The Growth Share Matrix 2026: What BCG Measured Later

The four rules under the BCG matrix, written in 1970

The original essay is called "The Product Portfolio", signed by Bruce Henderson, dated 1 January 1970 by the firm, checked 7 October 2026 at bcg.com. It opens on the purpose: "To be successful, a company should have a portfolio of products with different growth rates and different market shares. The portfolio composition is a function of the balance between cash flows."

Then it gives four rules, and the first one carries the whole edifice: "Margins and cash generated are a function of market share. High margins and high market share go together. This is a matter of common observation, explained by the experience curve effect."

Read that qualifier again. The author calls his own founding rule a matter of common observation, and offers the experience curve as the explanation rather than as evidence. That is an honest way to state an assumption, and it is the sentence most summaries of this framework leave out.

Two details from the same text are worth keeping. The fourth quadrant is named in the essay: "Products with low market share and slow growth are “pets.”" The word dogs appears nowhere in it. And no threshold appears either, for growth or for share. The grid as published has four boxes and no dividing lines, which means any BCG matrix chart or diagram you see with lines on it carries thresholds somebody else chose. A blank BCG matrix sample, four quadrants with unlabeled axes, is closer to the 1970 document than any ruled version of it.

The four BCG matrix quadrants, from cash cow to pet

The essay names the four boxes and defines each one by its cash behavior rather than by its place on a chart. The four definitions below are quoted from that same 1970 text, read again on 7 October 2026.

A cash cow holds high share in a slow market. "Characteristically, they generate large amounts of cash, in excess of the reinvestment required to maintain share. This excess need not, and should not, be reinvested in those products." The essay attaches a condition most summaries drop: "if the rate of return exceeds the growth rate, the cash cannot be reinvested indefinitely, except by depressing returns."

A star holds high share in a fast market. "It nearly always shows reported profits, but it may or may not generate all of its own cash." Its destination is stated outright: "The star eventually becomes the cash cow, providing high volume, high margin, high stability, security, and cash throwoff for reinvestment elsewhere."

A question mark holds low share in a fast market, and the essay is blunt about the odds. "They almost always require far more cash than they can generate. If cash is not supplied, they fall behind and die." Holding share wins nothing: "Even when the cash is supplied, if they only hold their share, they are still pets when the growth stops."

A pet holds low share in a slow market. "They may show an accounting profit, but the profit must be reinvested to maintain share, leaving no cash throwoff. The product is essentially worthless, except in liquidation."

What the BCG matrix evaluates, then, is cash behavior rather than product quality. The four labels are a cash forecast wearing a position, which is why the text says "All products eventually become either cash cows or pets." Putting a unit in a quadrant is a claim about the cash it will throw off, and that claim is the one to say out loud and defend.

Two farms as a BCG matrix example, same margin, different share

Oliver Wyman publishes a case in which a private equity client picks between the only two dairy farms in an Illinois county, checked 30 September 2026 at oliverwyman.com. The figures are all in the text, so the arithmetic is checkable.

By hand. Farm one: 10,000 cows at 100 units is 1,000,000 units; at 20 dollars that is 20,000,000 dollars of revenue; unit margin is 20 minus 10, so 10 dollars, giving 10,000,000 dollars of contribution; less 5,000,000 of fixed cost leaves 5,000,000 dollars of profit, which is 25 percent of 20,000,000. Farm two: 5,000 cows at 80 units is 400,000 units; at 30 dollars that is 12,000,000 dollars; unit margin is 30 minus 15, so 15 dollars, giving 6,000,000 dollars of contribution; less 3,000,000 leaves 3,000,000 dollars, which is 25 percent of 12,000,000.

The firm's own published answer says the same thing: "Farm one and farm two have the same profit margin. On absolute dollar terms, farm one has $2 million in profit more than farm two." Both statements check out, and 5 million less 3 million is indeed 2 million.

So here is a published case, from a firm, in which the volume leader holds two and a half times the volume of its only rival and earns exactly the same margin. The first of Henderson's four rules does not hold in it.

Share of what, exactly

The horizontal axis needs a share figure, and this case shows how much the choice of base moves it.

Each figure by hand. On cows, 10,000 over 15,000 is 66.7 percent, rounded to one decimal. On volume, 1,000,000 over 1,400,000 is 71.4 percent. On revenue, 20 over 32 is 62.5 percent. Relative to the rival, the three ratios are 2.00, 2.50 and 1.67.

The firm's published answer states that farm one has "more than 70% of the market share in terms of the number of cows and milk production volume". On volume that holds at 71.4 percent. On cows the same three data points give 66.7 percent. We publish both readings and leave the sentence as the firm wrote it, because the arithmetic is what a candidate can check and the wording is theirs.

The practical lesson survives whichever base you prefer. A single unstated base moves the leader's relative position from 1.67 to 2.50, which on most drawn versions of the grid is the difference between a comfortable position and a dominant one. Name the base out loud before you place anything.

The 2014 study that revisited the BCG matrix

In June 2014 the firm's Strategy Institute published a review of its own matrix, checked 30 September 2026 at bcg.com. The method is stated in its notes: all publicly listed US companies from 1980 through 2012 from Compustat, with relative market share defined as a company's share divided by the share of the industry's third ranked company, and average time in a quadrant computed for 1988 to 1992 and again for 2008 to 2012.

Four figures come out of it, and all four are the firm's own.

1. Circulation doubled in speed. "This was true in 75 percent of industries, reflecting the higher rate of change in business overall. In those industries, the average time spent in a quadrant halved: from four years in 1992 to less than two years in 2012."
2. Cash cows got smaller. "in our analysis of public companies, the share of total profits captured by cash cows in 2012 was 25 percent lower than it was in 1982".
3. The mature phase shortened. "the duration of that later part of the life cycle declined as well, on average by 55 percent in those industries that witnessed faster matrix circulation".
4. Pets multiplied. "our analysis found that the number of pets increased by almost 50 percent in 30 years".

Then the conclusion about the grid itself: "The matrix also requires a new measure of competitiveness to replace its horizontal axis now that market share is no longer a strong predictor of performance."

The firm keeps the matrix and asks for the bottom axis to be replaced, without naming the replacement. That is the state of the art according to its author, and our research file records that no substitute measure is published.

The funding conversation this grid still wins

One thing the grid does that a profit tree does not. It forces a company with several units to say which unit is paying for which, rather than judging each unit as if it stood alone. The 1970 essay is explicit that this is the purpose, and the balanced portfolio it describes has stars, cash cows and question marks in it.

McKinsey's nine box model was built for the same job in the same decade and says something useful about how to use either one. Its page describes judging a unit on "the attractiveness of the relevant industry and the unit’s competitive strength within that industry", then adds: "Sorting units into these three categories is an essential starting point for the analysis, but judgment is required to weigh the trade-offs involved", checked 30 September 2026 at mckinsey.com. A sort is a starting point. The trade off is the work.

In a case interview that translates into one question and one number. Ask which units fund which, then compute the cash each throws off and each consumes. For the two farms above, the answer is that neither funds the other and the client owns neither yet, so the grid has nothing to allocate and the question is a straight investment comparison. The grid's silence there is informative.

Three related pages. The growth question this grid is routinely confused with is covered in Ansoff against the growth share matrix. The composite score behind the nine box model is covered in market attractiveness. And the assumption in Henderson's first rule has a page of its own in scale against profitability.

What it costs on a scored structure

The weights below are ours, read from the rubric code in src/lib/grading. No consulting firm publishes a rubric of its own, which is one of our steadiest findings.