Attention Is Becoming a Competitive Advantage

TL;DR

• You did not choose to spend ninety minutes on your phone. You chose the next twenty seconds, a few hundred times over.

• Two research reviews covering about 195,000 people found the same thing. More short video, weaker attention and weaker self-control.

• But that is a link, not a cause. Two of the most quoted numbers in this whole field, brain drain and the 23-minute recovery, do not survive checking. Anyone who sounds certain here is selling you something.

• The real cost to a business is not distraction. It is giving up. You recover from distraction. You do not recover from a habit of quitting.

• AI makes removing effort cheap. Some effort is waste. Some effort is the work. Remove the second kind, and your company gets thinner, not faster.

• Attention is now the thing in short supply, and you cannot buy more of it. Check who controls it, the same way you check who controls your data.

---

THE PROBLEM

What I think is really going on

You did not choose to spend ninety minutes on your phone last night. You chose the next twenty seconds. Then you made that choice a few hundred more times, and ninety minutes were gone.

That is the part people miss. We count screen time because it is easy to count. It is the wrong number. What matters is where your attention goes, and who decides. With short video, the deciding has been designed out. A ninety-minute film asks you to commit. Short video never asks. It has no stopping point, so you never get a moment to choose again.

I am not writing this as a wellbeing point. I run an AI and automation business. My whole job is removing effort from how companies work. So this lands on my desk as a business question. Which effort do I remove for a client, and which effort are they paying me to protect?

Takeaway: screen time is the wrong measure. The real question is who removed your choice.

---

THE EVIDENCE

What the research says

Here is what I could check myself, because I read the studies instead of the headlines.

Nguyen and colleagues reviewed the field in Psychological Bulletin. Seventy-one studies looked at seventy combined, 98,299 people. More short-video use went with weaker thinking skills at r = -0.34, a moderate link. Self-control, the ability to stop yourself, was the weakest at r = -0.41. A second team, Tang and colleagues, combined 58 studies and 96,676 people in the Journal of Medical Internet Research and found the same direction on stress, anxiety, and low mood. Two teams, different sources, about 195,000 people, pointing the same way.

That tells you something about people. It tells you nothing about companies, so I went and found the workplace evidence.

Gloria Mark at UC Irvine has measured how long people stay on one screen at work for twenty years, stopwatches first, software logging later. In 2004, the average was two and a half minutes. By 2012, 75 seconds. In her most recent measurements, 47 seconds. Others repeated it and got 44 seconds, and a median of 40.

Sophie Leroy showed why that matters in Organizational Behavior and Human Decision Processes in 2009. When you switch tasks, part of your attention stays behind on the last one. She called it attention residue. It is worst when the last task was left unfinished, and it drags down your accuracy and speed on the new task for as long as you are on it. It does not wear off after a minute.

So the real cost of an interruption is not the interruption. It is the quality of everything that comes after it.

Takeaway: the phone evidence is about people. The workplace evidence is about your P&L.

---

THE OTHER SIDE OF THE EVIDENCE

Now the part nobody quotes

The caveats are serious. Both short-video reviews rely on snapshots. Researchers asked people at one moment how much they watch and how they feel. That shows a link, not a cause. People who are struggling may simply watch more. Most likely, it runs both ways.

The most quoted study in the field has a bigger problem. Ward, Duke, Gneezy, and Bos published Brain Drain in the Journal of the Association for Consumer Research in 2017, reporting that simply having your phone in sight uses up thinking capacity. In 2022, Ruiz Pardo and Minda repeated the second experiment, announcing their method in advance so they could not move the goalposts, and published it in Acta Psychologica. It did not work.

The workplace numbers get abused too, and this one I nearly got wrong myself. You have seen the claim that refocusing after an interruption takes 23 minutes and 15 seconds, credited to Mark, Gudith, and Klocke in 2008. That paper does not contain the number. It found the opposite: interrupted people finished slightly faster, 20.3 minutes against 22.8, at no cost to quality but with much higher stress and effort. The 23-minute figure comes from an interview Mark gave in 2006. The documented figure is about 25 minutes, from her 2005 study with Gonzalez and Harris, which also found people spent around 11 minutes on a piece of work before switching, and that 57 percent of work got interrupted.

The French numbers are stretched the same way. In a study of 5,633 people aged 15 to 29, run in April 2025 by Institut Montaigne, Institut Terram and Mutualite Francaise, 69 percent of the heaviest users reported trouble concentrating against 45 percent of the lightest. Those are answers on a depression questionnaire, not diagnosed attention disorders.

