Latest / Future of Work Tech with Fexingo: Remote Tools, AI Productivity, and Workplace Software / How Workplace Heat Maps Predict Employee Churn
Transcript
- Lucas: So there's this fascinating dataset coming out of a mid-size SaaS company in Austin. They installed desk occupancy sensors and badge readers across their three floors, and they found something that sounds almost too tidy to be true: employees who were about to quit started visiting the breakroom 40 percent more often in the six weeks before they gave notice. Luna: Wait — they were hanging out by the coffee machine more? That's the signal? Lucas: That's part of it. The theory is that disengaging employees start seeking social comfort — they gravitate toward common areas, linger by the snack wall, chat with whoever's around. Meanwhile, their actual desk occupancy drops. The heat map shows a kind of 'cooling' around their primary work zone. Luna: So the company essentially watched people's physical drift and used it to predict departure. That's either brilliant or terrifying — or both. Lucas: Right, and that's exactly the tension we're going to unpack today. Workplace heat maps aren't new — real estate teams have used them for years to figure out which desks to cut. But now HR teams are starting to look at the same data as a behavioral early-warning system. Luna: I remember last year, a Gartner report said that 60 percent of large enterprises already use some form of people analytics — but heat maps specifically feel more… intimate. Lucas: Yeah, it's the spatial dimension. A badge swipe tells you someone entered the building. A heat map tells you where they went, how long they stayed, who they sat near, and when they left. Multiply that across weeks and you get a behavioral pattern that's surprisingly predictive. Luna: What's the actual accuracy? I've seen claims that these models can predict churn with 80 percent precision three months out. Lucas: That number comes from a study by researchers at Carnegie Mellon and Microsoft, published in 2024. They looked at badge-swipe and Wi-Fi connection data from a large tech firm and found that a drop in 'network diversity' — meaning the number of different colleagues you interact with — was a stronger predictor than even manager feedback or engagement survey scores. Luna: So it's not just that you go to the breakroom more — it's that you stop talking to people outside your immediate team. Lucas: Exactly. The heat map reveals shrinking social circles. Someone who used to roam three floors and attend five meetings a day suddenly stays on one floor, goes to two meetings, and spends lunch at their desk. That pattern, combined with declining badge swipes on Fridays or Mondays, is what the models flag. Luna: And what does the company do with that flag? I imagine the response isn't always graceful. Lucas: The best practice I've seen is from a company called Envoy — they make workplace visitor management software, but they also use their own occupancy data internally. When the heat map shows a notable shift, a manager has a low-key check-in: 'Hey, I noticed you've been working from a different area lately. Everything okay?' No mention of the data. Luna: That feels human. But I also know there are companies that don't tell employees the data is being used for retention — they just say it's for space planning. Lucas: That's the privacy fault line. In the European Union, the General Data Protection Regulation would likely require explicit consent for this kind of behavioral analysis. In the United States, it's more of a patchwork. And even if it's legal, there's an ethical question: do you want to work somewhere that monitors your breakroom visits? Luna: I think most people would say no, until you frame it as 'we caught three high performers who were quietly disengaging and we saved them before they quit.' Then suddenly the trade-off looks different. Lucas: That's exactly the case a company called BetterCloud made public last year. They used badge data and Slack activity — not heat maps, but similar logic — to identify employees who were 'quiet quitting' before they actually left. They intervened with retention bonuses and role changes, and they claimed a 12 percent reduction in voluntary turnover within six months. Luna: 12 percent is significant. But I wonder: did the interventions work because the employees felt seen, or because they got more money? And does the heat map know the difference? Lucas: That's the limitation. The heat map can tell you someone is drifting away, but it can't tell you why. Maybe they're disengaging because of a bad manager — no retention bonus fixes that. Maybe they have a chronic health issue and need a different work arrangement — the heat map just sees a pattern change. Luna: So the data is a signal, not a diagnosis. The danger is when companies treat it as the latter, and start making decisions — like demotions or PIPs — based purely on occupancy drops. Lucas: Right. And there's already a vendor ecosystem pushing that. Companies like VergeSense and Density install