Peopleware: Productive Projects and Teams

Miles: The middle of the nineteen eighties. Two consultants in New York invite software organizations to enter a contest. Each one sends its programmers in pairs, and every programmer gets the same specification for a medium sized program to design, code and test.

Miles: The two halves of a pair never work together. Each sits at their own desk, in their own office, during ordinary hours, with their usual tools, and logs the time it takes.

Miles: When the scores come back, the programming language barely matters. Ten years of experience doesn't beat two. Pay hardly tracks the results at all. What tracks them is the company that sent you. Two partners who never worked together finished, on average, within twenty one percent of each other.

Miles: Today: Peopleware, from Tom DeMarco and Timothy Lister. In two sentences: the hard problems in project work are about people far more than technology, and most offices are laid out in a way that stops people thinking. It made the case for a quiet room with a door you can close on productivity grounds, decades before anybody called that a sensory need, and Nora's going to argue that makes it the most useful thing we've cited yet.

Miles: Sixty Books. Presented by Stoar Consulting and Coaching.

Miles: Book forty five of Sixty Books, and I'm Miles. The arrangement hasn't changed: one management title at a time, sixty picked from the last hundred years, each one read for the leaders and teams whose minds work on another pattern, and for a working week that now has AI sitting in it. Nora has the research, and this time the book ran an experiment of its own.

Miles: Nora, last time you left us on a study where colleagues walking up to the desk pulled people off their work more often than any alert. So the room, and not only the inbox.

Nora: And the room is this book, in nineteen eighty seven. Back then the interruption was a phone on the desk and a colleague with a question, but the claim hasn't moved. Thinking work needs long stretches nobody breaks into, and most workplaces are built to break into them.

Miles: So the case was in print before most offices had email.

Nora: In print, with a table of numbers, and the open floor plan spread anyway. By the third edition, in twenty thirteen, the authors had added a chapter on email. It's titled E, then evil in brackets, then mail.

Miles: You can hear the brackets.

Nora: And one of its subheadings is Repeal Passive Consent, which would sit comfortably in last episode's book.

Miles: Tell me how the contest ran.

Nora: The authors had been running public productivity surveys since nineteen seventy seven, and from nineteen eighty four they turned them into a competition they called the Coding War Games. The figures I'm giving come from the second edition, nineteen ninety nine, and cover nineteen eighty four to eighty six: more than six hundred developers from ninety two companies.

Miles: And what came out at the top?

Nora: The best organization worked eleven point one times faster than the worst, and every one of its competitors passed the acceptance test. The top quarter of the field averaged two point six times the performance of the bottom quarter.

Miles: And what didn't count?

Nora: Language mostly didn't, apart from assembly, where people fell well behind. Experience didn't, once someone had six months in the language they were using. Nearly a third finished with no defects at all, and they paid nothing in speed for it. And pay barely followed performance. The better performing half earned less than ten percent more than the other half and performed nearly twice as well.

Miles: Then the pairs.

Nora: The finding this part of the book leans on. Both members of a pair came from one organization and never worked together, and their scores differed by only twenty one percent on average. The authors read that as clustering. Strong performers gather in some organizations and weak ones in others, so something about the place is doing part of the work.

Miles: Which is where the desks come in.

Nora: Before the exercise, every participant filled in a questionnaire about their own workspace. Top quarter against bottom quarter. Dedicated floor space, seventy eight square feet against forty six. Acceptably quiet, fifty seven percent said yes against twenty nine. Acceptably private, sixty two against nineteen. Can silence your phone, fifty two against ten. Can divert your calls, seventy six against nineteen. And people often interrupt you needlessly: thirty eight percent of the top quarter said yes, and seventy six percent of the bottom.

Miles: Every line of that is about control.

Nora: Keep that, because our segment stands on it. And the authors printed the caution themselves, right after the table. The data doesn't prove that a better workplace makes people perform better. It may only show that better performers drift toward organizations that give them better workplaces.

Miles: The authors.

Nora: Tom DeMarco first. Born in Hazleton, Pennsylvania, in nineteen forty. Electrical engineering at Cornell, a master's from Columbia, a diploma from the University of Paris. He joined Bell Labs in nineteen sixty three and worked on the first large electronic switching system for telephone exchanges, and in the seventies he became one of the leading names in structured analysis, with a book on it in nineteen seventy eight.

Miles: And fiction, I gather.

Nora: Several novels, including The Deadline, a novel about running a project. He took the Warnier Prize in nineteen eighty six and the Stevens Award in nineteen ninety nine, and he lives in Camden, Maine.

Miles: Second.

