Last week I was invited to speak at the Australian Rostering and Scheduling Summit. That may sound like an unusual conference and an unusual place for me to speak, but I was absolutely thrilled to be invited and even more excited that they accepted my proposed topic to speak on, which was (… drum roll please…) “The Ethics of Using Data to Manage Employees”; AKA employee data.

Now, I have firmly placed serious questions about my geekiness in your heads, I’d like to confess a lifelong passion for this topic of data – the good, the bad, the potential, and the dark side. It fascinates me. Inspires me. Scares me. And challenges me to think it forward.

Touch points 

Here’s just a couple of touch points in my historical love affair with the topic of data, all of which happened in teams I led:-

  • One of the first Australian companies to use ‘bot interviews’ in recruitment.
  • One of the first Australian companies to use metadata correlations to analyse and predict success post recruitment.
  • One of the first Australian companies to put full datasets into Talent processes, right to Board level, and link assessment (psychometrics, 360, diversity, behavioural and other data) to making talent decisions.
  • One of the first Australian companies to look at correlations from engagement and NPS scores to business performance and also to leadership assessment, over time.
  • One of the first Australian companies to look at using absence data, and patterns, to predict disengagement and resignation, and to consider allowing leaders opportunities to intervene (the jury stayed out as to how far we could go on this one given the wide variance in leader capability and intent).

You see, I see data and the technology behind it, as a source of awesome potential. That said, I’m not silly. I can well see the potential risks….or some of them, anyway.

Let me give you a summary of where I see this at the moment.

But, before I do, work with me a little. Think of two leaders.

Two Leaders 

Leader A is that leader you love. Smart and able to wrap their head around tough topics and data, a great coach, and phenomenally empathetic. All wrapped up together. They build a great team, and create an environment of trust. They start with an assumption that people are good, and their job is to create the right environment to thrive.

Of course, all those lovely characteristics, rarely exist in one person, but let’s just imagine they do for a moment.

Leader B is, unfortunately, that leader we’ve all had. They’re pretty hopeless with numbers, shadow box at practically anything, don’t provide clarity or resolve conflict and as a consequence, their team sits on the edge of warfare most of the time. They start with an assumption that people are bad, and need to be controlled.

Sadly, those characteristics often find themselves together in one person, so not as much imagination is required.

3 Questions

Now, with those leaders in our heads, let’s look at data, or more specifically employment data, and let’s look at three questions –

  1. What data we collect?
  2. Why we collect it?
  3. What Leader A and Leader B might do with it?


Rather than write a tome, (there’s an awful lot of data out there and its growing) I’m going to visit some thoughts in point form, and I’m only going to talk about data we actually already collect.

  1. Attendance Data.

What employee data is there? Attendance. Days and hours your work, absences, and patterns of absences. With embedded ID chips, (as some companies already allow) then we can see everywhere they go, not just work.

  1. Bio Data from Wearables

What employee data is there? Heart rate, blood pressure, geo-locations, exercise.

Why we collect it? We want to make sure our team is well, and getting good exercise, and aren’t experiencing huge amounts of anxiety at work.

What might Leader A do with this employee data? Leader A, will watch the data as a team. They’ll be conscious of privacy and be interested in the aggregate across their team(s). They’ll encourage education on health and well-being, and will choose topics that best address the aggregate development areas. If they have one person under extreme stress, they’ll reach out to them quietly, and offer some more support. We all need that sometimes. 

What might Leader B do with this employee data? Leader B will check the data weekly, and announce in the team meeting who’s the fittest and ‘most well’ and who’s the laggard. They’ll note with interest, what time people sleep, and call out who sleeping well and who’s not. They’ll speak to the recruiters, and mention that fitness is really important, and they might even mention it to their Insurance Division, that they’re certainly seeing some unfit people on their team. They might correlate fitness to performance, or they might not, but they might assume good heart rate and low blood pressure will lead to a poor employee. There’s no evidence of this, but it might be good for the insurance team that they request employee data for better understanding of who should be employed and who should not.

  1. Facial Recognition and Movement

What employee data is there? We can use Facial Recognition as a security screen, and linked to Attendance. We can also use it, and some companies are testing this, to spot employees as they come into work, and actually look at their ‘mood’ based on their facial expression. We can also use ‘eye movement’ and other facial movements in VR to better understand reactions to situations, and as an additional data point potentially indicating stress or anxiety.

