98: How to predict employee behaviour using Predictive Behavioural Analytics, with Peter Dorrington, Anthrolytics.io

Peter explains how his groundbreaking work combines data science with behavioural science to provide insights into employee and customer behaviours.

We explore the concept of empathy at scale, predictive analytics, and the ethical considerations surrounding data privacy.

Guest Bio:

Peter Dorrington is the Chief Strategy Officer and Co-Founder of Anthrolytics, a company that combines behavioural science with data science to offer predictive insights into employee and customer behaviours. With over 30 years of experience in data science, Peter is known for inventing predictive behavioural analytics, helping organisations create environments where people thrive and profits soar.

Key Quotes:

– “Empathy is the biggest driver of loyalty when other factors are broadly equal.”

– “Just because you can, doesn’t mean you should. Ethical considerations are paramount in data science.”

Resources Mentioned:

– https://anthrolytics.io

– Peter Dorrington on LinkedIn: https://linkedin.com/in/peterdorrington

If you enjoyed this episode, please leave us a review and subscribe to stay updated on future episodes. Have a question or topic you’d like us to cover? Reach out to us on social media or through our website.

Join us next week for our “This Week in Work” episode, where we round up the latest in the workplace world and answer your questions. Don’t forget to subscribe and leave a review! See you on Tuesday!

---

General Support with Mental Health and Well-being

If you have been affected by any of the themes in this episode, or are currently struggling with your mental health, the following resources may be useful.

 Mind website: https://www.mind.org.uk/information-support/

If you are feeling in distress or despair, including feelings of suicide, please do consider calling the Samaritans for free on 116 123 (UK) or email jo@samaritans.org (Rest of World)

Resources

All the links mentioned in the show.

Connect with your hosts

• Connect with Al on LinkedIn 
• Connect with Leanne on LinkedIn
• Join the discussion about this episode on LinkedIn
• Email: podcast@TruthLiesandWork.com
• Follow us on Instagram @truthlieswork
• Chat with us on Twitter @truthlieswork
• YouTube channel for the podcast @TruthLiesWork
• Check us out on TikTok (LOL!!!) @truthlieswork

Related Episodes

Loved this episode? Here are some more you might like:

The Transcript

⚠️ NOTE: This is an automated transcript, so it might not always be 100% accurate!

Peter Dorrington: And I think a lot of it comes down to trust. I’m not going to abuse our relationship because trust arrives on foot and leaves in a bus with all your mates. And I think that, you know, it’s hard won and easily lost. And if you do something with this technology, which is an abuse, which is unethical, we do our best to Actually make that impossible, as I say, because we don’t take PII.

We don’t look at anything which is confidential. We don’t want anything that is embarrassing for people. We are genuinely trying to give them a better experiences. We don’t work with any organizations that I don’t feel strongly meet that ethical framework. And so because what we do is unique, I don’t think at the moment anybody else could do it.

So short answer, just because you can, doesn’t mean you should.

Leanne Elliott: Welcome to Truth, Lies, and Work, the award winning psychology podcast brought to you by the HubSpot podcast network, the audio destination for business professionals. My name is Leigh Anne. I’m a business psychologist.

Al Elliott: My name is Al. I’m a business owner.

Leanne Elliott: And we’re here to help you simplify the science of work.

Al Elliott: Yes. Welcome back. This is Thursday. This is our interview edition. You know this, you know this. Tuesday, workplace surgery. Roundup in Work, Thursday’s Interviews. Leigh.

Leanne Elliott: Yes, today we have such an interesting conversation for you. What would you say if I told you there is technology out there that will tell you which member of your team is going to quit before they’ve even made that decision?

Can you imagine the headaches you would avoid if you could identify tensions in your team Before the first argument even breaks out. And what about stress? How killer would your employer brand be? If you could tell that ambitious Gen Z candidate that you nurture wellbeing by reducing stress in the workplace before anyone even begins to notice that increasing workload is having a psychological impact, I’m guessing at this point, your left eyebrow is firmly raised.

Well, today we will be exploring all of that. All of these possibilities as we dive into the world of predictive behavioral analytics. Our guest today is Peter Darrington. Peter is an experienced management expert, an international keynote speaker and co founder of Anthrolytics Limited, a SaaS platform that combines behavioral science with data science to provide companies with insights into operational and external factors that drive employees behavior and link that importantly to financial outcomes.

It is a potentially groundbreaking opportunity, an opportunity for businesses to create environments in which people thrive and profits soar.

Al Elliott: Yes. Yes. Today we’re going to talk about empathy at scale and how we can achieve that in the digital age. We’re going to get Peter to explain exactly what those predictive behavioral analytics are and why it’s the basis of everything he does.

We look at what to do with this data once we’ve actually collected it. And finally, we’re going to grill Peter on the privacy issue. Issues or privacy? Privacy? I’m not sure which one it is. Privacy issues around this kind of data project.