Nora: Timothy Lister, born in nineteen forty nine, based in Manhattan, who works on design, software risk and the human side of technical work. The two of them wrote Waltzing With Bears, a book on project risk, in two thousand three. Both are principals of the Atlantic Systems Guild, a consulting firm they started in New York in the nineteen eighties with four colleagues.

Miles: And where the book came from.

Nora: The third edition's preface places it on a long flight to Sydney in the eighties, where the two of them talked their way to one idea. The major problems of systems work aren't so much technological as sociological. First edition, Dorset House, nineteen eighty seven, a hundred and eighty eight pages. The second edition, in nineteen ninety nine, kept the original and added a closing section of eight chapters. The third, in twenty thirteen, added several new ones, on leadership, meetings and email among them.

Miles: Walk it front to back.

Nora: Part one is managing people as people. The opening chapter makes the claim from the plane: projects mostly fail for sociological reasons, and managers drift toward the technical problems because those are easier to work on, not because they matter more.

Miles: Then the office.

Nora: Part two, which is the heart of it. It opens on the furniture police, the people who plan office space for tidiness and cost rather than for the work done in it. One subheading is The Uniform Plastic Basement.

Miles: That's a subheading.

Nora: It is. They weren't writing for the facilities department. Then comes the chapter called You Never Get Anything Done around Here between Nine and Five, which is where the war games live. Then the money chapter, which leans on a study a large computer company ran before building a laboratory for its programmers: cutting space below a minimum costs more in lost work than it saves in rent.

Miles: And flow.

Nora: The chapter's called Brain Time versus Body Time, and it borrows the psychologists' word for deep, almost meditative absorption, where hours pass without your noticing. That's three years before the Csikszentmihalyi book we read in episode twenty four. The book's case, from the first edition on, is that you can't switch it on. It takes fifteen minutes or more of concentration to settle in, and every interruption costs a fresh immersion, not just the minutes the interruption took.

Miles: So a two minute call isn't two minutes.

Nora: Not by this book's arithmetic. And then the move I'd keep from the whole thing. They propose counting it. The E-factor is the ratio of uninterrupted hours to the hours your body was present. Not hours at the desk. Hours at the desk that nobody broke into.

Miles: Hard to pad, that number.

Nora: Which is why it's useful. Next comes a chapter on the telephone, with a subheading called A Modified Telephone Ethic, and then one titled Bring Back the Door. After that a chapter built on A Pattern Language, the architect Christopher Alexander's book, which turns into four patterns for a workspace: space tailored from a kit, a window for everyone, indoor and outdoor space, and public space.

Miles: Part three?

Nora: The right people, on hiring and turnover, which I'll leave on the shelf. Part four is teams, and two ideas from it get quoted most. The jelled team, whose output is greater than its members' added together. And teamicide, a list of the ways organizations kill teams without setting out to. Defensive management, bureaucracy, physical separation, fragmenting people's time across projects, the quality reduced product, phony deadlines, and what they call clique control.

Miles: Each of those sounds like a policy somebody signed.

Nora: Most of them were. The third edition adds a chapter that goes back over that ground, with motivational posters and overtime in it. Then part five holds the newer chapters on meetings and email, and between them one called The Ultimate Management Sin Is, which answers itself in the text: wasting people's time. Part six closes on why it's supposed to be fun to work here.

Miles: The evidence. Start with the contest.

Nora: Which I want to describe exactly, before anybody quotes it. The Coding War Games are the authors' own data, from a competition they ran. Organizations chose to enter. The workspace answers are each participant's rating of their own office, on a questionnaire. And the result compares the top and bottom quarters of the field, which makes it a correlation, not an experiment. The fullest account is in the book. There's a five page paper at a software engineering conference in nineteen eighty five, and as far as I can find, it never went to a journal where outsiders could check the data.

Miles: So, a survey with a stopwatch.

Nora: That's fair. And their own caution is the right reading. Better people may pick better offices, and nothing in the design can tell those two stories apart.

Miles: And the famous ten to one?

Nora: The book opens that discussion with someone else's line about a ten to one difference among software organizations. Its own figures are eleven point one between the best and worst organizations, and two point six between the top and bottom quarters. If you've heard ten to one told as a fact about lone programmers, this isn't that study. The finding here is that performance bunched up by company.

Miles: Then what came after.

Nora: Four peer reviewed pieces, from four directions. Kim and de Dear, Journal of Environmental Psychology, twenty thirteen. A survey database run from Berkeley, more than forty two thousand office workers in three hundred and three buildings, mostly American. Enclosed private offices beat open plan on most measures of the indoor environment, sound privacy above all, and what open plan gained in ease of talking to colleagues was smaller than what it lost to noise and lack of privacy.

