Before You Size a Position, Answer These Four Questions

Most investors treat sizing as the last thing that happens after research is done. You've read the annual reports, listened to the concalls, run the numbers, formed a view. Now you decide how much to buy.

That sequencing is the problem.

Sizing isn't what comes after the research. It's the final act of the research itself. The moment you commit capital is the moment you reveal — to yourself, if no one else — how much the research actually produced. And if that moment arrives with any one of four questions unanswered, what you have isn't a sizing decision. It's a guess dressed in the clothing of conviction.

Why are these four questions a test rather than a checklist?

A checklist is something you complete before proceeding. You answer each item, tick it off, move on. The assumption is that completing the list qualifies you to size.

A test is diagnostic. Answering each question clearly is evidence that the research reached a real conclusion. Failing any one of them doesn't mean you need to do more work before buying. It means the research hasn't resolved yet — and the right response is a smaller position, not a longer reading list.

The four questions exist because they force confrontations that normal research lets you avoid. Most research is cumulative: more information, more familiarity, more confidence. The four questions are reductive: they strip away the accumulated warmth and ask what actually remains.

Question 1: Can you state the thesis in one sentence, without notes, and stand behind it if the stock falls 30%?

The first part — one sentence, no notes — tests whether the thesis is a real claim or a collection of positives that feel like a claim.

Most investors can produce a paragraph about why they own a stock. Fewer can produce a sentence. The paragraph is easy because it can absorb contradiction and ambiguity — you lead with the good, mention the risks briefly, balance back to positive. A sentence can't do that. It has to commit to one specific assertion about why this business will compound in a specific way because of a specific structural condition.

If the sentence isn't there, the thesis isn't sharp enough yet.

The second part — standing behind it after a 30% fall — tests whether the thesis is durable or comfort-dependent. A lot of what investors call conviction is actually comfort. The stock has been going up, the narrative is positive, the concall was strong. That's not conviction. That's momentum with a story attached.

A thesis that only holds when the price is cooperating was never a thesis. Ask yourself: if this stock falls 30% over the next six months and nothing in the business has changed, would I still say the same sentence? If the honest answer is "I'd start to question it," the sentence isn't load-bearing enough to size on.

Question 2: Can you name the KPIs that derive directly from the thesis — with specific thresholds?

This question tests thesis precision, and it's the most revealing one.

A vague thesis produces vague KPIs. If you find yourself reaching for industry metrics, macroeconomic indicators, or broad financial ratios that apply to any business in the sector — revenue growth, PAT margins, RoE — rather than metrics that derive directly from the specific claim the thesis makes, the thesis isn't specific enough.

The KPIs should follow from the thesis the way the output follows from the input. If the thesis is about margin expansion driven by a specific product mix shift, the KPI is the revenue contribution from that product category and the gross margin on it — not the overall PAT number. If the thesis is about a regulatory moat protecting pricing, the KPI is realised fee per unit and any regulatory commentary — not the market cap or the P/E.

The threshold matters as much as the metric. "Margins should be healthy" is not a KPI. "Execution margins below 14% for two consecutive quarters would suggest the cost structure assumption is wrong" is a KPI. The threshold is what makes it falsifiable — without it, you're just watching a number move without knowing what it's telling you.

If you can't write KPIs with thresholds that derive directly from the thesis, go back to the thesis first. The problem isn't the monitoring system. It's that the claim underneath hasn't been made precise enough to be measured.

Question 3: Have you written a specific exit condition that is not a price level?

This question tests whether the thesis is falsifiable.

A thesis without a named failure mode isn't an investment claim. It's an emotional orientation toward a business. And emotional orientations are extremely durable — they can survive any number of contradicting quarters, because nothing was ever specified that would break them.

The exit condition is the protection against that. It forces you, before ownership begins, to name the scenario in which you were wrong. Not "the stock falls 40%." Not "the sector goes out of favour." A specific, observable business event — the kind that would appear in a quarterly result or a concall — that would tell you the load-bearing assumption in the thesis has failed.

Writing this is uncomfortable. It makes the investment feel more contingent. That discomfort is the point. An investor who genuinely understands why they own a stock finds exit conditions relatively easy to write — because the load-bearing assumption is already named. The difficulty of writing the exit condition is itself a diagnostic: if you can't write it, either the assumption isn't named yet, or you don't want to specify it because naming the failure mode makes the thesis feel less safe.

The Exit Conditions Framework covers the full mechanics — the three categories, how to use conditions when the stock is moving, and what happens when a trigger fires. This question is asking a simpler thing: does a written exit condition exist at all?

Question 4: What kind of business is this at this stage of its journey — and what does that mean for how you hold it?

This is the question that separates a position from a thesis.

A business can be genuinely good and still require completely different behaviour depending on where it sits in its lifecycle. An early-stage compounder proving out its model for the first time demands a different holding posture than a mature business generating steady free cash flows with a long track record. A transformation play — where the entire thesis rests on management executing a strategic shift — requires different monitoring than an established franchise with a durable moat and predictable incremental economics.

The reason this matters isn't sizing. It's everything else.

How much tolerance do you have for a quarter that disappoints? A mature compounder missing one quarter on a timing issue is noise. An early-growth business missing its key operating metric for the second consecutive quarter is a signal worth taking seriously. The same event carries different weight depending on what stage of the journey the business is in.

How quickly should a worrying concall comment make you act? For a transformation play where management execution is the entire thesis, a shift in tone from the promoter is load-bearing information. For an established franchise with thirty years of operating history, one cautious concall rarely changes the thesis materially.

When does it make sense to add? A business still in early growth that keeps hitting its milestones earns more capital as the thesis strengthens. A mature compounder held for the long term probably warrants a different add logic — not milestone-based but valuation-based.

None of this is resolved by naming a percentage. It's resolved by being honest about what kind of business you're holding and what that business needs from you as an owner over the time horizon that actually matches the thesis.

If you can't answer this question — if the business feels like it could be any of these things depending on the quarter — the thesis probably isn't specific enough to have generated a clear answer to Question 1 either. The four questions are connected. Clarity on one tends to produce clarity on the others. Vagueness propagates the same way.

What does "unclear on any one" actually mean in practice?

The default when any question is unclear is not "do more research and then decide on size." It is a tracking position — deliberately small, held to maintain engagement with the business — with a pre-written add condition: what specifically would have to happen for me to build this to a full position?

That add condition is the fifth implicit question. It keeps the research open without forcing a capital commitment the conviction hasn't earned. It turns an unsized interest into a monitored hypothesis.

When all four questions are answered clearly, sizing becomes less of a decision and more of an expression. The thesis is specific, the KPIs are derived from it, the exit condition is written, the business stage is named and understood. At that point the right size tends to be obvious — not because portfolio construction math produces a number, but because you know exactly how much you believe it and what you're prepared to do with that belief.

The four questions don't tell you what to buy. They tell you how much of what you know is actually actionable.