later this year.

Finally, we would give people their own data in ways that would help them develop

“Give people their own data.” I am not sure how far Reb Rebele means to take this, but there is the beginning of a movement that believes that users should own their own data and lend it to the organization(s) they are currently working with. This turns the current orthodoxy on its head. In most companies, it is assumed that people data is the property of the organization and not the individual. The leading companies in the skill area, Degreed and TeamFit are aligned on this and see learning records and skill records as something that belongs to the individual and travels with them along the course of their career. See “Ownership of data in a collaborative age: Three unworkable approaches and a way forward.”

“Data privacy issues are better conceived in terms of data ownership.”