What is the ground effect problem in a team?

The ground effect problem is when a team mistakes a temporary, condition-dependent advantage for durable capability. Borrowed from aviation—where an aircraft gets free lift near the ground that disappears as it climbs—it describes how small, co-located teams get a performance boost from high shared context and near-zero coordination cost. That boost comes from the environment, not the process. When the team scales or goes remote, those conditions vanish, the free lift disappears, and the team stalls even though headcount and effort went up.

Why do small teams perform so well before they scale?

Because proximity does work that doesn't show up on any org chart. In a small, close team, context is ambient: people overhear, ask in seconds, and stay aligned without anyone writing it down. Coordination is nearly free, so almost all the energy goes into output. That's the co-located team advantage. The risk is crediting the speed entirely to talent and process, when much of it is the environment doing part of the job for free.

Why does adding people slow a team down?

Because communication paths grow faster than people do—roughly n(n−1)/2. Six people have 15 possible links; fifty have 1,225. Output grows about linearly with headcount, but coordination cost grows as a curve, so past a certain size each new hire adds more coordination drag than throughput. That gap is the core of Brooks's law, and it's why the coordination cost of scaling teams can quietly overwhelm the extra capacity you just added.

Does going remote cause the ground effect problem?

Going remote doesn't cause it, but it exposes it fast. Remote work removes the ambient context—the overhearing, the chair-swivel question, the hallway fix—that small co-located teams ran on without noticing. That shared context doesn't migrate into tools by itself; it turns into meetings, longer threads, and interruptions. Losing shared context going remote is often the moment a team discovers how much of its old speed was proximity rather than process.

How do you keep a small team's speed as you grow?

You can't recreate one room at scale, but you can manufacture what the room gave you: shared context. Write decisions and ownership down before you need them, design async communication deliberately instead of defaulting to meetings, and keep the work, conversation, and record on one durable surface so context stops fragmenting across apps. The goal is to build real "airspeed"—externalized, durable shared context—before growth lifts you out of the cushion that was carrying you.

The Ground Effect Problem: Why Scaling Teams Stall

# The Ground Effect Problem: Why Scaling Teams Stall

Watch a heavily loaded plane take off on a hot day and you'll sometimes see something that looks like a small miracle and is actually a trap. The aircraft lifts off, climbs a few feet, and then—instead of soaring—settles back toward the runway, mushing along just above the asphalt. It got airborne. It is not, in any real sense, flying.

Pilots have a name for both the miracle and the trap, and it's the same name: ground effect. Close to the runway, the ground interrupts the airflow spilling off the wingtips and hands the plane a slug of extra lift for free. It feels like climb performance. Then the plane tries to leave that cushion, the free lift evaporates, and whatever holds it up now has to come from the wings alone. The aerodynamics is unforgiving: an aircraft can become airborne in ground effect at a speed too low to actually fly once it climbs out of it. The FAA's own handbook warns pilots not to confuse the cushion for airspeed they haven't earned.

Most teams have a ground effect, too. When you're small and close, you get a slug of free performance—and you almost always mistake it for the wing. This is a deep-dive into that mistake: what the lift really was, why it disappears the moment you scale or go remote, how to tell if you're flying on a cushion right now, and how to build the airspeed that actually keeps a team aloft.

What ground effect actually is

Here's the part worth getting exactly right, because the whole metaphor turns on it. Ground effect is not "being good at takeoffs." It's a temporary aerodynamic gift that exists *only* near the surface. The ground blocks the wingtip vortices and the downwash behind the wing, which cuts induced drag and lets the wing make the same lift at a lower angle of attack. Translation: close to the ground, the air does part of your job. Leave the ground, and the air stops helping.

The lethal version of this is well documented in aviation. A plane that's overloaded or short on runway can stagger into the air inside the cushion, the pilot pulls back expecting it to climb, and it simply won't—because it never had the airspeed to fly, only the airspeed to float. The cushion created, in the FAA's framing, "a false sense of adequate climb performance." The crash isn't caused by the takeoff. It's caused by trusting the takeoff.

One honesty note before we map this onto people: a team is not an airfoil, and this is a model, not a law of physics. Humans learn, adapt, and remember; air does none of that. But the *shape* of the trap—free performance that exists only under specific conditions, and a failure that arrives precisely when you leave them—describes a lot of teams better than the org chart does. The ground effect in your company is every advantage that quietly depends on being small and close. The danger is identical to the airplane's: mistaking the cushion for the wing.

The free lift was proximity, not process

Think about why a five-person startup feels superhuman. Six people in one room—or one tight Slack channel and the same three time zones—share almost everything by default. Context isn't documented; it's *ambient*. You overhear the support call. You catch the objection at lunch. You turn your chair around and ask, and the answer costs eight seconds. Nobody writes a status update because everybody already knows the status.

That co-located team advantage is real, and it's enormous. But notice what's actually producing it. It isn't your process—you barely have one. It isn't your management—there's nothing to manage. The small team speed you're so proud of is coming from the *environment*: high shared context and a coordination cost of nearly zero. That's the ground. The air is doing part of your job.

This is the seductive part, and where the trap is laid. Because the lift is invisible and free, you credit it to the wing. "We ship fast because we're a great team." Maybe. Or maybe you ship fast because every person who needs to agree on something is within earshot of everyone else, and agreement is free. You won't be able to tell the difference until you climb—and by then the test is live.

Climb-out: why teams slow down as they scale

The climb-out is the moment you leave ground effect—the conditions that were secretly carrying you. There are two of them, and most growing companies hit both at once: you add people, and you go distributed.

