About
Most companies are not mispriced because the market failed to read the filings. They are mispriced because nobody is in charge.
Ownership has become passive. The largest holders of most listed companies cannot sell and do not want to engage, so nobody on the register has both the incentive and the standing to say the obvious thing. Into that vacuum steps the consultant, and out comes a 192 page strategy deck that nobody owns, nobody executes, and nobody is accountable for. Outsourced strategy is the great plague of the listed company. A firm that has to buy its own thinking has already lost the argument.
The result is a large population of businesses that are not failing and not improving, carrying structures assembled for a world that no longer exists, waiting for someone to behave like an owner.
The opposite case is rarer and more valuable. Occasionally you find a company where someone has already worked out where the market is going and is quietly building for it, years before anyone notices. Those are worth more than any restructuring, and they are just as badly covered.
This publication is about both. Finding the companies being built properly, and the ones where the thinking has not been done.
Why public markets
Every year there are fewer listed companies and fewer people looking at them. Research budgets were unbundled, coverage below large cap collapsed, and capital moved into private markets and into indexes that buy without looking. What is left is a market with more money in it and fewer eyes on it, full of companies nobody is holding to account.
Private markets have the opposite problem. Every asset is picked over by people with more capital and better access than I will have for years. The public market is the one place where a single person willing to do unglamorous work still has an edge over institutions, because the institutions have stopped doing it.
That is the most obvious place to do serious work that I can find.
The model is usually the problem
The most persistent mispricings are not about numbers. They are about the wrong story being told, and often the company is telling it about itself.
Businesses frequently do not know why they work. A management team attributes success to the product when the actual engine is distribution, or to the brand when it is a regulatory quirk that made a channel cheap. Investors inherit that explanation, model the wrong thing, and price the wrong risk. Everyone is looking at real numbers through a broken lens.
This runs both ways. A company can be far better than its own account of itself, sitting on a structural advantage nobody has named. Or much worse, with a moat everyone believes in that turns out to have been an accident of timing.
So a large part of the work here is not forecasting. It is working out what is actually going on, then checking that against what the company believes and what the market believes. The gap between those three is where most of the opportunity lives.
How I work
Most of the work is reading things nobody wants to read.
Patent filings and their citation trails. Job ads and who is being hired where. Customs and shipping records. Regulatory submissions. Municipal planning documents. Supplier lists buried in an appendix. Conference abstracts. The Finnish and Swedish filings that no international analyst opens because they are in the wrong language. Almost everything that matters about a company is already public. It is just tedious to find, tedious to connect, and nobody is paid to bother.
Then it is reasoning. Taking a dozen unrelated fragments and working out what they must imply, then testing that against everything else available. Most of my best work started as a small inconsistency that did not fit the accepted story, and ended several weeks later somewhere nobody expected.
When a question genuinely needs a model, I will build one and go to considerable lengths to get it right. I have run a revenue forecast on one company for fifteen consecutive quarters, published before each report, with an average error of around two percent. But the model is the exception, not the method. Most of the time the answer is sitting in plain sight in a document nobody opened.
The current focus is where AI is quietly reshaping the economics of businesses whose reported numbers have not caught up yet. I do not think of this as a technology cycle. I think intelligence is substrate independent, that there is no meaningful line between what a mind does and what a machine can do, and that we are living through the compounding phase of that fact. The right response is to build faster, not to slow down. Most companies have not internalised any of it, and the distance between what is happening to them and what they are disclosing is where the interesting work is.
How I think about companies
I read companies the way an owner would, not the way a trader does.
The question is never only what is this worth. It is what is the good business inside here, where is its market actually going, and does the company understand that.
Sometimes the answer is that it does. Then the work is to see it earlier and more precisely than anyone else, and hold on. That is the most enjoyable version of this job and the one that compounds.
More often nobody has laid out where the industry is heading or what follows from it. Sometimes that implies a separation, sometimes a different shape, sometimes just a clear statement of what the company is for. The work is not persuasion. It is doing the thinking that should already have been done, in enough detail that acting on it becomes the obvious next step.
Strategy is not a document. It is a decision someone is accountable for. I come from a family that has built things in Finnish industry for four generations, and I have spent enough time in boardrooms to know that inaction is rarely about information.
Where this started
It began with a single Swedish small cap that nobody covered. I wanted to know how far ahead of a quarter you could actually see if you were willing to do the work, so I built a model to predict its revenue from public web data, published the prediction before every report, and let the results grade themselves.
That company was Haypp Group, and the work turned out to be about something larger than one name. The market read it as a discount retailer. I argued it was building an advertising channel in a category where advertising is otherwise illegal, which is a different business with a different value. Same numbers, different lens.
Everything here is an extension of that. How much is knowable in advance, and how little of it anyone bothers to know.
Who reads this
Institutional investors, fund managers, analysts, and operators, alongside people who simply enjoy watching a thesis get built in public. Companies I have written about read it too, which is a useful discipline: nothing goes out that I would not defend to the management team in question.
What you get
Deep dives on companies almost nobody covers, built from primary data rather than from other people’s research.
Running measurement of structural change in businesses that have not yet acknowledged it, published as forecasts that can be checked.
And occasionally, a case where the accepted story about a company turns out to be wrong in a way that matters.
Free subscribers get the majority of the writing. Paid subscribers get the full research, the positions I take and why, and the work I would not put in front of an audience I do not know.
