Whatever You Build Will Drift by Thursday

Structures that work do not hold themselves. A week after you install one, someone makes a decision outside the agreed lane and a scheduled update does not go out. That is not evidence the structure failed. It is the ordinary condition of anything running under load.

You put something in place. Maybe it is a standing update every Monday so people stop hearing news third-hand, or a rule about which decisions come to you and which do not, or an agreement that anything affecting two teams gets said out loud before it gets scheduled. You thought it through, you explained it, and the first week ran close to the way you pictured it. People used the new lane. The update went out on time. Someone told you it was helpful.

Then a Thursday arrives. A decision that belonged in the new lane gets made in a hallway, because the person needed an answer that afternoon and you were in something else. The Monday update goes out on Wednesday, and the following week it does not go out at all. A month in, a version of the old habit is running quietly alongside the thing you installed, and nobody has mentioned it, because nobody experienced any single instance of it as a violation.

Most leaders read that sequence as a verdict. Either the structure was wrong for this team, or the team did not take it seriously, or they themselves did not push hard enough. All three conclusions lead to the same action, which is abandoning something that was mostly working and designing a replacement. The replacement drifts on roughly the same schedule, and after two or three rounds the unit learns that new structures last about a month, which is a far more expensive lesson than any of the failures that taught it.

Drift is what use looks like

You designed for an ordinary week. What arrives is the actual week, which contains someone out sick, a deadline that moved, a request from above due Friday, and a person who needed a decision while you were unreachable. Each of those produces a local exception, and each exception is correct at the moment somebody makes it. Nobody is cutting a corner. People are getting the work done with the structure temporarily set aside, which is what any competent person does when the two are in conflict.

The trouble is that nobody records the exception, because recording it would suggest something went wrong, and nothing did. Six defensible exceptions later, the working practice has a shape nobody designed and nobody agreed to, assembled entirely out of reasonable individual choices. That is drift, and it is not a discipline problem. It is the predictable result of putting a fixed structure into a variable environment, which is the only environment there is.

Timing is the part worth knowing. Drift is nearly invisible for about a week and close to permanent after about a month, and the difference between those two states is social rather than technical. Inside the first week, everyone still remembers the agreed version clearly enough that naming the gap costs nothing. After a month, the exception has been repeated enough times to have become the practice, and naming it means telling several colleagues they have been doing something wrong for weeks. Almost nobody takes that on voluntarily, so the structure stays formally in place and quietly unused.

Three ways it shows up before anyone calls it a problem

The exception nobody calls an exception anymore. The first time a decision skipped the agreed lane, two people noticed and one mentioned it. By the fifth time, it is simply how urgent decisions get made here, and the lane now covers the decisions that were never urgent, which needed it least. Ask whether the structure is being followed and people will say yes and mean it, because they are thinking of the cases it still covers.

The form that outlives the function. The standing meeting still runs for its full thirty minutes, and no one has raised anything difficult in it since April. The written update still goes out, and it now contains only items everyone receiving it already knew. Attendance and delivery both look healthy on any measure you might apply, which is why this version of drift can run for a year without anyone reporting it.

The parallel channel. A small group has found a faster way to get the same thing done, usually two people and a direct message, and it works well enough that they are not going to stop. The structure has not been rejected. It has been routed around by people trying to be effective, and the workaround is useful information about which part of the design was too slow.

Three moves

Schedule the check rather than waiting to notice the failure. When you install something, put two short reviews on the calendar at the same time, one at two weeks and one at six. Ten minutes each. The purpose is not to enforce anything. It is to create a moment where naming a gap is expected rather than volunteered, which is the entire difference between catching drift at day nine and discovering it in February.

Say what actually happened without treating it as a breach. The wording carries most of the weight here. There is a large distance between asking why people stopped following the process and saying that the agreement was one thing, the practice for three weeks has been another, and the question is which of the two is better. The second version invites the person who invented the workaround to explain why, and that explanation is usually the most useful thing anyone will tell you about your own design. Handle drift as maintenance and people report it early. Handle it once as a compliance matter and you will not hear about the next one.

Decide in advance which parts are load-bearing. Most structures contain one or two elements that carry the whole purpose and several more that are convenience. If the point of the Monday update is that nothing significant reaches the team through a hallway first, then the timing is convenience, the completeness is load-bearing, and an update that goes out Wednesday with everything in it has not drifted at all. Leaders who have not made that distinction defend the calendar slot as hard as the thing that matters, and their teams learn that all of it is equally rigid.

The part that is about you

You will drift first, and your drift is the one nobody on your team will name. When you take the hallway decision because you are pressed for time, you are not observed as breaking the agreement. You are observed as amending it, since you are the person with the standing to do that, and the amendment takes effect whether or not you intended one. That is how a structure with your genuine support collapses inside a month, and it will look to you like the team let it go.

There is a harder question underneath the practical one. Some structures get abandoned because their drift is inconvenient, and some get abandoned because holding them would require the leader to make a decision they would rather not make. A rule about which decisions come to you is easy to install and uncomfortable to keep, because keeping it means letting something get decided in a way you would not have chosen. If you find yourself building a third replacement for the same structure, the design is probably not the thing that keeps failing.

Three questions

Name a structure you put in place more than three months ago that is still running the way you described it on the day you installed it. If you cannot name one, everything you have built has either drifted or been replaced, and you have been treating both as normal.

For the most recent thing you installed, which single element is load-bearing? Say it in one sentence. If the sentence needs four clauses, your team has no way of knowing which part to protect when the week goes badly.

When did somebody last tell you, without being asked, that an agreed practice had stopped happening? If nobody has in the past six months, your unit is not free of drift. It is not reporting it, and the second condition is the one that costs you.