Adding people is the one with hard math behind it. Communication paths between *n* people grow as n(n−1)/2—not in a line, but as a curve. A team of 6 has 15 possible links. At 12 people it's 66. At 50 it's 1,225. Output, meanwhile, grows roughly linearly with headcount. So the coordination cost of scaling teams climbs far faster than the work does, which is the quiet engine inside Brooks's law: adding people to a late project can make it later. The cushion that made coordination free is gone, and you're paying full price for every one of those 1,225 connections.

Going remote pulls the other support out from under you—the ambient context. The casual overhearing, the chair-swivel question, the hallway fix all vanish, and they don't migrate cleanly into tools. They turn into scheduled meetings, longer threads, and "let me loop you in." Microsoft's 2025 Work Trend Index found knowledge workers now get interrupted by a meeting, email, or message roughly every two minutes, and that the share of meetings spanning multiple time zones keeps climbing as teams distribute. That's not a tooling problem. That's the cost of manufacturing, one message at a time, the shared context that proximity used to hand you for free.

Put the two together and you get the team version of the airplane settling back onto the runway. The numbers say you should be flying—more engineers, more hours, more tools. The feeling on the ground says you're slower than you were at six people. Both are true. This is why teams slow down as they scale: the visible inputs went up while the invisible lift went away, and nobody put the lift on the books.

How to tell if your team is in ground effect

The dangerous thing about ground effect is that, from inside the cockpit, floating and flying feel almost the same—right up until you need to climb. Same with teams. Here are the tells that your scaling team performance is running on a cushion, not a wing.

Your speed is tied to specific people being awake

If work only moves when two or three particular people are online and reachable, you don't have a fast process—you have a small, well-connected core doing the coordinating in their heads. That's pure ground effect. It works beautifully until those people go on vacation, get promoted, or simply can't scale their attention past the next ten hires.

Onboarding has quietly gotten brutal

When you were small, a new hire absorbed context by osmosis and was useful in a week. If new people now take months to get oriented and keep tripping over decisions nobody wrote down, that's the cushion thinning out. The early magic was never your onboarding program—it was the proximity that made an onboarding program unnecessary. (This is close kin to the cold-start problem: every new person needs context the team never had to externalize before.)

Decisions evaporate the second they're made

In a room, a decision is overheard and remembered by everyone who matters. Distributed and scaled, the same decision gets made in one call and is invisible to the six people who needed it by Thursday. If your team keeps relitigating settled questions, you're watching context that used to be free fail to transfer.

You're adding tools and getting slower

The instinct when the lift fades is to bolt on software—another tracker, another notetaker, another AI bot in the meeting. If the app count keeps rising and the velocity keeps falling, you're paying the cost of context-switching to rebuild, badly, the shared awareness a small room gave you for nothing.

Building real airspeed before you scale

The pilot's fix for ground effect is not "stay near the runway forever." It's: get real airspeed *before* you try to climb. The team version is the same. You can't recreate a six-person room at sixty people—but you can deliberately manufacture the thing the room was giving you for free, which was shared context. That manufactured context is your airspeed.

Three moves build it. First, write the context down *before* you need it—decisions, the reasoning behind them, and who owns what—so the answer survives the person who knew it leaving the channel. Second, design async on purpose instead of letting it happen to you: a clear default channel, real status that doesn't require a meeting to extract, and the discipline to stop using a 30-person call to do a 3-person decision. Third, give the team one durable surface where the work, the conversation, and the record actually live together, so context stops scattering across a dozen apps and re-fragmenting every week.

That last one is the hardest, and it's where most stacks make the problem worse, not better. The reigning answer to lost context is to staple tools together—video here, canvas there, an AI notetaker that joins like a stranger and makes everyone self-censor. Each tool is a separate surface with the others bolted on, so the context you're trying to preserve splinters across all of them. We built Coommit on the opposite bet: put the video call, the collaborative canvas, and the contextual AI on one surface, so the AI already sees the work and hears the conversation instead of being invited in after the fact. The point isn't more features. It's manufacturing the closeness a small co-located team had by accident—so the lift doesn't vanish the moment you climb.

When ground effect is worth keeping

Now the steelman, because the cushion isn't the enemy—mistaking it for the wing is. Plenty of great teams stay deliberately close, and they're right to. Two-pizza teams, tight founding pods, a strike team locked in a room for a launch: these keep coordination cost near zero on purpose, and they fly fast because of it. The lift is real. There's nothing wrong with using ground effect when you've chosen it with eyes open.

The skill is knowing which mode you're in. A small team that *intends* to stay small can lean on proximity indefinitely—that's not a trap, it's a strategy. The trap is only sprung when you intend to scale, keep flying on the cushion, and never build the airspeed the climb requires. So the question to ask before you grow isn't "are we fast?" You are; the ground is helping. The question is: "if we doubled and half of us went remote tomorrow, what carries this team—and is any of it written down?" If the honest answer is "the people who happen to sit near each other," you're in ground effect, and you have flying to do before you climb.

The runway you can't stay on

The seductive lie of a great small team is that the speed is *yours*—a permanent property of the people, portable to any size and any geography. For the part that's genuinely talent, it is. For the part that was proximity, it isn't, and the cruel design of the ground effect is that you can't tell the two apart while you're still close to the ground. The cushion feels exactly like competence until the moment you ask it to climb.

Growth is the climb, whether you planned it or not. More people, more distance, more time zones—every one of them lifts you out of the regime that was quietly carrying you. The teams that stall are the ones that read their early speed as airspeed and pull back on the stick expecting more of the same. The teams that make it are the ones that looked down, saw how much of the lift was the runway, and built real airspeed before they ran out of it. You don't get to stay near the ground forever. So the only question that matters is whether you'll have the airspeed when the cushion lets go.