Leanne Elliott: Let’s go and meet Peter Durrington.

Peter Dorrington: I’m the chief strategy officer and co founder of Anthrolytics, a new startup company.

But my main area of expertise and what I’m known for is being the inventor of predictive behavioral analytics. So, um, my background, 30 years as a data science professional advising businesses on how to use data and analytics in everything from understanding customers to understanding employees and suppliers.

But 10 years ago, I started to wonder, Can’t we do a better job of understanding individual people? And that’s when I got into behavioural science. So Antalytics is that company that’s combining data science with behavioural science and is something I called predictive behavioural analytics. And I’ll explain that, hopefully not too technically, as we go on.

The vast majority of human beings have an ability to see something from somebody else’s point of view. And we typically call that analytics. Empathy. So being able to walk in somebody else’s shoes. Now there are actually three parts of that. The first of which is understanding. What the other person is feeling and why we call that cognitive empathy, then there’s our emotional response to that.

Um, and that’s our emotional empathy. So we feel a bit about what that person is feeling. Now that’s different from sympathy, which is I have feelings, for example, of being sorry. that something’s happening to you, but they’re not the emotions you’re feeling. But the most important part is compassion, which is where we’re moved to take action.

Now, we can do that one to one, and many businesses have invested a lot of money in emotional intelligence training, empathy training, so that their human beings can have a good experience with another human being. But when we all moved into the digital world, we broke that human connection. And many businesses have recognized that their big differentiator in that world is that experience, both for their employees and for their customers.

And they’re trying to recreate it. At the digital level. So instead of being able to do it one to one, they want to be able to do it for thousands or even millions of people all at the same time. Um, and clearly that requires something different. So the challenge with traditional approaches to empathy is It scales on the basis of how many people you have.

But when you’re in the digital world, you need scaling that is not linked to the number of employees that you have, or the number of customers. So empathy at scale is where we take the understanding of what people are feeling, so that’s the cognitive empathy bit. And we can do that with modern technology, and I’ll talk about that just in a moment, and then taking actions as a result of that, and we can automate some of those actions.

Obviously, what we can’t do with systems is teach computers how to feel, but what we can do is teach them to feel. What people feel and why, and what they’re supposed to do about it. And we’ll still leave the human in the loop bit for the bits that are complicated, very emotional, where you have to deal with somebody on those one to one calls.

For example, customer services, great place to be able to display empathy. And, um, you know, the evidence is quite clear that empathy is the biggest driver of loyalty when the other factors are broadly equal. So if the price is roughly the same, convenience is roughly the same, Empathy is what’s driving people’s decisions to work for you, stay working for you, stay motivated and productive.

So sorry, quite a long answer there. But empathy as a scale is trying to replicate some of what human beings can do at the digital scale.

Al Elliott: Tell me what would a company or an organisation look like who understood empathy at scale and were implementing systems? We’ll come into that in a second. How would that compare to someone who has never heard of this term and doesn’t do it?

Peter Dorrington: The organisations that are having success with this firstly have Lots of data. You’d need data to train the systems, and I’ll explain how this all works in a moment. Um, and you also need to be of a fairly large scale, because if you’ve got three employees, you can probably manage that yourself. But if you’ve got 50, 000 employees and particularly If your employees aren’t all working in an office, so if they’re distributed around the country or even internationally and they’re remote working, one of the big benefits for this, and this is where it’s created a lot of interest, is being able to remotely diagnose, I don’t think Peter’s in a great place right now.

I might not see a colleague for days on end, possibly even weeks. So apart from feelings of loneliness, where’s my support network? And how do old idiots like me who are taught management by walking around translate that? into the remote working or hybrid working environment. So they tend to be larger organizations with a lot of data.

Now they don’t have to have huge amounts of data. There are other techniques that we could use, but that at the moment seems to be the most popular deciding factor. I’ve got a lot of employees, we’re suffering from, for example, the attrition problem, so we pay really well, we’ve got great benefits, and we’re still hemorrhaging staff, and we’re not quite sure why.

Leanne Elliott: I love this idea of empathy at scale, and especially through the lens of, of behaviours and behavioural science and data analytics. I think empathy can often seem a bit fluffy, dare I say it, but I think anyone that has that view has fundamentally misunderstood empathy. The role of empathy in organizational life.

There is a growing amount of research that highlights the huge impact empathy has on employee engagement, wellbeing, and performance. Just to give you a very brief overview. In 2021, there was a report by Catalyst that found that employees with empath, Empathetic managers are more likely to report being innovative.

There’s a 2018 study published by the Journal of Organizational Behavior that reports that employees with empathetic supervisors demonstrated higher levels of job performance. And when it comes to turnover, the really big problem, a report by Business Solver found that a whopping 92 percent of employees would be more likely to Stay with that current employer.

If they empathized with their needs, I can’t stress it enough. Empathetic leadership and empathetic organizational practices are critical to the success of your people and your business and predictive behavioral analytics offers a really powerful solution to embed empathy into everyday working life.

