What is the ultimate goal of People Management? Happier people? Well, maybe, for a purpose. And what is that purpose? Performance. We in HR all work on producing the best environment, choosing the best people, working out the best processes, rules, and regulations, conducting the best development programs, all in order to help our employees achieve the best level of performance they can so that our organization can be successful.
This implies that Performance is the principal metric that we use. At the individual level, everything concludes in performance. Performance is the building block of organizational success. While well-performing people still need to be organized and managed to form a successful organization, management and organization are actually tasks that are performed by individuals, so these requirements become part of their performance.
Surely we haven’t exhausted one hundred years of organizational theory above, but one thing we should agree on: Performance is one of the most important personal metrics, possibly the single most important one. Every definition of Performance might differ a bit from the other, and what is common in all the different definitions is possibly the aim to express the topmost expectation towards each employee: “how much” (or “how well”) this person does what she or he has been employed to accomplish. We actually often mix up Performance with Productivity and even with Effectiveness and Efficiency – forgive me for not going into the details this time.
I would like to focus on the reliability issues that frequently concern Performance measurements. In many cases, what we witness and should try to circumvent, is that we have some data which is called Performance but it just doesn’t seem to provide usable information about the actual performance of individuals. The symptoms that are the most easily noticed are small range: for example, almost everyone having performance values between 8 and 10 on a scale of 10 – which may be very effective when scoring boxing matches but less useful in HR. (; with a strong skewedness (similar to the small range case: the scale is far from being evenly used).These basic flaws can fundamentally undermine the usability and validity of all the calculations that utilize our Performance measurements. It is important to note that missing values, even if they look harmless, can bias or distort Performance based calculi when occurring systematically. In other words, when there is a hidden pattern or correlation with other factors in the occurrence of missing data.
There are problems with Performance; problems that exist in the information that the numbers hold and are less easily discoverable and often more dangerous. It happens that the numbers “just don’t make sense”. These numbers are there, they have been produced in some way, but when you think about the real person A or the real person B you have a hard time associating the measurements with your personal impressions of them. This latent feeling may be your first and only clue that the data is not reliable. What can you do? Go and investigate how the Performance measure was produced. In the case of such positions as production/assembly line workers or salespeople it is certainly possible to obtain hard, totally reliable numbers that testify to the Performance of the individual. But when it comes to everyone else in the company structure you cannot measure Performance in such a simple way. For example, what office workers do, generally, is processing information, often in creative ways. Some information comes in, they add their expertise to it, then they send the new, enriched information to other people – who then add their expertise. This process can be repeated as the information wends its way up the management ladder. A graphic designer will take a brief as input and will output drawings. A manager will take business intelligence and spit out decisions. These are not very easily measurable activities. One can try to measure the quantities that occur, but the results will not evaluate the quality aspect. Measuring the quality of the output only is not an objective measurement of Performance. Trying to measure both quantity and quality is what I would call the desirable practice, but it is not easy to accomplish.
In many cases, for obvious business reasons, management decides to measure the monetary success of departments, teams, or where possible, of individuals. After all, the purpose of every employee should be adding to the overall success of their company (team, etc.), right? That is very true, however the problem with this kind of measurement is that it very often factors too many conditions into the measurement that are external to the subject – that is: stuff that doesn’t depend on them. Someone’s performance might come out low because their department has not made enough money in the last quarter. Some other people might have constantly high performance because they happen to work in a profitable business unit where market demand is high.
Finally, we often learn that what a company calls Performance measurement is rather the product of 360-style assessment – they might take the part of the evaluation that was done by the direct superior(s) and call it “Performance measure”. Often it’s done in an even more subjective way by simply assigning a number that is freely decided by the line manager of each employee. The problem with this approach is exactly its subjectivity: these are opinions, however important they may be, but they are not measurements. Let me be clear: we at Assessment Systems use 360-based data as one of the foundational parts of developmental evaluations, and we call it Reputation. But we are careful to keep Reputation and Performance measures separate, and resist the ever-present temptation to confuse or conflate them.
I believe the best advice would be to not blindly trust your Performance data. Don’t take it for granted, as when it’s flawed, all the analysis that you based on it will be flawed, which will result in flawed decisions and misguided strategies. What can you do in order to adjust your Performance reporting system and make it more usable? Don’t be afraid to reinvent it! Experiment, run A/B tests, try different approaches and see how each works. Try using different measures for different job families (but remember that it’s a double edged sword as we mentioned in our recent blog post “Less ammo…”). If needed, construct a complex Performance calculation method. Use personal goals defined in peer reviews as indicators. Perhaps most importantly: clarify the expectations assigned to each job family. Take the time to examine which are the behaviors that will best serve the success of the company, organize them into expectation systems and keep those expectation systems both transparent and up-to-date.
If everything fails – first of all admit that everything has failed and don’t cling on the false illusion that you have the holy grail of Performance measurement in your hands because it is likely that any sane person in your company already has an idea that you don’t. Throw away your Performance numbers - temporarily omit them from your business intelligence analyses. Go back to the drawing board and build a system which is usable. Your HR, your data scientists, and more or less anyone around will be more than glad and able to help you with that – not to mention your HR consultant. While not having a working Performance measurement system, use your 360 reviews (but don’t call them Performance data), or even better, measure and use Engagement (arguably the closest precursor of Performance) because it can be measured rather easily and reliably.
HR R&D specialist