Miles: So the trade the open plan promised didn't pay off.

Nora: On satisfaction, in that database. Then Bernstein and Turban, twenty eighteen, which episode twelve gave you as a round seventy percent. The detail's worth having. Two headquarters of large multinationals, measured before and after a move to open space, with employees wearing sensor badges that logged face to face conversation. At the first, fifty two people, face to face time fell seventy two percent, while email rose fifty six percent and instant messages sixty seven. At the second, a hundred people, face to face time fell by sixty seven to seventy one percent and email rose by between a fifth and a half.

Miles: They took the walls down and people stopped talking.

Nora: In two companies, measured by badges. The authors' reading is that people withdrew from colleagues they could now see all day, and moved to the screen. Third, Pejtersen and colleagues, twenty eleven. A national Danish survey with two thousand four hundred and three office workers in it. After adjusting for age, smoking, exercise and more, people in open offices of more than six reported sixty two percent more days off sick than people in single offices. Self reported and measured once, so there's no direction in it.

Miles: And the fourth.

Nora: The one on flow. Altmann, Trafton and Hambrick, Journal of Experimental Psychology: General, twenty fourteen. A lab task where people had to hold their place in a sequence of steps. Interruptions averaging two point eight seconds doubled the rate of sequence errors afterwards. At four point four seconds, the rate tripled.

Miles: Seconds.

Nora: Yeah. Seconds, in a lab. It backs the book's mechanism, that what an interruption costs you is your place, not the time it took. The fifteen minutes is the book's own figure, and I couldn't find a study that pins it down.

Miles: So how much of the case is left standing?

Nora: The direction holds up better than the data it started from. In the largest survey, noise and lost privacy are where open plans do worst, and the collaboration case for them has the weakest support of anything here. What's thin is output. Most of the later work measures satisfaction, conversation or sick days, not how much good work got done.

Miles: Now the tier statement.

Nora: Peopleware is practitioner literature, two consultants writing for managers, not peer reviewed research, and its best known data come from a contest the authors ran and scored themselves, with their own warning printed beside it that it proves less than it seems. Its claims about quiet and privacy have since found support in peer reviewed surveys and field studies, which is more than most of this shelf can say. The argument has aged better than its evidence, and none of that is a verdict on either author.

Miles: Our half. And you promised the most useful citation in the series.

Nora: I'll make that case, and it's an argument. Look again at what that table measured. Floor space, yes, but mostly quiet, privacy, a phone you can silence, calls you can divert, not being interrupted for nothing. That's control over what reaches you. The book asked for it in nineteen eighty seven, for every programmer, on productivity grounds, and never framed any of it as a matter of the senses.

Miles: Why do the grounds matter?

Nora: Because a request built on productivity doesn't require anyone to explain themselves. A manager can offer everyone a door, a protected hour, a way to divert calls, and the people who need it most never have to say why.

Miles: Then the evidence on need.

Nora: Group findings, three of them. Williams, Suzman and Woynaroski, Ear and Hearing, twenty twenty one, a meta-analysis on autism and decreased sound tolerance, meaning everyday sound experienced as too loud or painful. Pooling interview and questionnaire studies, about thirteen thousand autistic people, mostly children, the estimate was forty one percent at the time of study and sixty one percent at some point in life.

Miles: That's a majority, over a lifetime.

Nora: In those samples, which skew young. For ADHD, Panagiotidi, Overton and Stafford, twenty eighteen, two hundred and thirty four people from the general population: the more ADHD traits they reported, the more sensory difficulties they reported, in every sense tested. Online questionnaires, measured once.

Miles: And the third.

Nora: The one that argues for control rather than silence. Nigg and colleagues, twenty twenty four, a meta-analysis of randomized lab studies using white or pink noise. In children and young adults with ADHD or high ADHD symptoms, thirteen studies and three hundred and thirty five participants, background noise gave a small improvement on attention tasks. In the comparison groups without ADHD, eleven studies, the same noise made performance worse.

Miles: So the same sound helps one person and costs the one beside them.

Nora: Yeah. The same sound, on those lab tasks. Which is why I'd say the book got the design right without aiming at it. One setting for an open floor can't suit both. A door, headphones you're allowed to wear, a phone that can go quiet, a say in where you sit. Control lets each person set their own.

Miles: Has anybody asked employees what they want?

Nora: One survey I'd use. Tomczak and Ziemianski, twenty twenty three, a hundred and forty autistic adults in Poland, recruited through community networks and passed along from there. Out of twelve proposed workplace adjustments, the two rated highest, both at four point seven nine out of five, were sensory: an office arranged to limit overstimulation, and a workspace people could personalize. A self selected sample, so read it as what this group asked for, not as a rate.