Al Elliott: So as we mentioned before, the basis of analytics is this phrase predictive behavioral analytics, but I wanted to know what this was and how it can help with say quiet quitting or surprise resignations. I’ve got written down here a phrase predictive behavioral analytics. Is that the basis of it? And how are you getting data from someone sitting at home and understanding how they feel.

Peter Dorrington: So this was the bit that took me three years of failure to finally crack. Like everybody else, I tried to come up with a prediction that was based on using lots and lots of data and feeding it into artificial intelligence and regression modeling and all of those great data science things. None of them worked.

And it was incredibly frustrating, but the real problem was Complexity. Human beings are incredibly complex. And one of the things that said my original question was, how can I think of people as individuals, not as big homogenous groups? So, um, after three years of failure, I went back to the drawing board and I came up with an approach which has two very distinct phases.

The first phase is to use where people are talking, tools to understand what are they talking about. And these tools are all very widely available. This is natural language processing, for example. But more importantly, how do they feel about what they’re talking about? So being able to identify underlying emotions.

So as an employee, for example, I might say to a colleague, you know, I was called in on my day off to work a shift because we had people off sick. And, you know, I was really upset because I was intending to spend time with my family. There’s a lot of emotions in that. So those moments where enough people talk about broadly the same thing and generates emotions, I call those moments that matter.

Now everybody has a definition for that, but for me it’s specifically what are things, interactions or events, that generate emotions. And we find that by listening to what people talk about, not by snooping on them, by the way, but reading the things where they would expect it to be analysed. So, for example, surveys where there’s a, you know, any comment section or people’s evaluations of an interaction or even a transcript of a call.

You know, they say on those calls, this may be recorded for training purposes. That’s perfect. That’s exactly what I want. So moments that matter, any interaction or event that generates an emotional response. So the first big problem is it’s impossible in many cases to draw a direct line between a cause and an effect.

I had a bad discussion with an employee. They immediately resign. That rarely happens. Normally what happens is over time, those negative experiences accumulate and all the emotions that are associated with those, and eventually we reach a point where my wife said, That’s it. I’ve had enough. You’ve been getting on my nerves for 40 years.

It’s time that you need to leave. That’s the first part. Now, how do I make it predictive? Imagine that journey over time. And I do something I know that, for example, a customer doesn’t like. Like I put my prices up. Now I can hear from a few customers we really didn’t like that, that made me angry, or I was surprised, so I identify putting the prices up as a negative moment that matters and it generates these emotions.

I don’t have to phone every customer and ask them, how did you feel about that? Now I’ve got a group that says, well, if, you know, a sizable or a significant number of people say it’s important and how they feel about it, I can make an assumption that it’s important and significant to other people. And what I do is saying, how did Peter feel yesterday?

And how has this moment that matters affected or changed the way that they feel? And then to make it predictive, and is Peter approaching a threshold? Where he’s likely to do something that we can observe like resign or actually take on more responsibilities or give some discretionary effort. So the first thing to do is to say is to work out what’s important to people.

And by the way. We often don’t overtly say what those things are. We might say one thing in a survey, but actually feel something quite different. But once we know what those are, then what we do is for every employee, we look at, well, have any of those moments that matter happened to any of our employees?

And how will that have affected the way that they feel? So Peter was angry yesterday. We’ve just done this. He’s going to be enraged today. And that’s the point at which he’s probably going to throw his resignation letter down on the desk and walk out the door. Now, the good thing about this is long before you reach that threshold point, we’re observing your day to day journey as an individual.

And we can say, actually, we think Peter’s getting to the point where he’s feeling a bit angry, a bit down or whatever the emotion might be. Let’s intervene. before that becomes a crisis. So I don’t have to wait until Peter is displaying very overt signs of being stressed out or angry or very depressed. I can actually say Peter’s on the journey, but we can head him off at the pass if we take some corrective actions or we build on some of the things that he really enjoys.

So I don’t go back and reanalyze Peter’s whole history. I just say, how did Peter feel yesterday? What’s happened since? And now how does he feel today? In banking, we found there was a group of customers I called the overlooked and neglected. Now, this is an empigraphic segment. These are customers who don’t feel great about the bank.

They’re not actively looking to switch. So they’re perceived as being loyal because they’ve been around for a long time. But if you were to talk to them, they would say things like, well, you never approach me unless you want to sell me something. You reserve all the best offers for all the new clients.

You know, I don’t get any real value out of this. And we did some experiments. So what we did is we took those group of Historically loyal, but mildly disaffected customers. They were very disengaged with marketing. They don’t buy any more product. And we moved them to the point where they got to neutral.

They said, well, actually, you’re no better or worse than any other bank. You’re doing okay. That had an incredible positive impact on things like people taking up more products and their responsiveness to marketing campaigns. So we just took them from that mildly disaffected group, made them feel neutral.

