It usually begins with a kindness. Someone has had a rough few weeks, so we tell them to slip off early on Friday. It costs nothing, and we are small enough that being decent about it feels like the whole point. Then the early finish becomes the whole afternoon. Then the whole of Friday. A few months on, one of my best people is in the office half the week and "working from home" the rest, a phrase that has started to arrive wrapped in invisible quotation marks. Nobody decided any of this. It drifted. And one ordinary morning I catch myself resenting people I genuinely like.
If you have run a small business, you know the feeling. It is not a personal failing, and it is not a sign you hired badly. It is structural, and there is a good deal of research behind it.
Why small teams go soft
Accountability is hard in a small company for the same reason a small company is a good place to be. The relationships are real. Robin Dunbar's work suggests we each carry only so many close relationships, layered in rings of roughly five, fifteen, and fifty people. A company of twenty or thirty sits right inside those rings. You do not have "direct reports." You have people whose kids you know by name and whose weak spots you have quietly learned to work around.
That closeness is an asset until the moment you have to hold one of them to a standard. The research here is uncomfortably consistent. We go easy on people we like. Edward Thorndike named it the "halo effect" a century ago, our warmth toward a person coloring how we judge their actual work. Later studies found that more agreeable managers rate people more leniently to keep the relationship intact, and that managers form an in-group of the people they are close to and cut them more slack on discipline. None of it is corruption. It is friendship doing what friendship does. But stack it up and the standard bends toward whoever the owner is fondest of, and the hard conversation keeps getting pushed back, because having it would cost something personal.
The uncomfortable thing about scale
It is often easier to hold the line in a big, impersonal organization than in a small and personal one. Big companies can run more efficiently, but what looks like a small-business problem is really a small-group one, and the same quiet nepotism grows inside any tight team, even one buried in a large firm. A big company can let someone go and have it read as the organization deciding, the policy requiring it. There is a machine to stand behind. In a fifteen-person firm there is no machine. There is just you, on a Tuesday, telling someone you have shared a hundred lunches with that it is not working out.
Sebastian Junger's "Tribe" is a useful way to see what a small group is for. Humans evolved in tight bands where everyone relied on everyone, and standards held through belonging and plain social pressure rather than any rulebook. That is a small company at its best, and it is also the trap. The same bonds that make people show up for each other make it hard to hold any one of them to account. Scale runs the other way. Grow large and anonymous and accountability can quietly vanish, or you can rebuild it with an impersonal system, so enforcing a standard no longer rides on a friendship surviving it. A startup keeps all of the closeness and none of that machinery. That is why the job is harder there.
But scale is no free lunch either
Grow up and get bureaucratic is not the answer either. Big groups carry their own disease. People help less as the crowd grows, and pull less hard on a rope in a group of eight than they do alone, and both effects are well documented. Big organizations are quietly full of people who have worked out that their own contribution is neither visible nor needed. So size settles nothing on its own. What drives accountability is whether the individual is visible by name, and whether there is a real process around them. Gallup found that engagement held up even on large teams when the manager gave real feedback every week, and collapsed without it even on small ones. A small company gets the visibility for free. What it lacks is the process. A big company has the process and loses the visibility. The one that builds both wins.
Which is really a question about your wave
You get to pick your wave. There is no "right" business, only the one you have chosen and the swell that comes with it. You set your expectations and you choose your trade-offs on purpose. Balance is a choice, not an accident. And the rule I keep coming back to is that you should not try to ride two different waves at once.
Staying small, building a firm that never goes much past fifty people, is a completely legitimate wave, and often a fine one. But picking it means taking the conditions that come with it. Choose the closeness and you also choose the accountability problem I have been describing. You do not get to keep the closeness and also expect the frictionless enforcement that only impersonal scale provides. You have to build the discipline by hand, precisely because you have opted out of the machine that would build it for you. The good news is that this is doable, and doable without turning into the boss nobody wanted to become.
First, borrow the impersonality. The most useful thing a small-business owner can take from big companies is not coldness. It is a written standard that lives outside the relationship. "The role needs someone in the office Monday through Thursday" is a fact about the job. "I am let down that you are never here" is a fact about the friendship. The first is enforceable without spending anything personal. The second is not. Write the expectations down before you ever need them.
Second, make feedback a rhythm rather than an ambush. Hard conversations feel personal in small firms mostly because they only happen once something has already gone wrong. Put a light, regular cadence in place, a proper one-on-one every couple of weeks with real substance in it. Gallup's finding is that the cadence itself carries the accountability. It also means the difficult message arrives as the conversation you always have, not as the boss finally losing patience.
Third, name the norm before it drifts. The Friday creep happens because nobody ever decided the policy. It defaulted into being. So decide it out loud, as policy. Here is the arrangement, here is what working from home actually means for output and availability, here is how we will know it is working. Ambiguity is what quietly kills accountability, and small teams run on ambiguity because settling things in plain words feels too formal among friends. Settle them anyway.
Fourth, protect what makes each person visible. Loafing feeds on work that cannot be traced back to a name. Give every outcome an owner. In a small team this is easy, and you should use the advantage without embarrassment, because it is the one thing the big firms would love to have and mostly do not.
Fifth, hold the kindness and the clarity at the same time. The best research on strong teams describes pairing real safety with real standards, not trading one for the other. The cool boss keeps the safety and quietly drops the standard. The tyrant does the reverse. The actual job is both at once. I am on your side, and this needs to change. In practice it sounds like, "I think you are excellent, which is exactly why I am going to be straight with you about this."
And in the end you have to have the conversation yourself, because there is nobody to hand it to. No HR department, no faceless organization to carry the blame. That is uncomfortable, and it is also the whole point. Anchor the conversation to the written standard rather than to your own hurt feelings. The standard is what makes it fair. Your willingness to own it is what makes it land.
Being liked and being respected are not the same thing. The trap of the cool boss is trading the second for the first and then losing both. A team that watches the slow fade go unchallenged loses respect for the person who let it. In time it stops much liking the place either. You started small because you wanted the relationships. Keep them. Just do not mistake letting the standard slide for being good to work for. The kindest thing you can do for people you care about is to be someone whose word about the work still means something.
Pick your wave. Then build the discipline that wave asks for, because at fifteen people nobody hands it to you. That was always the deal. And it is a good one.
Sources and further reading. Robin Dunbar on the social brain and the layers beneath 150. Sebastian Junger, "Tribe: On Homecoming and Belonging" (2016). E. L. Thorndike, "A Constant Error in Psychological Ratings" (1920). Cheng, Hui, and Cascio on leniency and relationship maintenance (Frontiers in Psychology, 2017). Graen and Uhl-Bien on leader-member exchange (1995). Darley and Latané on diffusion of responsibility (1968). Latané, Williams, and Harkins on social loafing (1979). Gallup, "Span of Control" (2025). Amy Edmondson on psychological safety (1999).