Why we collect it? Technically, it’s an ID option, that saves cards and chips being used. As we start to collect it for ‘moods’ or state of mind, we will need to look at what ‘normal’ is, for both individuals and the collective. With VR, it is used to understand reactions to scenarios and situations (e.g., comfort, discomfort, averting eye contact, etc.)

What might Leader A do with this employee data? Security ID aside, Leader A may be given direct access to the ‘facial expression’ information during VR training, and will use it within broader analysis and understanding of reactions to scenarios. They will use aggregate team data, during the team discussion around scenarios. If a person looks particularly anxious, they may open a casual conversation to check in.

What might Leader B do with this employee data? Leader B, may look at the facial expression data on the team as they arrive, and go directly to speak to people who look anxious or concerned, as they arrive at their desk. It will be an awkward and unexpectedly personal conversation, potentially when the person is already feeling someone out of sorts that day, or not, when their facial expression was due to not having an umbrella that day when it was raining.

That’s just three pieces of data. All with potential practical positive uses, and all with potential oddness. That’s before we add Social Media Data, which many organisations are collecting and trying to make sense of, and a plethora of other data possibilities we could use.

Why we collect it? We originally collected the data for pay and performance, but we now use it for patterns of absences, and we can use these patterns to predict disengagement and eventually resignation. We can say things like ‘People who have four Fridays off in a row, are somewhat disengaged, and after six Fridays, will resign 80% of the time”.

What might Leader A do with this employee data? Leader A, will watch the data, and note two people working excessive hours, and someone else who may be less engaged. They’ll check in with all three. They might discover, the person having Fridays off is actually totally engaged, but unfortunately, has an ill father and needs to change over to a four-day week for a few months, and they’ll help them do this. This might discover the two working excessive hours are working on a project they love and are powering through it. They’ll make an agreement about how long the extended hours will continue and then follow up to make sure things are under control.

What might Leader B do with this employee data? Leader B, will look at the exception report, and pick up the person who’s taking Fridays off. Without checking in, they’ll decide they’re quite disengaged. They’ll assume it’s the same problem they’ve seen before. They’ll take the person off training, as they’re sure as hell don’t want to invest in someone who’s got one foot out the door. They’ll stop spending time with the person, and the person will quickly work out that they’re on the outer. Unbeknown to their manager, this will add to the stress of trying to care for an ill father.

Now let’s put all three together.

The Leader, A or B, will know your geo-location, bio data, health, and your facial expressions every minute you’re at work, and potentially your geolocation and bio data outside of work. They can add the psychometric data from recruitment or promotions. They can add performance scores, salary data, interview examples, and your social media views and posts. They could your data from a previous employer, or two, long before you learnt to lead. They could add the data from the teams you’ve led, across multiple companies, and check how they’re all travelling now to make sure your impact on others is what you say it is.

In fact, they have so much data on each person that they have to be overwhelmed, and then there are ten people on their team, and they all impact each other, so multiply it many times over.

Do we know what capabilities, mind-set and bandwidth Leaders will require to deal with all this data?   Do we know what potential correlations mean, if anything? Do we have norms that make sense with individual difference, let alone when we add in all our different disabilities?

That’s even looking though the eyes of our friend, the good Leader A, let alone adding in the ‘how can I control people’ mind-set of Leader B.

The good news is that at the Rostering Conference, I was surrounded by a lot of people who love this topic and are already thinking forward.

There was great knowledge and a good debate in the room.

No one wants fear mongering, but we do need open conversations. ’

We want better understanding of not only intended consequences, but also unintended. Not just ‘what is this data for?’ but equally, ‘how could it be misused?

‘what data should we never add it to?’ and ‘what do we need to train Leaders in, so they use the data competently and appropriately’?.

Encouragingly, around the room, there seemed to be general consensus that ‘opt in’ was the way to go. Explain to people what you want to collect and why, and explain the potential uses. If people agree, then you can collect and use the data as you intend. But you cannot sell it, or on-forward it, or add in other potentially unrelated data, without their permission and knowledge as well.

Of course, amongst all that agreement, there was another very human question – “What if you don’t want to opt in, but everyone else does? Will there be there social pressure to ‘opt in’, is it really your choice, and will you not be included in the team if you don’t ‘opt in’?”

And that’s the reality of employee data and any rules we put in place. There’s the written rules, the plans, and the very positive intentions. Then there’s reality, and accepting that it’s never quite as perfect as we planned.