That I think everybody would understand. That’s intuitive. What was counterintuitive was we took a very similar control group that were neutral. And we said, well, what happens if we make them happy? So we spend a lot of money investing in campaigns to make those people feel really good. Hey, you’re a great bank.

You know, I really enjoyed that offer. You sent me, we got them to the point where they were happy. It costs quite a bit of money to get them there because we had to give them a lot of value to get there. They didn’t buy a single darn thing extra. So from a campaign point of view, it’s a disaster. We’ve spent a lot of money making people happy, but it had no behavioural effect.

So that’s really where we’re getting to this point to say if we can make predictions about people approaching thresholds, but we don’t wait till they reach a crisis point. We intervene or jump in when it’s earlier, and it’s usually cheaper and easier to affect a change. So don’t wait until somebody’s got one foot out the door and they’re leaving to try and win them back.

At best, you’ll buy a couple more months of their attention. But realistically, they’ve checked out. But if you knew somebody was feeling a little bit down and perhaps they’re not feeling great about the company, long before it’s that crisis point, intervene. See if there’s something we can do as a business.

that would give them a better work experience, better employee experience. And many of those things actually don’t cost a lot. They’re really just making people feel more valued, more understood. There’s the empathy thing again. And that we are a humane organization that really gets them. And that’s actually for most of us, what we really want out of the world.

And we know that businesses have limitations, they can’t pay us everything we would like, but a lot of values approach to it. Well, I think they get me and they’re trying to give me a work experience or an employee experience that makes me feel that way, that I’m valued and understood and listened to and incorporated into the process.

What we’re obviously trying to do is reduce things like attrition. So people who resign, the unplanned. Staff turnover. But what you could see that before people walk out the door, there’s often a prehistory to that, which is they start taking time off. They, you know, they just call in sick. They go home a bit early.

They disengage. They’re not as motivated or productive as they were before. And all of those are predictable. So the precursors to the point where they actually walk out the door, we can do things like tell you how many of your employees are probably going to take an unplanned day off in the next, you know, 30, 60, 90 days.

And we’ll assign a probability to that, as I say, because it’s not an absolute de facto, 100 percent accurate model. And there are one final limitation is you may be just having a really bad day. Reversed out of your driveway in the morning and ran over the cat. There’s no way I know that, right? But today is not going to be a good day for you.

So if I happen to talk to you today, I’ll get some of that benefit. So of course, there are going to be lots of things that happen in people’s lives that we couldn’t possibly know about. But on balance, when we look at the things that we can control and the things that we can deliver, this is an incredibly powerful addition to what you already know.

To give a more human experience, a more empathetic experience to your employees, your customers, your business partners, basically. other humans.

Al Elliott: Okay. So we’ve collected all this data, but what now, how do we know what’s important and what actually needs action? Let’s get back to the interview. So if you are able to choose these segments, how does someone actually decide which segment to spend their time on?

Whether it be improving engagement, um, sales, marketing, that

kind

of thing.

Peter Dorrington: Let me talk a moment about Empigraphic Segmentation and how that fits with what every other business already does in what I call Strategic Segmentation. So Empigraphic Segmentation are cohorts of people that have broadly similar sets of feelings.

Now, what I mean by that is at the core of what we do, we use eight universal emotions we know everybody on the planet feels. Um, what’s different about empigraphic segments? One of the big differences is that people move in and out of them because their emotional state keeps changing. So if you’re a high rolling customer or you’re a long tenure employee, that status doesn’t change day to day.

You know, you tend to have characteristics which are persistent. But obviously, with an empigraphic segment, a week ago, I felt really terrible. But today I’m on top of the world. I’m in a completely different group now, a completely different cohort. So I’m in the happy clappy segment of people. So that volatility, that change, some of which we want to influence, we want to put people into positions where they do feel good, that we’re giving them the experiences that they want.

But we have to temper that with ROI. So, Empigraphic Segments are about feelings. We add it to what you already know, like RFM, if I’m looking at that. For employees, we add it to, um, what we know about employees, what they like, what they value, but, you know, what their shift pattern was. Have they worked five straight antisocial hours shift in a row when they’re supposed to be working office hours?

Because we know they’re not going to like that. And the answer, intuitively, is a lot depends upon how they feel about you. If they love you, they’ll probably forgive you, but if they already hate you this might be the final act that destroys your relationship.

Al Elliott: Let’s talk about the actual employees and employee engagement and that.

So from what I’ve understood from what you’ve said, is that if we run this from an employee engagement or potential risk of attrition, sort of a sort of angle, then you’re able to take, if I’ve got a thousand employees, you’re able to say to us, Say to me, there are six people who you need to worry about this week because something’s going to happen.

Have I understood that?

Peter Dorrington: Yes, exactly. Now, as I say, it’s a probability. So, um, because yeah, some people are extremely resilient. One of the things about infographic segments, this, this was a weird wrinkle that came out in the research. I was asked. Do different kinds of people respond differently to their emotions?