Miles: Cautions.

Nora: Two. First, the join is mine. Nobody has tested private, controllable workspace with neurodivergent employees' work as the outcome. A scoping review this year, Sanders and Le Rossignol in the journal Work, found twenty two pieces of writing on ADHD employees in open plan offices and called the empirical research a significant gap, with most of what exists anecdote and opinion. I've set a correlational table from the eighties beside some sensory research and drawn the line myself, and that's an argument, not a finding.

Miles: And the second.

Nora: Control has to be real. A quiet room you book three days ahead, or a closed door your manager reads as not being a team player, is the furniture police with a kinder policy. And the book lists physical separation as a teamicide, so the aim was never isolation. It's a door the person chooses when to close.

Miles: Fair enough.

Miles: Sixty seconds, the way this show reads it. Peopleware says the hard part of project work is people, that thinking needs long stretches nobody breaks into, and that most offices are arranged against them. For anybody leading a mind that runs on a different pattern, or working inside one, here's Monday. Take your team's E-factor: count the uninterrupted hours in a week against the hours people were there, and treat a low number as a design problem rather than a character flaw. Offer control to everybody, a door or a quiet room that's free when it's needed, headphones without comment, phones that can go silent, so nobody has to disclose anything to get it. Keep shared space and private space both, the way the book's patterns do. And when somebody asks for quiet, hear it the way a productivity study from nineteen eighty seven would have heard it: as a request to do better work.

Miles: And the assistants.

Nora: The fixed caution on these tools comes after the closing music, same wording every time, so I'll leave it there and spend this on what a book about doors adds.

Miles: Go on.

Nora: The largest trial I could find of these tools in ordinary office work. Dillon, Jaffe, Immorlica and Stanton, a preprint, not yet through peer review. Seven thousand one hundred and thirty seven knowledge workers in sixty six large firms, some picked at random to get an AI assistant built into the office software they already used, between twenty twenty three and twenty twenty four. Across everyone offered it, email time fell by about an hour and a half a week, from almost twelve. Among the people who took it up, about two hours.

Miles: And meetings?

Nora: No change they could detect. It trimmed the inbox and left the calendar alone.

Miles: So the hours come back, but in crumbs.

Nora: Yeah. In crumbs, and that's the book's question. An hour handed back as ten six minute scraps isn't flow. So three uses, all on the door's side. Batching: let it gather messages and draft the replies, and open the inbox at two set times rather than every time something knocks. Accounting: have it read last month's calendar and count the stretches of two hours or more that nothing touched, which is your E-factor without a stopwatch. And triage: have it summarize a long thread so you can decide whether it needs you now or at four.

Miles: And what to watch for.

Nora: The assistant becoming one more interruption, with its own pop ups and suggestions. Give it the treatment the book gave the telephone.

Miles: Leave us on something.

Nora: The book's own test, aimed at the tool. Don't ask how many minutes it saved. Ask whether it gave anybody an unbroken morning. If it didn't, the door still matters more than the software.

Miles: That's Peopleware.

Miles: Next time: Death by Meeting, Patrick Lencioni's book from two thousand four, and his name's back on the shelf after The Five Dysfunctions of a Team in episode thirteen. DeMarco and Lister guard the door against the phone and the inbox, but the other thing eating the week has a name and a recurring invite, and Lencioni wrote a whole book about it.

Nora: And the third edition here already has a subheading that says status meetings are about status, so he's arriving in the middle of an argument.

Miles: Nora, you made a door sound like a policy worth fighting for. Thank you.

Nora: Happy to do it.

Miles: More soon.

Miles: Sixty Books is put together by Stoar Consulting and Coaching. Logan Williamson is a Licensed Professional Counselor in Texas and Louisiana. This was one book read closely, for the people who run teams and the people on them. It wasn't coaching, it wasn't counseling, and it wasn't a reading of you.

Nora: The episode page holds a source for every claim, with the transcript alongside. Book forty six arrives in two days.

Miles: Two disclosures before you go. The voices in this episode are synthetic: Miles and Nora aren't people, hold no license, and said nothing about anyone listening. AI helps draft each script, and Logan Williamson, a Licensed Professional Counselor in Texas and Louisiana, at Stoar Consulting and Coaching, then edits it and approves it.

Nora: And one note about the AI segment. The evidence that these tools help neurodivergent people at work is early. There are controlled trials now, but they measure things like conversation skill rather than how a job actually goes, and for work itself most of what exists is individual providers reporting success. So take it as promising and not proven. The tools also get rebuilt every few months, so whatever you set up today is built on something that moves.