Outliers and operational excellence: Rethinking our approach to knowledge worker peak performance
- 7 Min Read
Cole Napper speaks to HRD Connect on how organizations can shift their approach to measuring peak performance amongst knowledge workers and how placing a greater focus on outliers can drive outsized organizational engagement, performance, and productivity.
Operational excellence models such as Lean Production and Six Sigma have reshaped our approach to performance in manufacturing roles over the past 40 years. Knowledge-working roles are next in line, and this requires a re-think of how we approach and measure peak performance through people analytics.
Cole Napper, Founder and Host of Directionally Correct & VP of People Analytics at Orgnostic, argues organizations must make several shifts to achieve operational excellence in their knowledge-working divisions. Napper presents his recommendations in this Q&A with HRD, including:
- Replacing normal distributions with power law distributions
- Removing forced distribution performance reviews and focusing instead on outliers
- Refocusing individual performance reviews on divisional or team-level performance
1) What impact does differentiating between typical and peak performance have on employees and the organization?
Cole Napper: When people think about performance, they usually focus on the low-hanging fruit. Things that are easily quantifiable. Think of sales or manufacturing, where jobs are already highly data-fied. The challenge comes when differentiating between typical and peak performance when you’re looking at the knowledge-working space.
I use the concept of operational excellence, which has been in vogue for 30 or 40 years since the revolution of the Toyota Production System and things like Lean and Six Sigma. These systems look for ways to reduce waste or minimize errors, and how to make a 1% or 2% gain in typical performance among employees can lead to seeing an outsized impact on the organization’s performance.
But operational excellence literature historically didn’t account for the role peak performance plays in adjusting those averages or baselines. We’re now seeing research that performance does not follow a normal distribution. Instead, it fits according to a power law distribution, or in layperson’s English, the 80-20 rule. 20% of your employees account for 80% of your organization’s performance. It’s not to say that operational excellence is not important. Getting the 2% gain in productivity across your team is incredibly important. But we need a better understanding of how the peak performance of a small number of employees can make outsized impacts on an organization too.
2) How can HR leaders discern which metrics they should track to get a dynamic view of peak vs typical performance?
Cole Napper: There are multiple layers to this. A lot of typical people metrics that organizations track don’t provide a lens into peak performance. How many employees do we have? How many of them quit? How engaged are they? That’s just a quantitative decomposition of your workforce. But when you start to focus on peak performance, a good rule of thumb is that average is the enemy of greatness.
This means looking for outliers. In terms of the knowledge workers, this is where quantifying peak performance gets very tricky. Organizations measure typical performance once a year or every six months through highly biased performance review processes. There’s very little differentiation and sometimes even a forced distribution of performance. But if someone has delivered the most impactful project, presentation, or performance of their career, how is this reflected in the performance review process? It wouldn’t be.
3) What model or methodology would you recommend when using people analytics to determine peak performance?
Cole Napper: Firstly, do not try to fit performance into a normal distribution. This is not going to be an effective way of modeling your data. Instead, use power law distributions and use outlier analysis to identify the few inputs that account for a substantial or inordinate amount of value to the organization.
I would also dip into qualitative insights from a case study, interviews, or focus groups to understand these peak employees and/or experiences. Who are they, and how can we replicate that performance amongst the masses?
One area organizations typically under-assess is focus time. Compare the performance and productivity of someone on back-to-back meetings eight hours a day, five days a week, compared to someone with the ability to have two hours of focus time each day. The odds of them being able to put in an outlier or peak performance goes up dramatically. It’s important to note, however, that focus time is a necessary, but not sufficient, condition to see peak performance because there’s still going to be variability in two people who engage in the same amount of focus time in one time period. One may have demonstrated peak performance, and the other may not.
4) How can HR leaders utilize these insights to improve employee engagement, performance, or productivity?
Cole Napper: Most leaders look for negative spikes in engagement. Where engagement scores went down, and there is a burning platform to focus on improving areas that are underperforming. But who are the leaders studying functions with inordinately high engagement? Why don’t we study the peak performance leader who’s bringing up engagement scores of an entire department and applying what is learned to underperforming areas, rather than studying underperforming areas and trying to get them to improve?
In terms of performance itself, we need to move away from a forced distribution model of performance for ratings on one-to-five scales. Instead, reward employees who’ve demonstrated peak performance in areas where it matters for the organization. That doesn’t mean that everyone suddenly gets a five-out-of-five performance rating. And for productivity, again, look at what we call 10x teams. Too often we focus on analyzing individuals when we could analyze entire teams and business units. Conduct a store-by-store analysis or cohort analysis to look at groups of individuals and use hierarchical linear modeling to show which teams or divisions are off the charts on a linear scale and the factors contributing to that team’s effectiveness.
5) How can people analytics help HR leaders strike a balance between rewarding peak performance without neglecting the development of the broader “typical” workforce?
Cole Napper: I have a radical viewpoint on this. Take the example of any individual who starts a new task as a novice. They’re underperforming and not even achieving average performance. Most learning and development programs would teach them how to go from beginner to average. Using a 90-10 distribution of how organizations spend their L&D budgets in 90% of cases, that’s fine.
But what if 10% of L&D budgets were dedicated to taking someone from a beginner to a peak performer? If we focused a little more of our educational and coaching efforts with this mindset, we would see radically different results in the effectiveness of these types of coaching and L&D programs.
6) What emerging trends or innovations should HR leaders be prepared for in performance management?
Cole Napper: People talk about AI in very “magical” terms, like it’s going to fix everything. I would replace AI with the word automation. AI is going to automate a lot of work we do today, and the knowledge-working space can learn a lot from manufacturing over the last 30 years, where automation has been widely adopted. The impact on the manufacturing workforce has been a barbell distribution, so the extremely technical and extremely manual work remains, but automation has hollowed out a lot of the middle-skilled labor positions. This means in the knowledge-working fields, peak performance is going to matter much more at one end of the distribution, and softer skills will matter much more on the other end of the spectrum.