And I think somebody was thinking about, you know, the wealthy versus those that are in poverty or people who live in the East versus those of us in the West. Um, The answer was a little bit more complex in that everybody tends, if you’re angry, you tend to act like an angry person, irrespective of other circumstances.

But if I’m somebody with a sense of entitlement, like I know that I’m a really valued employee, I might get angry easily. easier or more quickly than somebody who’s hanging on to their job by their fingernails and absolutely need the job. Now, I’m not saying they don’t feel the anger, but they either may display it differently or they may take, they may be slower to anger.

Now, The prediction part. We look at all 1, 000 employees and we put them in what I call a disposition score. That’s just one number, but it represents the sum total of all of the emotions and the disposition would go from minus 100, you know, I’m surprised they’re still here, to plus 100. These people are chomping at the bit.

They’ll come in on their days off. They’ll bring their families in to help. So everywhere between there. Now what we can do is saying for every employee, we’re recalculating that score. Okay. based upon what’s happened to them, how that’s affecting the way that they feel. And we know that there are certain empigraphic segments which are very close to the point where they will do one of those actions.

Now let’s take one that’s big on people’s minds at the moment. Attrition. People leaving. Unexpected resignations. There may be many reasons why people leave. Some people have a really bad shouting match with their manager and they just walk out. That’s a bit like a lightning strike. You know, that’s very difficult to predict, but what we can do is say this person’s had a A relationship over the last couple of weeks, which has just not been working, or they were passed over for a promotion, or they didn’t get the bonus that they expected.

There are loads of things. So that anger might be one reason, simplistically, why people leave. Other people might leave because they’re really anxious about their job and say, I need to jump before something bad happens. So the outcome is the same. The causes. Towards those outcomes are different, but they’re rarely one thing.

And this is a say where so many people get it wrong. What they look at was what was the last thing to happen before that person resigned or they walked out and they said, well, that caused it. And it’s rarely that was the case. That was just the last step in a long journey of disappointment or despair.

You know, that was the final straw moment. So drawing too much significant. on the effect of that can make organizations do really weird things. You know, they think, okay, clearly the problem is here that, you know, we changed the working hours and half a dozen people resigned as a result of that. It must’ve been that that caused it.

No, that was probably the final straw for five of them. That’d be for one person can’t do that. So what we do with leaders and managers. Is we give you a heads up that we say, this is the emotional makeup of your team right now, whether it’s five people or 500 people, some of them are in a good place, some of them we think are in a bad place.

More importantly, we can tell you what the emotional composition of that is. Some of these people are angry, some of them are afraid, some of them may have been disappointed or they’re disgusted by you’ve announced the new corporate strategy. That was a good one, actually, because I looked at, um, a merchant bank.

And they announced their strategy in the city. And they had an employee engagement survey shortly after that. And everybody said what you would normally expect. Yeah, rah, rah, I’m totally with it. My manager is wonderful, and so on. But when you listen to the subtext, the conversations that were going on, that wasn’t how they felt, felt in reality.

You know, some of them say, I’m really worried about this. I don’t think this is the right direction. By the way, my boss is awful. And I think this is the bit where Without snooping, we’re not reading their emails, we’re not analysing anything they wouldn’t expect to be analysed, but we can start to identify some of those moments, say actually this team’s performance is being negatively impacted by the influence of their leadership.

You know, they’ve got some people in this team, some of the changes we’ve made have affected them so adversely and they were already in a difficult place that this has put them to the point where they’re definitely in the danger zone. So what we would normally do is say of your thousand employees. This is the probability of these people leaving within 30 days, the probability within 60 and 90 and so on.

And obviously, prior to that, we say there’s a probability they’ll probably take an unplanned day off, or they’ll go home early, or their productivity will start soon. You see the precursor behavior typically before you see the big moment. the, you know, the business outcome that we’re interested in.

Al Elliott: You said, um, very early on, you said it was counterintuitive and you gave me a result.

I think it was the bank that was counterintuitive. Is there anything else that surprised you whilst you’re rolling this out? Any other surprises you came across?

Peter Dorrington: Well, the first one, as I said, was about happiness and also organisations say we’re going to monitor employee happiness. And actually, there isn’t a direct correlation between some of the things of everything, as I said, like Happy customers don’t, or happy employees don’t always stay.

Happy customers don’t always buy. Um, the, I think the big one was actually how poor surveys are at predicting what is going to happen next. A survey tends to be a snapshot of a moment in time, and it does have all of the bias in there. And I am full of admiration for people who can write good surveys. Um, but I think that, um, as I say, they’re a snapshot.

It’s how I felt three months ago. May be completely different today. One final thing I think I would add to that as well is don’t stop doing those kinds of things. Don’t stop looking at employee satisfaction, but do really genuinely listen to what people are saying and look for the common things. I mean, you can deal with people one to one.

