Is the growth-share matrix (aka BCG matrix) useful?

The growth-share matrix was created by BCG, the consultancy, in the late 1960s based on work with a paper mill. Paper manufacturing is dependent on economies of scale and economies of learning, so market share went on the x-axis. Companies want to grow, so growth went on the y-axis. This model turned out to be immensely successful for BCG and it was implemented by many companies in the 1970s. 

The model

Here is one of the first writings about the model. The key message was that companies needed a balanced portfolio of products with different growth rate and market share. No product will remain a cash cow forever, so it was important to also invest in new products. A virtuous circle would occur when cash from cash cows was invested in questions marks, which were grown into stars. Note that BCG's early writing talked about products and not business units. However, the model quickly became a corporate strategy model, in which business units were plotted and not products.

Bruce Henderson, CEO BCG, presenting the growth-share matrix in the late 1960. He has began to draw what is called the success sequence.  Cash from the cash cow is spent on the question marks, which then grows into stars. The arrow left to be drawn moves from the star to the cash cow. Source: BCG.

In the 1970s, there was a stronger focus on manufacturing and no emerging markets. The model was a product of its time, i.e. economies of scale was still the main source of competitive advantage (at least in theoretical reasoning). The low growth environment created a willing audience of corporate executives. The model turned out to be unimaginatively successful. Management consulting became a growth industry. Clients had never seen a matrix before in their lives, so the simple two-by-two became a revelation. Unfortunately, executives also became deluded by the normative advice implicit in the model, e.g. find promising question marks, do the hard work by investing in building market share, divest dogs, then finally, relax, smoke a cigar and watch your stars mature into cash cows.

BCG had realised the problem with the growth-share matrix. In 1981, they introduced the advantage matrix. In this new matrix, the old growth-share matrix was relegated to one of the quadrants, labelled Volume. The new matrix was hard to understand and was not used, at least not outside BCG. BCG's style in the 1970s was academic so credit should be given to the consultants for realising and writing about the problems associated with the growth-share matrix.

Unfortunately, in management, there are often fads and certain models take on a life of their own. BCG's internal writings on the growth-share model were not widely available at the time. (Today you can find a selection on the BCG website.) Instead, the model entered strategic management textbooks because it was used so extensively by large companies. Most of the details were lost. Smaller consultancies continued to use the model well into the 1990s. Currently, the model is described on hundreds of website.

Problems

Here are four central problems with the model. The problems are so severe that one probably should not use the growth-share matrix at all anymore.

1. The model assumes that the only source of competitive advantage is having a high market share. In fact, it is not sufficient to have a large market share, it has to be the highest market share (i.e. relative market share above 1). We know today that market share is not the only source of competitive advantage. Another source is uniqueness. If you are unique you do not need to be the largest; you can even be successful as a  niche player.
2. The model assumes that growing market share is necessary. However, growing market share can be costly, e.g. lowering prices or by buying a competitors. So even in industries in which market share leads to competitive advantage, the model will push some companies into making dangerous choices regarding market share.
3. The model assumes that there are no synergies between products. In reality, there are often linkages between products. Products might both share production capacity and removing one product might not result in correspondingly large cost savings. In addition, there might be synergies between upstream and downstream companies in the industry supply chain (e.g. forest plantation upstream and pulp and saw mills downstream).
4. The model assumes that cash cows are rather stable. The model correctly assumes that cash cows do not last forever. However, it erroneously assumes that the decline of the cash cows will be gradual. Disruptive innovations or other immediate threats to to the cash cow do not exist. In reality, it might make sense to invest further resources in strengthening the cash cow instead of investing in question marks.

There are additional, but less severe, problems. The model assume that companies should be cash neutral. However, it is possible for companies to borrow funds. The model assumes that companies should not try to grow market share in low growth industries, but instead divest the dogs.

It can be interesting to plot a company's different business units solely to get a visual representation of the company, but all normative implications of the model should be discarded.