Finding Companies at the Point of Inflection

We recently launched our proprietary Potential J Curve Setups scan.

In this post, we will explain what the scan is, how we built it, and where we think fundamental scanning is heading.

J curve scan

The idea behind a J curve is simple. Some companies spend years doing the difficult work before their earnings begin to accelerate. They build capacity, win customers, develop new products, repay debt, or enter a new market. For a long time, the financial statements may look ordinary.

Then something changes.

The new plant starts operating. The order book begins converting into revenue. Utilisation rises. Margins expand. Profit starts growing faster than sales. The earnings curve turns upward.

That is the shape we are trying to find.

The scan looks for companies where something material is changing in the business and where the early work has already been done. The trigger could be a new product, a new factory, a large customer approval, a filled order book, or a structural change in the industry.

The word “new” is important here. A company growing at the same rate it has grown for years may be a good business, but it is not necessarily a J curve setup. We are looking for a change in the path of the business.

This was an ambitious project. Fortunately, Investorstack has spent the last few months building a deeper research layer for companies. We used the research reports, valuation models, thesis, Growth Triggers, Bear Cases, KPI data, capex tracking, and other company level research already available on the platform.

We applied these frameworks across the companies in our coverage universe to identify businesses where a visible change could lead to stronger earnings over the next few years.

The result is the Potential J Curve Setups scan.

We have also built checks into the process to reduce false positives. A company does not qualify just because its profit has grown sharply. The improvement needs to come from something real in the business, and there needs to be a reason to believe it can continue.

My own investing style combines momentum with a material change in the business. I want to find companies that are beginning to look better fundamentally and are also showing signs of strength on the charts. The J-curve scan focuses on the first part: identifying what is changing inside the business.

I run a portfolio of around 15 companies. Ten of them currently appear in this scan. That does not make the scan right, but it is a useful personal check on whether the framework is finding the kind of businesses I would study myself.

The process behind the scan is more important than the label. The rest of this post explains how we judge a potential J-curve and how the framework separates a genuine business inflection from a temporary earnings jump.

Draw the letter J on a piece of paper. It sits flat for a while, then it turns a corner and shoots upward. That is the shape we are hunting.

Think about a real business. A company decides to build a second factory. It spends two years and a lot of money building it. During those two years the profit does not go up. If anything it goes down, because the company is spending. The chart is flat. Nobody is interested.

Then the factory opens. It starts making things. It starts selling them. Suddenly the same company, with the same staff and the same customers, is earning far more than it was, and the cost of running it barely moved. The line turns the corner.

The flat part is boring and nobody watches it. The steep part is exciting and everybody watches it. The money is made by the people who were already there during the boring part.

You will see the word inflection everywhere in this scan, so it is worth ten seconds.

An inflection point is simply the place where the line stops being flat and starts going up. The bend. The corner of the J.

That is it. There is no more to it than that. When we say a company is “at an inflection”, we mean the boring flat part looks like it is ending and the climbing part looks like it is starting.

Why we care so much about the bend: everything before it is cheap to buy because it looks boring, and everything after it is expensive to buy because it looks obvious. The bend is the only place where you can see what is coming and still be early.

Here is the part that matters most, and the part most people get wrong.

A company reporting that its profit tripled is not automatically a J-curve. Profit can triple for boring, temporary, meaningless reasons. It can triple because last year was terrible. It can triple because the company sold a building. It can triple because a tax refund landed.

For it to be a real J-curve, the extra profit has to come from something physical that actually happened, and that you can put a date on:

• A new factory got switched on.

• An order book got filled.

• A big customer approved the product after two years of testing.

• Prices went up and stayed up.

• A pile of debt got paid off, so the interest bill shrank.

Something real, that happened, on a date. If nobody can point to the thing that changed, there is no J-curve, no matter how good the numbers look.

A J-curve does not appear from nowhere. It walks through the same six steps every single time, in the same order.

1. Trigger. Something changes. This is the starting gun. A plant is announced, a customer is won, a rule changes, a competitor shuts down, a debt gets repaid.

2. Capacity and operating leverage. The company now has more ability to produce than it is using. This is the coiled spring. A factory that can make 100 units but is making 40 has 60 units of profit waiting inside it, because the rent and the staff are already paid for.

3. Revenue acceleration. Sales start growing faster than they were growing before. Note the word faster. A company growing 15% every year forever is not accelerating. A company that went 8%, then 12%, then 19% is.

4. Margin expansion. The company starts keeping more of each rupee it sells. This usually follows step 2, because the fixed costs were already being paid.

5. Profit acceleration. Now the bottom line jumps. This is the part everybody can see.

6. Re-rating. The market notices, decides this is a better business than it thought, and agrees to pay a higher price for the same rupee of profit.

