Key to analyzing your employee’s performance? Data Velocity.

9 min readApr 5, 2020

I initially posted this on LinkedIn August 2016. Starting to slowly move posts over here to Medium

Annual Performance Management

The performance management process can be painful and time-consuming. I know that’s an understatement for a lot of companies out there. A 2014 Deloitte University article found that 58% of companies surveyed said their performance reviews were a waste of time. It hasn’t gotten better since then.

Deloitte also opened up the conversation on performance management in Spring of 2015 with the HBR article detailing how and why they’ve dropped traditional annual reviews. The HBR article, Reinventing Performance Management, gives a great review of their transformation. It notes that “[Deloitte] found that completing the forms, holding the meetings, and creating the ratings consumed close to 2 million hours a year”, which is stunning for a company that bills by the consultant hour. And Deloitte’s far from alone. Let’s look at a few others that have changed course in the past few years:

Accenture | Accenture CEO explains why he’s overhauling performance reviews
Adobe | How Adobe improved performance by eliminating performance reviews
GE | Why GE had to kill its performance review after three decades
Goldman Sachs | Goldman is about to make life less stressful for employees
IBM | IBM changes its employee reviews
Microsoft | Microsoft abandons stack ranking of employees
SAP | SAP, Maker of performance review software, ditches performance Reviews

Looking over the list, this is not a group of “tradition be damned” startups looking to throw out the old without considering the consequences. These are companies that pilot, test, and analyze new programs before diving in. This is spreading past this group as well. In a recent post, Joe Ungemah cites Towers Watson as finding that “14% of companies have eliminated performance ratings, while an additional 24% are considering to go ratingless”.

Each of the articles above cites a number of reasons why these companies are dropping old systems (employee mental health, lack of results, enormous time sink, etc.). These authors cover that side of the argument better than I can. With that in mind, I think there are a few reasons for the switch that have so far been unsaid. In particular, one of the main reasons I see for switching to a new performance management methodology is the massive increase you get in the velocity of performance data.

One of the main reasons I see for switching to a new performance management methodology is the massive increase you get in the velocity of performance data.

What is the Velocity of Data?

For those unfamiliar, let me take a step back and explain why I keep saying Velocity. The three V’s of big data are the defining properties / dimensions of big data. These qualities, are what have been increasing as the world goes digital, which is what creates big data problems and allows for big data solutions.

The three Vs are Volume, Variety, and Velocity (sometimes value or veracity are included, but I’m setting them aside for now). Volume is the amount of data we have available, Variety is the type of data, and Velocity is the speed at which we receive the data. How much data, what types of data, and how quickly do you collect it?

HR has massive data sets about the company which are growing daily. As the volume of HR data has increased and technology has developed, we’ve seen companies start to leverage that data with HR analytics. By analyzing that high volume of data with algorithms, HR can uncover insights that the human eye wouldn’t be able to assess. Humans are not good at finding needles in haystacks, even when the haystack is in millions of columns and rows. However, algorithms excel at it when applied properly and have started to produce fascinating insights from the high volume of data already collected.

The variety of HR data is increasing as well. Text analytics, voice processing, and video analytics are the core backbone to a number of HR startups (Kanjoya and HireVue jump to mind). We no longer have to convert data down to a more structured form before we can analyze it and gather insights. As our capability to include a larger variety of data increases, more of our existing HR data is included in HR analytics projects. As HireVue shows in their demos, if we can input video interview insights into our predictive models we get a fascinating return.

The area within HR that has had little to no movement so far has been Velocity. I’ve been at conferences where I’ve heard HR practitioners complain that HR has “Slow Data”. Compared directly to consumer data HR data practically crawls. People don’t quit their job at the same rate as they quit products and they aren’t hired as often as they make purchases. I’ve also heard people say that we don’t have the Velocity of data required for proper big data analysis. One of the biggest villains of this slow data parade of horribles is the annual performance review.

This is where you’re asking, “Richard, are you sure we want faster data in HR when we barely analyze the data we have?”. Regarding the current “slow data” that HR has on performance management, yes, there is a lot of it we’re not currently using, but it’s often incomplete (employees come and go), outdated (analysis can take months), and heavily influenced by recency bias. Peter Drucker famously said “there is nothing quite so useless as doing with great efficiency something that should not be done at all”. I think that applies well to some of the current methods of collecting and analyzing annual performance data.

On top of that, the fact that we have “slow data” isn’t something permanent or outside of our control. To complain about it is to normalize the problem without taking action. We can control the velocity of our data.

Why hasn’t Velocity kept pace for HR data?

I’d argue that Velocity has been stagnant because it requires HR to change the way we’ve done things. In order to increase velocity, HR will need to collect data FASTER. In saying that, I know I just gave a few HR practitioners mild heart attacks.