sensors that can detect presence down to the inch and they market their churn prediction algorithms directly to HR. The pitch is usually 'stop regrettable attrition before it happens.' Luna: 'Regrettable attrition' — I hate that term. It implies the company knows which departures are acceptable and which aren't. Lucas: It's a loaded phrase. But it does reflect the reality that some turnover is beneficial — low performers leaving is often a net positive. The hard part is distinguishing between a disengaged high performer and someone who's just having a bad month. Luna: And a heat map can't tell you that. It can only tell you where someone walks. Lucas: Which brings us to the other big issue: false positives. If the system flags an employee as a flight risk based on two weeks of unusual movement patterns, and the manager confronts them about it, you've created a self-fulfilling prophecy. The employee feels surveilled and distrusted, and that actually pushes them toward the door. Luna: There's research on that too, right? A 2023 paper from Harvard Business School found that when employees learned their movements were being tracked, trust in management dropped by 25 percent on average. Lucas: Yes, and interestingly, that effect was strongest among high performers — the very people the system is trying to retain. They felt insulted, like the company assumed they were slacking. Luna: So the cure might be worse than the disease, unless you handle it with extreme care. And I'd argue most companies aren't that careful. Lucas: Agreed. But I don't want to throw the whole concept out. When done transparently — with opt-in consent, anonymized aggregates, and a clear policy that individual-level data is never used punitively — heat maps can actually improve the workplace. They can reveal that the second floor is too loud, or that the collaboration zones are underused, and that helps everyone. Luna: That's the original purpose of heat maps: space optimization. Not people optimization. I think the ethical line is crossed when you shift from measuring furniture usage to measuring human behavior. Lucas: Exactly. The furniture doesn't care if you watch it. People do. Luna: Speaking of watching — if today's conversation gave you something useful to think about, that's the whole goal of this show. These episodes stay ad-free because of listeners who value the content. If that's you, you can support the show at buy me a coffee dot com slash fexingo. It's a simple way to keep this independent. Lucas: Absolutely. It genuinely helps us keep the focus on substance, not sponsors. And on that note — let's talk about one more piece of the puzzle: what happens when the heat map data starts feeding into performance reviews? That's already happening at some companies. Luna: I've heard of a few. One fintech in New York apparently includes a 'collaboration score' derived from seat proximity and meeting room usage in quarterly reviews. Employees can see their own score but not how it's calculated. Lucas: That's the nightmare scenario. A black-box algorithm assigning a collaboration score based on where you sit. If you're on a team that works mostly asynchronously and doesn't need to meet face to face, your score looks low, even if you're extremely productive. Luna: And then you get penalized for a work style the company itself enabled by having an open floor plan. It's a systems thinking failure. Lucas: Right. The sensor data is a mirror — it reflects the environment you've built. If you use it to judge individuals, you're blaming people for the system's design flaws. The better use is to adjust the system. Luna: So what's the best practice for a company that wants to use heat maps ethically? Give us a rule of thumb. Lucas: Three things. One: disclose everything. Tell employees exactly what sensors are collecting, how often, and for what purpose. Two: aggregate the data. Only look at patterns across groups of ten or more, never individuals, unless the individual has explicitly opted in for coaching. Three: never tie the data to compensation or promotion. Use it for space design, scheduling, and maybe — maybe — as a conversation starter with a manager trained to use it gently. Luna: And let employees see their own data. I think that changes the dynamic completely. Lucas: Yes. When you give people access to their own movement patterns, they often self-correct. They realize they've been isolated and start seeking out colleagues. That's empowerment, not surveillance. Luna: So the heat map can be a tool for connection or a tool for control. The difference is transparency and agency. Lucas: Exactly. And right now, most companies are somewhere in the middle. The ones that get it right will see retention improve — and not just because they caught people before they quit, but because they built a workplace worth staying in. Luna: I think that's the real metric that matters. Thanks for diving into this one. Lucas: Thanks for the great questions. That's all for today. We'll be back next time with another angle on the future of work tech.