At the end of every month, someone in the studio sits down to work out what everyone gets paid. It takes a while. They go back through the attendance, such as it is, counting who was in and who wasn't, who took leave and how much, whose leave crossed into loss of pay, what advances went out mid-month and need adjusting. Some of it is in a register, some in WhatsApp, some in memory. It gets pieced together, the numbers come out, people get paid. And then the next month, the same person sits down and does the whole thing again, from the start, as if it had never been done before.
It isn't only payroll. A new team member inherits a project and spends their first week reconstructing it, reading old message threads to work out what was agreed and why. A client asks about a decision made eight months ago and someone has to go digging to reassemble the reasoning. Somebody needs to know which version of a drawing is current and ends up comparing files to be sure. Each of these is the same act. Information that once existed, clearly, at the moment it was created, being rebuilt later from fragments because it was never set down anywhere the studio could return to.
The studio is not short of information. It had all of it, once, at the exact moment each thing happened, when the leave was taken, when the decision was made, when the drawing was replaced. What it lacks is a place that kept any of it. So the knowledge has to be produced twice. Once when it happens, and again, harder, when it's needed, and the second time is always the worse version, assembled from memory and scattered messages long after the details have gone soft.
And that is the part worth sitting with, because the second version is not just slower to produce. It is less true. Memory does not store a decision and hand it back intact. It reconstructs, and where the details have faded it fills the gaps with what seems plausible, and it does this so smoothly that the reconstruction feels exactly as solid as a real memory would.
So when someone asks why the beam moved on that project, or why the fee was set where it was, or why a stage ran long, the answer that comes back is often not what actually happened. It is a reasonable story, assembled after the fact, that everyone then treats as fact. The studio isn't just redoing work. It is quietly replacing what happened with a plausible version of what happened, and acting on the copy.
You can watch this play out on a single question. Eight months into a build, someone on site asks why a structural decision was made the way it was, because it now matters for what comes next. Nobody is certain. The person who handled it has a rough memory, the messages from the time are half-findable, and out of those pieces a reason gets assembled that sounds right.
The next decision gets made on top of that reason. If the reconstructed version happens to match what really happened, no harm done. If it doesn't, the studio has just built its next move on a detail it invented, and it will never know, because the invented detail felt precisely as reliable as a true one.
None of this shows up as a problem, because the cost never arrives in one place. It is spread thin across a hundred small reconstructions a month, each of which feels like nothing more than part of the job. Nobody in the studio ever sees a line that says this is what we spent rebuilding things we already knew. They just feel busy, and slightly behind, and vaguely unsure of a lot of the details of their own projects, and none of it connects into a single visible cost, so nothing ever gets done about it.
The fix was never a better memory, or more discipline about writing things down after the fact, which is just the reconstruction moved earlier. It is that the work should leave a record as it happens. The leave marked when it's taken, the decision noted where the project can hold it, the drawing version kept with the reason it changed, so that when the question comes eight months later there is an actual answer waiting instead of a story to be built.
This is most of what Projectsmate is for. Not to help a studio remember, but to make sure the studio never has to, because the record formed itself the first time and was still there the second. The information a practice needs already exists, in full, at the moment it happens. The only real question is whether the studio keeps it, or throws it away and pays to rebuild it for the rest of its life.