That’s empathy. We normally do that anyway. But if you can find those common drivers, you can make systemic change in your organization. So rather than deal with the consequences. you can actually deal with the causes. And I think that for me, as a data scientist, was the most important thing. We didn’t have to wait until we could see the effect that followed the cause.

We could identify the cause and predict the effect.

Al Elliott: Is it the Heisenberg principle where if you measure something, it can alter the result? Is that the right one?

Peter Dorrington: So if you tell people that you’re going to come in for an interview, we want to do all this kind of work, you will get a different result than if you do it on their natural behaviour.

So this is why I say we listen to the conversations they’re naturally having, not as a snooping, but actually just being part of that. I was talking to somebody yesterday, she used to go outside and smoke. And when I was there, I’d hear people talking about what was going on inside the company. So yes, if you are over focus groups, for example, the worst.

People know they’re being watched. So they have limited utility. And the problem is bias, which we try to get out of the system as much as we possibly can. What is the smallest group that this is effectively working for? Well, I think if you’ve got less than 500 employees, it’s probably not economic to put the investment in to do the initial study.

Now I’m not saying don’t do it. Much of what I’ve talked about, you can do manually. But there is an overhead in doing that. But larger organizations, so if we go from the upper end of the small to medium enterprise and beyond, or where you have a complex environment, that’s really, I think, where most people get the benefit.

But some of the things I’ve talked about, you can do that in a company of five people. Um, and, you know, I’m not saying that’s probably the best use of your time, but if you’re interested in what’s driving your employee behavior. That’s one way to go.

Leanne Elliott: As a psychologist, my world is governed by ethics and ethical practice, ethical considerations.

When you add on that, I regularly help clients with running surveys, collecting data. I have to uphold the highest of standards when it comes to data privacy and confidentiality. Data of any kind can be misused. So hearing about Answerlytics and the millions of data points being monitored, I had a lot of questions about the ethical use of the platform.

What are the ethical considerations of having so much data at your fingertips and how are people and their thoughts and feelings being protected? As a psychologist, my red flag is up. I’ll ask Peter about this and how they mitigate the risk of organizations misusing data.

Al Elliott: You just said there, we don’t read their emails or anything, but I’m curious, how the hell are you getting this data in the first place?

And the second part of the question is, if you do implement this, do you tell all your employees that this is running behind the scenes, or is it something you should keep quiet?

Peter Dorrington: So firstly, I’ll take the second part first. I’m all for transparency. Be absolutely open with what you’re doing. Um, but also the reason you’re doing it.

So the organizations we work genuinely want to deliver a better employee experience. They want to understand why do people leave when we’ve, you know, we’ve put beanbags in the break rooms. We’ve given them great pay and benefits. They’ve got the long term incentive options and so on, yet they still leave.

Clearly there’s something else going on. So be very transparent about what you’re doing. Now the data bit. Firstly, we do not accept any personally identifiable information. So all we want to know is that employee abc123 is the employee that we saw yesterday and what’s happened to them since. We don’t need to know that that’s Peter, we don’t need to know specifics about individual things and we’re definitely not opening his email and reading them to see what he’s saying in closed quarters.

What we’re doing though to get the initial data about what’s important is reading things which we could say they’re in the public forum. So people would expect them to be read and analyzed. So, you know, if it’s a call center, we look at the transcripts and what the agent says about what that call was really bad.

Or, you know, we look at a manager’s notes saying, I just did the weekly review and Peter’s review this week wasn’t great. Um, because we know that Peter would probably feel something about that. So employees would expect us to analyze those, especially when we tell them We do it so that we can figure out what you care about, so that we can actually help you by delivering more of the things that you care about and try and lessen the impact of the things you don’t.

So be transparent, do the initial research with transcripts because you, when people talk about things, you remember I said that question, how would you describe this? That comes from a completely different place about how would you rate this? So if I asked you on a 1 to 10 scale, how would you rate your experience today?

What happens is that goes into your conscious. Logical part of your brain. You think about it. You think about why do you ask the question? What does a 1 to 10 scale look like? What do I actually feel? Is there any adjustment? But when you ask somebody to talk or tell a little micro story, we only need a couple of sentences from somebody to say what they naturally do is talk about what was important to them and then they use emotions to tell the other person why it was important to them, all of which is detectable.

So we don’t do anything secretive. We don’t do things that are, you know, um, people would get, uh, embarrassed about. It’s a bit like the employment engagement survey, we say working a late shift makes people feel some negative emotions. The way we figure out Peter worked in late shift is we look at the time recording system and say employee ABC 123 worked a late shift last night, probably made him feel a little bit more negative than he did before.

Now, will some organizations. get this wrong, I have no doubt about it, mostly because as human beings, we’re very fallible. And I’ll use myself as an example. As a manager, I learned all the things about management you would expect, but I’m not great at reading people. I have to have it pretty much in my face to do that.