Now the important bit. By the time you reach step 5, it is too late to be an edge. When profit explodes, it is in the newspapers, it is in every screener, and everybody knows. The value of this scan is entirely in steps 1 to 3, where the reason is visible but the profit has not arrived.

That is why “profit up 200%” on its own is worth nothing to us, and why a company can be an excellent candidate with completely ordinary current numbers.

The scan does not just guess which step a company is on. It marks each of the six steps with one of four labels:

“Confirmed” is a high bar on purpose. Money actually spent. Capacity actually commissioned. Orders actually booked. A promise is not a confirmation.

Every company on the path is put into one of three stages. This is the single most useful thing in the scan, so read it slowly.

Stage 1: Base building. The plant is being built. The order book is filling. The debt is being repaid. But sales still look completely normal, so the market ignores the company entirely. This is not a rejection. Some of the best entries in the world live here. It is just early, and early is uncomfortable.

Stage 2: Inflection. The corner. Growth starts improving, the new capacity starts filling up, and management changes how it talks. The tell is small but reliable: they stop saying “we expect” and start saying “we are seeing”.

Stage 3: Acceleration. The engine is running. Sales grow 20 to 30%, profit grows 40 to 70%, returns improve and debt falls. There is a specific fingerprint that proves this is real operating leverage rather than just a good year:

Profit growth is faster than EBITDA growth, which is faster than sales growth.

EBITDA, if you have not met the word before, is what the business earns from actually operating, before interest, tax and the accounting charge for wear and tear. It sits between sales at the top and profit at the bottom.

So that sentence just means: sales grew, the operating earnings grew faster than sales, and the final profit grew faster still. When those three line up in that order, the extra sales really are falling through to the bottom line instead of being eaten on the way down.

Now the counter-intuitive part, and the thing most people get backwards.

Stage 3 is not the best stage. Stage 2 is.

Acceleration sounds like the exciting one, and it is, but by then the move has largely happened and the price usually reflects it. A company in acceleration is only still interesting if the price has not yet caught up. Inflection is where the evidence is strong enough to trust and the price has not fully reacted. That is the sweet spot.

If a company fits none of the three, it is marked off-path and it does not appear.

Behind every company in this scan, seven questions get answered first. The score you see is only a summary. These questions are the actual analysis.

1. What changed? If nothing changed, there is no J-curve, however wonderful the business is. A great company that is doing exactly what it did last year is a great company, not a J-curve.

2. Is the market or sector growing? It is much easier to grow when the whole industry is growing. Swimming with the current beats swimming against it.

3. Has the company invested ahead of demand? Did it build the factory before it had the orders? This is the brave, expensive decision that creates the flat part of the J. It is also the thing that makes the steep part possible.

4. When does the new capacity actually start contributing? Not “soon”. A quarter. A month. A date. This question is what turns a nice story into something you can check.

5. Can revenue grow more than 20% for two to three years? One good year is not a J-curve. The shape needs a run of them.

6. Can margins improve at the same time? Growing sales while margins shrink is just buying revenue. The J needs both moving the right way together.

7. Is the price still leaving room? If everything above is true but the market already worked it out and priced it in, the opportunity has passed. This question is asked honestly, and the answer is reported as an observation. The scan never tells you to buy or sell anything.

Because a big profit jump is exciting, it is worth being suspicious of one. Before any earnings jump is believed, the scan checks where it actually came from:

• A terrible base, so this year looks huge only because last year was awful

• An inventory gain

• Exceptional or one-off income

• A currency gain

• A tax reversal

• A single one-off order that will not repeat

• Other income, which means money not made from the actual business

• A temporary commodity price benefit

The test is simple: if you strip the one-off out, is the core business still accelerating? If the answer is no, the check is marked unclean and the company does not make the list.

The reverse is also true and matters just as much. A company can be firmly on the path while printing an ugly quarter. A real factory ramp underneath a messy result is completely normal. The scan separates the two rather than reacting to the headline number.

Each company is scored out of nine. One point each, no half points, and a point is only given when something specific can be pointed at.

The scan currently covers around 140 companies from a universe of nearly 5,000 companies. You can open the thesis for any company on the list and understand why we think it is approaching an inflection point.

This is not a one off scan. Companies will enter and leave as their thesis, data, or business conditions change. We have also added a return-since-entry column so you can track how the companies perform after they enter the scan.

The future of scanning, in our view, is not just screening for one number or one technical condition. It is using the rich data available on each company to understand what is changing in the business, where the next inflection could come from, and whether the company is ready for stronger earnings.

That is what we are trying to build with the Potential J Curve Setups scan.

Investorstack