To explain, collecting data faster in the current HR environment isn’t an option. No other part of the business has a more ornery, hard to pin down, and sometimes hostile data source than what HR has in the employee population. With the tools at our disposal HR departments are stretched to their limits and it is hard enough to collect the slow data that is already collected.

Running the current annual review process more often is also not an option. There are companies that have gone to semi-annual, but I can’t imagine that the process picks up enough extra data to evolve into a high velocity big data enabled solution. It also doubles rather than removes the inefficiencies that bog down performance management today.

So what can we do to increase velocity? If you look closely at the links above, the companies that are “dropping annual performance reviews” entirely aren’t actually eliminating performance management or even the review process, but instead have developed technologies which enable them to switch to a high velocity alternative.

GE has developed an in-house performance management app, Deloitte has regular check-ins and surveys managed by a central system, and Microsoft has been purchasing HR analytics startups who specialize in this very work and building out their own solutions. It’s a falsehood to perpetuate the myth that they’ve eliminated performance management. They’ve just shifted gears to a faster method.

Performance management in high gear

So what does the world look like on high velocity HR data? A unique property of employee performance data that often gets overlooked is that performance falls in a time series of data. Each point relates to the one prior and the one after and there are ebbs and flows. If you’re not assessing fast enough, you might get a very different result each time as changes occur.

Think about your fitbit or step tracker if you have one. If it only read out to you once a year how many steps you had taken, it wouldn’t give you much insight into how to change your habits. The critical piece of that technology is the velocity of the data collection. That enables you to know when you’ve been sitting too long reading HR analytics articles and that you should get up and take a walk.

When you can collect higher velocity data the gaps between data points in a time series shrink, which then lets a learning algorithms better understand the phenomena of employee performance. When an algorithm can make sense of your data across time, that’s when you can start to make predictions or better segment the employee population.

High velocity data is also key to automating employee performance analysis. High velocity processes require you to rely more on algorithms and less on human opinions which puts you on the road to automating many of the steps of the former performance management process

The Bad News

I want to focus on the current tools for a minute. HR is evolving to a new data-minded approach, but sometimes the tools are not keeping up. One recommendation I’ve heard for combating recency bias in annual reviews was to stop any time your employee did something well and write it on a post-it. Then at the end of the year, review all of the post-its. While this may work for the annual review process we currently have, it’s not a practice that’s statistically sound or bias-proof. If HR is using post-its it will be blown up before we know it.

Here’s the bad news, without a digital technology that is designed for the new way of collecting performance data, I don’t see how a company could roll out a modern high velocity performance management system. It would be akin to trying to do astronomy without a telescope. You can want to do it, see the benefits of doing it, and still not be able to do it without the technology.

This can be frustrating. For the most part in applying HR analytics we’ve been playing catch up and have a last mover advantage in applying HR analytics. We have the data, we can borrow or steal the tools to analyze that data, and from there we can find solutions. Performance management is a little different though. The tools necessary to capture, refine, and analyze employee data at the speed necessary to perform real time analysis just haven’t existed before now. I don’t see finance or marketing developing them for us either since human performance management is a fairly unique HR topic.

Moving Forward

The options that I see to move forward are to build, buy, or be left behind. The field is moving this direction rapidly. Many of those companies listed above developed their new technology in-house which is fascinating. It’s a new age of HR process as a strategic advantage. Many of those systems are still in their pilot or early stages and we should hear some results soon regarding their effectiveness.

If you’re not able to build in-house or don’t see this being a strategic differentiator, you’re still in luck. While the HR tech vendors have historically focused upstream (hiring) or downstream (attrition), there are quite a few coming out with new and innovative technology to tackle the midstream (employee experience and performance management). To name a few, Zen Workforce, Humanyze, and People Matter are doing incredible work in this space.

With that, I’m excited to hear your thoughts on this shift in performance management. What roadblocks or other benefits do you see? I’d love to continue the conversation in the comments below. Please let me know what questions came up as you read through this blog post.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — —

As always, a huge thank you to the readers for taking the time to think through this with me and to all of the writers and HR analytics enthusiasts out there who share their ideas with the world. It’s an exciting time to be learning in this area.

If you’re new to HR analytics, but this article has sparked an interest, please check out my HR Analytics Starter Kit series below. My goal is to get as many people interested and talking about this space as I possibly can. I also post as often as I can on Twitter at @RichardRosenow. I’d love to connect with you on there as well and share cool things going on in the HR analytics and HR Tech space.

— Richard

https://www.linkedin.com/in/richardrosenow

https://www.twitter.com/RichardRosenow

Other articles I’ve published on Linkedin: