The sort was going fine until the founder put everything in Core.
Not out of stubbornness. That is the part people get wrong when they tell this story. She went item by item, gave each one a real answer, and the answers were true. The reporting module closed the deal that kept the company alive in year two. The onboarding flow was rebuilt three times before it worked, and it works. The integration nobody uses is used by two customers, and one of them is the reference that gets them into every enterprise conversation they have.
Every sentence she said was a fact. And at the end of ninety minutes the Bloat column had one item in it, a cron job nobody could identify, and the Cut list was empty.
She was not defending the system. She was remembering it. Those feel identical from the inside, and they are not the same thing.
Value is something you assess at a distance. You look at a part of the product, ask what it does for the business now, compare that to what it costs, and make a call. That works when you can see the part clearly, and it requires a gap between you and the thing.
The founder has no gap. They were in the room when each piece was built. They remember the fight it won, the customer it saved, the quarter it rescued, the argument they had to win to get it approved. That memory is real, it is accurate, and it attaches to the part permanently.
So the founder is not looking at the part. They are looking at the part plus its history, fused, and trying harder does not separate them. The history is what makes it feel essential.
This is Core Drift, and it is the same mechanism rather than a metaphor for it. Here the label lives in a memory. The value moved, the label stayed, held in place by the one person in the company with the most vivid and most accurate recollection of when it was correct.
Proximity does not make you sentimental. It welds what a thing was worth to what it is worth.
So the obvious question is whether you can correct for it once you know it is there. Mostly not, and not by trying.
A blind spot is not a bias you can lean against. A bias has a direction, and if you know the direction you can push back. A blind spot is a hole in what you can perceive, and it presents as normal vision. When you look at the reporting module and see something essential, you do not experience that as memory contaminating judgement. You experience it as looking at a reporting module and seeing that it is essential.
You can decide to be more sceptical, and you will spend that scepticism on the things you were already unsure about, which is precisely where the problem is not. The parts you are certain about stay certain, and those are the ones with the most history attached.
Which means a founder can read all ten of the pieces before this one, hold every shape in their head, run the practice exactly as designed, and still walk out of the room with an empty Cut list, having done everything right.
That is the limit of the method.
The fix is not more discipline. It is a different pair of eyes, and the reason is mechanical rather than moral.
Someone who was not there does not have the fusion. They see the reporting module and they see a reporting module: what it does, who touches it, what it costs. The deal it closed in year two is not attached to it, because they were not in the room when that happened. They are missing exactly one kind of information, and that is the entire point.
You do not need to hire anyone for this. In the sort I opened with, the person who broke it open was a principal engineer who had joined eight months earlier and worked on something else entirely. She had no history with the reporting module. She asked who was using it, the room went to look, and the answer was three accounts, two of which had churned. She was not smarter than anyone else in that room. She was the only person in it who had never had a reason to be proud of that module.
That seat is already in the practice. The sorting room needs someone who can ask why do we still have this without paying for it, and this is what that seat is for. Not politics. The room needs one person whose sight is not fused to the history, and the founder cannot be that person however honest they are.
What you cannot do is be that person for yourself. Distance is not an attitude. It is a fact about where you were standing.
I am not exempt from this and I have the receipt.
I had a German phone number on my site. Nobody called it. I had the data, I looked at the data, and I removed it, which is exactly the move this series spends eleven issues recommending.
A friend asked why it was gone. Because nobody uses it, I said. He told me it was not there to be used. It was there so that a stranger landing on the site could see that a real person with a real number was behind it.
I put it back. Enquiries went up afterwards, which is not proof of anything, and I am not going to pretend it is. One change, no control, a small sample.
That is not the part that matters. The part that matters is that I had no way to see it. I was assessing the number by what it did, because I was the one with the logs, and I could not see what it signalled to someone arriving cold. The founder sees the part plus its history. I saw the part plus its usage data, and both of those are information the outside does not have and cannot be distracted by.
Which is the uncomfortable half of this. Proximity does not only make you keep dead things. It makes you cut live ones, for good reasons, with data in your hand.
Now the reflex, because it will show up in the room, and if you are the founder it will show up inside your own chest.
They don't understand our product.
That sentence is usually true and usually irrelevant, and the trick is knowing which half you are holding. Yes, the outsider does not understand your product the way you do. They do not know the customer's name, the reason for the odd thing in the schema, the compromise from 2021. All true.
And none of it is what they are for.
An outside review of a system is not a trial of the decisions that built it. Nobody is asking whether you were right to build the reporting module. You were. That it was right when you built it is exactly why it is so hard to remove now.
The question is narrower. What does this thing do for the company today, and does that justify what it costs today. Answering it requires not knowing why it exists.
Which is why "you don't understand our product" is not a rebuttal. It is a description of the qualification.
And that closes the thread, because it takes us back to where this started.
The first thing I wrote in this series was about a room full of competent people making a decision that cost them a year, not because anyone was wrong, but because a question never got asked. That has been the argument the whole way through. The technical debt that was a product decision nobody wanted to make. The project risk sitting in plain sight because it was not technical and the room was watching the technical. The part still wearing the Core label after the value walked out. The thing nobody would remove because nobody could say what would break. The architecture that was right at the scale you were. The decision that was never written down and so got made again, expensively.
In most of them the question was sitting right there, it cost nothing to ask, and nobody asked it.
The last one is this. The missing question you are most likely to skip is not the one that is hard to answer. It is the one you cannot see is missing, because you were standing too close to the thing it was about.
That question exists in your product right now. You have walked past it more times than you would like. It is attached to something you are sure about, which is why you have never stopped to ask it, and it is why you will not find it by thinking harder.
You will find it when someone who was not there asks you why you still have that.
Cases in this series are composites. The patterns are real, the companies are not.
The full Core and Bloat™ method, four boxes and all, sits at dirkbauer.com/core-and-bloat.html.

