I get asked some version of "do you actually use this yourself?" often enough that it deserves a straight answer. Yes. I built it for my own account and I still use it that way. Here's honestly how, including the parts where I ignore it.
The short version: the model narrows the field, and I make the decision. It turns 2,300 companies into a list short enough that a person with a job can actually research it. That's the job it does. It doesn't tell me what to buy.
The Top 30 is a starting point, not a shopping list
I don't mindlessly buy the top of the list, and I'd be uneasy with anyone who did. What the Top 30 gives me is a signal to look more closely — thirty names the model thinks are strong, growing businesses at a fair price, which is a very different claim from thirty names I should own.
Two kinds of names show up there, and both are useful in different ways.
Some I've been watching for a long time already. Seeing one surface in the rankings is a nudge — the model independently thinks what I've been thinking, which is worth something even when it doesn't change my mind.
The other kind is the reason I keep reading it: companies I have never heard of. That's the part a screen does better than a person can. My own attention naturally circles the same few dozen businesses I already follow; the model has no such habit, and it reads all 2,300 with the same standard every day. Several of the most interesting things I've researched arrived this way — not as recommendations, but as here is a company that scores well and you have no idea why. Then I go find out why.
The part I use most
Search and compare, by a wide margin.
The typical sequence is that I've narrowed a decision down to three or four names — usually within the same industry, all of which look reasonable — and I need to see them next to each other. Reading four sets of financials one at a time and holding them all in your head is exactly the thing I built this to stop doing. The compare page puts the same metrics side by side, scored the same way, so the differences are visible instead of remembered.
That's the whole reason the site exists, really. Not the ranking — the ability to hold several companies up against each other on identical terms and see where they actually diverge.
What moves something up the watchlist
A good score gets a company looked at. What gets it taken seriously is agreement between independent things.
When the fundamentals look right and the DCF cross-check points the same way, a name moves up my watchlist. Two different methods reaching the same conclusion is a stronger signal than either alone, because they can fail in different directions — a screen can be fooled by a good trailing year, and a DCF can be fooled by optimistic assumptions, but both being fooled the same way at the same time is less likely.
When they disagree, that's informative too. It usually means I've misunderstood something about the business, and it's worth finding out what before doing anything.
How I open a position
Never all at once.
I decide up front how much money I want in a position in total, and then I buy it in pieces over weeks or months. The decision I'm making is how much of this company do I want to own — not is today the right day to own all of it. Those are separate questions, and I only think I can answer the first one.
Buying in tranches means I don't have to be right about timing, because I'm not making a timing call at all. If it falls after my first buy, the rest of my purchases are cheaper and I'm glad. If it rises, my first tranche is already working. Either way I avoid the specific misery of putting the whole position in on a single day and then watching that day turn out to have been the high.
It also gives me time to keep reading. A position built over two months is a position I've watched report at least once, and occasionally what I learn in that window stops me completing it.
Themes, when I want deliberate exposure
The themes pages get used differently. Rather than "what's good today," they answer "I want more exposure to this idea — who's actually in it, and how do they score?"
A real example. I wanted exposure to cybersecurity, and it came down to Palo Alto Networks and CrowdStrike. I went back and forth. Both are genuinely strong businesses, they score closely, and the honest conclusion I reached was that I did not have a good enough reason to prefer one over the other.
So I bought both.
That's not a clever strategy and I'm not presenting it as one. It's what you do when the analysis genuinely doesn't separate two names and you'd rather own the thesis than manufacture a conviction you don't have. The alternative — picking one and inventing a rationale after the fact — is a worse habit than splitting the position.
When I don't understand a business
This comes up constantly, because the model surfaces companies I've never heard of by design.
My response is not to skip them. It's to go find out how the company actually makes money — what it sells, to whom, and why they keep buying. That research is the real work, and the score is just what pointed me at it.
It's been worth doing. Snowflake is a name I'd never have arrived at on my own; I had to sit down and learn what the business actually was before I could have any view on it at all. Digging into companies I didn't understand has produced some of the best positions I've owned.
The rule isn't "avoid what you don't understand." It's "don't buy what you don't understand — go understand it first." Those sound similar and lead to completely different portfolios. The first one keeps you inside whatever you happened to already know. The second one keeps expanding it.
Where the model has nothing to say
It doesn't know anything about my situation. It doesn't know what I already own, what a new position would do to my concentration, what I paid for anything, my tax position, or my timeline. Those are the inputs that actually decide whether a good company is a good purchase for me, and none of them are in the score.
So the score is never the last step. It's close to the first one.
The one-sentence version
The model does the elimination. I do the judgment.
It's very good at the first job — reading everything, applying one standard, never getting bored or attached. It is incapable of the second, and any screen that claims otherwise is selling you something.
Disclosure: I hold positions in the companies named above. Nothing here is advice — it's a description of my own process, not a recommendation that anyone follow it.