So, um, Yeah, if I was to play cards, I don’t play poker because I can’t read people. I can’t tell if they’re bluffing or not. But I do like blackjack because there are very firm rules. Stand and learn from that. So, um, that’s normally that the hand in glove kind of thing. If we give you this information as an organization, Essentially, you’re being given great insights into how people are feeling, but you have to figure out how you’re going to respond to those in the same way that you have to figure out whether you want to save that relationship with that particular employee.

Maybe they’re, they’ve just been really difficult for five years. And we may decide trying to save them is actually harder than saying perhaps they’re not in the right place. What can we do to help them transition perhaps into a new role or a new organization, which would be a better fit? So we talk about well being a lot now.

So if we can give people insights into who’s feeling great, who’s feeling not quite so great, what can we do? Maybe we need to take Peter off the complaints line for a while and give him something which is a little less stressful. Then that’s an enormous insight. Now. One final thing on this AI and automation is having an enormous impact on the way that people feel and the way that they behave.

It does all of the simple, repetitive, low value task centric things that we as human beings used to spend a lot of time doing. So we think, hooray, that’s going to save us a lot of money because we don’t have people spending time changing passwords, for example. And I totally get the economic case for that.

But there’s been an unintended consequence to that. The work that’s left over often is much more stressful. tiring. It requires creativity because it’s something you didn’t expect before, otherwise you would have automated it. And if that’s your job, firefighting eight hours a day for five days a week, you’re going to burn out because nobody can do that for very long.

We have actually taken away their decompression time, the time when they would do admin, the time when they would actually reflect a little bit and come down off the ledge because they just go from one firefighting episode to another firefighting episode. And that’s really driving powerful conversations about how do we support staff in that environment where they’re giving much more value.

Than they ever could before because they can really focus on where they make a strategic influence, but it’s killing them. So we need to make sure we moderate that and use the tools in our toolbox to give people a more rounded experience. So they don’t go home at the end of the day, completely exhausted.

They don’t spend the whole of Sunday worrying about going into the office on Monday morning. I used to do that.

Al Elliott: You used to worry about going into the office on Monday. So clearly there’s something happened because you’ve now spoken for almost continually an hour on something with no loss, no loss of, of enthusiasm on it.

What’s, what was the inflection point for you in your career to go? This is what I want to work on.

Peter Dorrington: My disaffection actually came from firstly, how little progress we were making. We were tarting up data science. Let me give you one example, machine learning. It’s been around for ages, decades, you know, and I can remember working on it when we called it neural networks and variety of other guises from that.

Um, and also very few people understood how it worked and therefore were caught out by when the model didn’t predict something that was going to happen. Um, But when I had the a ha moment, which is for somebody, as I say, sometimes I feel like an outsider looking in on the human race. I said, actually, I can deal with people at a level now, which is important to them.

And we can have a balanced relationship because after all, contract of employment is in law between two equal parties. Um, often we feel like I have all the power or you have all the power. It was that ability to affect a real change, come up with something really new. At the end of the day, I’ve always liked looking forward and saying, what can I do that would change something?

So now what gets me so excited is I’m seeing this making real change to people’s lives, which is great. This is not abstract anymore. This is very practical, hands on. And secondly, we’ve only just started to scratch the surface of how This is going to impact the way. I like you think that AI is going to be the transformative technology of the decade.

It’s enormous. There’s no putting that genie back into the bottle, even if you wanted to. But we’re still at the point of understanding where it will be able to make a change. And the bit that AI struggles with is the uniqueness of human beings. You know, it can deal with us when we all look, work like homogenous groups.

But I want to get to the why can’t I have a segment of one person? Why can’t every employee belong to their own empigraphic segment? And we’re there now. So that’s why I’m so enthusiastic. I’ve actually, I did this, not because I wanted to invent something I could put my name to, but because I could see a problem and I wanted to find a solution.

And I tripped over the solution. I say, after three years of failing miserably, and now I’m excited about seeing it being deployed and making real change in people’s lives. What

Al Elliott: I think some people might be worried about is how does this not turn into Evil. I mean, you, you’re collecting so much data. How do you stop people from turning from going bad?

Okay.

Peter Dorrington: The dark psychology concern has come across and I’ve, I’ve actually looked at them and been involved in some of the discussions about the ethics of emotion AI. So that being able to identify emotions, let’s be honest. Um, if you have that level of insight in any technique, there is an ability to misuse it.

So from marketing point of view, I could use it to identify people who are vulnerable, for example, and then come up with a campaign to address vulnerable people. So any technology has the potential to To be misused or abused. But as I said, we’re very transparent about what we’re trying to do and why we’re doing it.

And the organizations we work with are also that transparent. We genuinely want to deliver better, humane, empathetic experiences for our customers, our employees and our business partners. At the end of the day, It makes good business sense, and I think a lot of it comes down to trust. I’m not going to abuse our relationship because trust arrives on foot and leaves in a bus with all your mates.