I am taking apart half of my own case on purpose. What survives is enough. Attention at work has measurably shrunk. Switching costs you more than the switch. And nobody has proved what caused any of it.

Takeaway: check the famous numbers before you quote them. Most of them are wrong.

---

WHAT IT COSTS

The real cost is giving up, not getting distracted

You recover from distraction. Something interrupts you; you come back, you carry on. Irritating, not serious.

Giving up is different, and it is the one I would worry about if I ran a large workforce. Learning maths. Building a company. Closing a hard deal. Fixing a broken system at three in the morning. All of it means staying with a problem after the interesting part is over and only the hard part is left.

Feeds train the opposite habit. Bored, so swipe. Difficult, so switch to something easier. Do that a few thousand times a year and you are not training distraction. You are training yourself to quit. And because how we feel about ourselves rests partly on what we finish, less finishing means less confidence.

That is not a wellbeing problem. It is a capability problem, and it will show up in your delivery numbers long before it shows up in a staff survey.

Takeaway: distraction costs you an hour. A habit of quitting costs you a capability.

---

WHAT IT COSTS YOU

Where this shows up at work

Teachers compete for attention against recommendation systems built by the best-funded AI teams on the planet. Everyone can see that fight. Companies are in the same one, and most have not noticed. Email, Slack, Teams, alerts, dashboards, and now AI assistants, each giving you a new reason to switch tasks.

After twenty-five years running IT operations and automation at HP, IBM, and DXC across more than twenty countries, my summary is blunt. Companies mistake being busy for producing something, and producing something for creating value.

I keep three things separate. Output is what got made. Outcome is what changed because of it. Impact is what that change is worth. Messages sent is output. Speed of reply is output. No profit line has ever moved because somebody answered faster.

Takeaway: stop rewarding speed of reply. Start measuring what actually changed.

---

WHAT TO DO

Which effort to remove and which to protect

AI raises the stakes, because it makes removing effort almost free. Once something is free, people stop asking whether it should be done.

Some effort is pure waste. Searching for files, formatting documents, chasing numbers, writing status reports. Remove all of it. That is my day job, and I am good at it.

Some effort is the work itself. Sitting with a problem you cannot yet solve. Testing an assumption you would rather leave alone. Making a judgement call you will defend in front of a board. Remove that, and you have not made the company faster. You have made it thinner, and thinner shows up nowhere until something breaks.

This is where I watch AI projects lose their value. I track it across four stages: promised, piloted, in production, made permanent. Most die between production and permanent, and the reason is seldom the technology. It is that nobody left in the chain had enough quiet thinking time to make the hard call about what had to change.

Takeaway: automate admin without hesitation. Protect thinking time like a contract.

---

THE CONCLUSION

Who controls your attention

I check who owns a company’s AI across five things: the model, the data, the people, the cost, and the way out. Miss one and you do not own the capability, whatever the vendor slide says.

Attention deserves the same check, and most companies would fail it. Who decides when your best engineer gets four hours without interruption? Not the engineer. Not you. In practice, it is a notification schedule set by a product manager at a company you do not employ, cannot influence, and have never met.

Three things you can do, none needing a budget. Swap one speed-of-reply measure for a real outcome measure on one team for three months, then listen to the conversations that follow. Put quiet thinking time in the calendar as a real commitment, because willpower does not survive a company built for interruption. And before you automate, ask whether the effort you are removing is admin or thinking. Admin goes. Thinking stays.

Information stopped being scarce long ago. Intelligence is getting cheap and plentiful. Attention is the scarce one now, and unlike the other two you cannot buy more of it. The companies that remove wasted effort while protecting useful effort will pull ahead, and nobody will see the gap on a dashboard until it is already large.

The question is no longer who owns our data. It is who owns our attention.

---

Takeaway: attention is the one input you cannot buy. Audit who owns it.

Where this came from, and where I disagree

This piece started with Pierre Valentin’s Malaise dans la generation Z (Gallimard, 2026). Several points in the opening are his, not mine, and I want that on the record. The contrast between committing to a long video and the almost unthinking swipe on a short video. The test of trying to remember the last three short videos you watched. The point, credited by him to Baptiste Detombe, that once one platform invents something addictive, the others copy it. The observation that a teacher facing thirty pupils now faces a majority who struggle to concentrate. And the chain from losing perseverance to losing confidence. His own wording, on page 49, is that the collapse of attention and perseverance makes the collapse of wellbeing easier. I disagree with him on three points, set out above. That the effect is proven. That the brain drain finding is solid. That the French figures describe attention disorders. Read the book anyway.