And I think that, you know, it’s hard won and easily lost. And if you do something with this technology, which is an abuse, which is unethical, we do our best to actually make that impossible, as I say, because we don’t take PII. We don’t look at anything which is confidential. We don’t want anything that is embarrassing for people.

We are genuinely trying to give them a better experiences. We don’t work with any organizations that I don’t feel strongly meet that ethical framework. And so, because what we do is unique, I don’t think at the moment anybody else could do it. So, short answer, just because you can, doesn’t mean you should.

And think about that from an ethical point of view. Because at the end of the day, you may ruin your own business. Employee base your own customer base by doing the very things which is a short term fix long term disaster

Al Elliott: Well Cambridge Analytica should prove that beyond any reasonable doubt that they found a loophole They used it and they basically screwed themselves in there.

So

Peter Dorrington: Yeah, what was interesting is what they did, you might argue, wasn’t actually immoral. What was wrong was they didn’t tell people that they were doing it. And especially they didn’t tell people they were doing it with second and third party data. So the data collection they did, I think, yeah, absolutely no question, that was not What should have happened?

What they were trying to do is to look at voting preferences. Well, we do that with marketing preferences. Who wants to buy our product? Where should we put a store? So I think there are some discussions that you can have around that. I keep it pretty straightforward. We give you actionable insights about the way that people feel and how that might influence what they do.

It’s up to you to figure out what to do with those insights, just like almost every analytical team that’s out there at the moment working match. The same way.

Leanne Elliott: Such an interesting conversation. Hopefully you enjoyed the slightly different format we did there to, to bring you into the, into the interview room.

Um, Al tell me, you spoke to Peter, I think for almost two hours overall. I spoke to him for, I think for about an hour before that. He’s such an interesting guy. What were your main takeaways? What would you want people to really understand about this, this platform and the work that Peter does?

Al Elliott: Well, first of all, I went into it, um, and I apologize, Peter, because I did go into it thinking, Hmm, is this going to be a bit dodgy?

Are they, are they listening to people’s conversations? Are they reading people’s emails? But Peter explained, No, I would encourage you to get on the phone with Peter if you are interested in this kind of thing, because he has such a great way of putting things. He’s so eloquent yet. doesn’t, even though he’s a genius, he doesn’t sort of baffle you with, um, uh, with facts and figures and stuff, which I wouldn’t have understood.

So I honestly could have spoke to him for about two or three hours. I think the main thing is I really liked how honest he was about things like, yeah, this could be misused, which is why we don’t let anyone use it unless we vet them very carefully. Um, yes, there is a potential for evil, which he covered there.

Um, and even just going into his personal life a little bit, you know, he didn’t use to like work and then he found this. Um, and obviously you can tell, I think that he absolutely loved it.

Leanne Elliott: What I love about Peter as well is, is how he really is bringing the science to human emotions and actually looking at how we can support people in the workplace by proactively identifying their needs ahead of time and meeting them.

It is such an interesting, interesting product. And yeah, as you said, I think Peter would welcome connections from anyone who has enjoyed the episode. As always, we will leave All the links for Peter in the show notes.

Al Elliott: So join us on Tuesday for our This Week in Work episode, where we do a roundup of the latest in the workplace world and also bring you the weekly workplace surgery, where I put your questions to Leanne.

Leanne Elliott: And just before we go, if you do want to find out more about Peter or Anthrolytics, here’s what to do next.

Peter Dorrington: The thing to do is check out the website, which is anthrolytics. io. Um, and we’re putting resources up there. My co founders say I publish too much, but. Actually, I think, well, this is a new area. So let’s, let’s do a little bit of what it is in awareness raising or look for me on LinkedIn.

There aren’t that many Peter Dorringtons out there. So Peter Dorrington on LinkedIn. Um, and I always answer emails. And if you send me a query via LinkedIn, I will respond. Um, I’m more than happy to share some of these insights and talk things through with people as I say. So go to the website anthrolytics.

io or look for me on LinkedIn and just ping me an IM.

Leanne Elliott: We will see you next week. Please do follow. Subscribe is genuinely the best way to support the show and enable us to bring you more content every week. We will see you on Tuesday.

Al Elliott: See you Tuesday. Don’t forget to subscribe. Don’t forget to subscribe, leave a review, subscribe, leave a review.

Bye.

Leanne Elliott: Bye.

And just before we go, if you do want to find out more about Peter or Anthrolytics, then here is what? Anthrolytics?

Al Elliott: Anthrolytics.

Leanne Elliott: I’m co founder of Anthrolytics. No. Yeah, that was right. Anthrolytics.

Al Elliott: Like analytics, but anthro.

Leanne Elliott: I keep saying empathetic wrong, don’t I? Empathic, is it? Empathetic. I would say empathetic,

Al Elliott: I think.

Leanne Elliott: Just make a very brief comment. Is that allowed?

Al Elliott: Of course. Did again, from empathy also positive. Oh, no, you have. I’ll shut up. Carry on.