Cows & Dogs

Introducing the BCG matrix

Friends,

Much like Keanu Reeves, I could probably spend the rest of my career doing matrices. In the world of strategic management and marketing consulting, they have long since become the epitome of commodities. Most everyone has their favorite, either of own or others’ design.

Having said that, most are far from original and few hold up to scrutiny. Even those considered most tried and tested, such as the Ansoff grid that we covered last week, have easily demonstrated weaknesses. And yet, although perhaps not to the extreme extent of the 1980s, a quick look at Google search trends show that their popularity remains stable with (as expected) repeated peaks during the annual autumn planning season.

The reason, I would suspect, is that matrices provide a sense of simplicity in a complex and frequently difficult-to-grasp corporate reality. Out of a myriad of various conditions that could have an impact on strategy, they focus on a handful that (in theory) are deemed to be of particular importance. Put differently, a matrix limits data by prioritizing specific information, much in the same way that certain KPIs do. To managers looking for clarity in uncertainty – and consultants looking to sell services to said managers – the perceived value of a matrix should therefore hardly come as a surprise.

Today’s object of analysis originated from precisely such a relationship. The Boston Consulting Group’s growth-share matrix (most commonly labelled the BCG matrix, though it is a child of many names) was first created by Alan Zakon during an acquisition strategy project for the Mead Paper Corporation in the late 1960s. After a lunch conversation with a colleague specializing in finance, he decided to take inspiration from investment portfolio planning in order to ‘sort out the losers’ from Mead’s six product groups and 45 operating divisions. As it happened, his approach turned out to be so successful that he was asked to present it across BCG. With the help of founder Bruce Henderson and other members of staff, the matrix eventually turned into the 2x2 with which many will be familiar.

The four quadrants of the BCG matrix represent specific combinations of market growth and (relative) market share.

• Low market growth, high market share is intended to equal a low-cost position with little need to add additional funds for, for example, product innovation. The market should be mature, company performance predictably stable, and one can ‘milk the cash cow’ for reinvestment elsewhere.

• High market growth, high market share means a star company that carries a significant return potential. However, it will also require an equally significant investment in order to maintain share or, even better, improve it further before the market matures and the company moves into the cash cow quadrant.

• High market growth, low market share indicates that there is a significant need for financial aid. Given the uncertainty that comes with the position, the strategist has to quickly figure out whether to invest or divest. A portfolio can only carry so many question marks; if a company fails to maintain growth, years of cash consumption can cause all sorts of future problems due to negative compounding.

• Low market growth, low market share, lastly, is where one typically finds what one might call pet companies; fun to have, perhaps, but no more than a modest contributor to (or drain on) overall funds. These should be either liquidated, divested or repositioned unless there are strong strategic reasons for not doing so (e.g., interdependencies with other business units).

From the reasoning above follows a potential redrawing of the matrix based on optimum cash flow as seen through the wider corporate lens:

The influences from stockbroking and investment management should be evident to anyone with corporate finance training. At the BCG grid’s core is a differentiated holding of interests in businesses with varying degrees of risk and opportunity, developed to determine investment requirements and cash flows for an overall portfolio balance.

Those of you with a background in product development or marketing may also observe that the matrix is rather reminiscent of Product Life Cycle (PLC) theory. During a market introduction, costs are often high, sales volumes are low, and profits are scarce; most would agree that it is a question mark stage. As the company grows, costs are reduced (often due to economies of scale) and sales volumes are increased, but so too is the intensity of the competition. If one is lucky, the stars might align. Eventually, the market matures. Costs are lowered further, production volumes increase, and it is time to milk the business for all that it is worth before the vertical saturates, profits drop, and the overall portfolio contribution diminishes along with them.

Personally, however, I would be careful not to take this too far. Although the two concepts may appear similar, the BCG grid is based on broader asset allocation while PLC is about individual products and their marketing strategies; each has its own unique contextual requirements, and product lifecycles are (as a rule) much shorter than business lifecycles.

I digress.

Despite the undeniable historical success of the growth-share matrix (having been around for more than 50 years), it would be impossible to ignore that it has, so to speak, moved into the dogs quadrant. The portfolio analogy of the firm is largely a thing of the past, and while Zakon and Henderson’s creation was taught in practically all business schools a couple of decades ago, the same is not true today (at least in Western markets; emergent market training appears to be slightly lagging)

Disregarding for a moment the necessity of reading broadly, I do not think this is particularly troublesome.

Underlying the structure of the matrix is an explicitly assumed link between market share, cash generation and profitability. Although this supposedly positive correlation may come across as familiar – no doubt due to the fact that it is a claim frequently made in textbooks and courses – competitor-orientated objectives have repeatedly been proven to actually be negatively correlated with both profits and ROI. As we have discussed before and I detailed in Strategy in Polemy, obsessing over rivals is not necessarily the best way to do business; neither striving for market share supremacy nor being the market leader is a guarantee for profitability. While share gains can bring a myriad of benefits, there exists no causal mechanism whereby they automatically translate to profit.

To make matters worse, Armstrong and Brodie found that use of the BCG matrix as a decision-making aid substantially reduced the profitability of subjects’ decisions. Slater and Zwirlein concluded, from a study of 129 firms, that those whose strategies were consistent with portfolio planning models had lower returns to shareholders. This echoed discoveries made by Capon, Farley and Hulbert; companies that used the matrix reported a lower return on capital.

Of course, we cannot know how those companies used the matrix. Often, tools such as BCG’s famous grid are taken to be more than they are, either because those selling them make overly grandiose claims (Henderson famously stated that his chart, with a projected position for five years out, would be ‘sufficient alone to tell a company's profitability, debt capacity, growth potential and competitive strength’, for example) or because those using them are unaware of any limitations. But what we do know is that the matrix is inherently reductionistic, which means that crucial context inevitably will be lost.

Long-time readers will know why this is. Markets and firms are complex adaptive systems and, as such, inherently cannot be reduced. Whether it is a 2x2 such as the BCG matrix, McKinsey’s 3x3 version, ADL’s 4x5 multifactor model or something altogether different is irrelevant – it is impossible to cover everything.

To be fair to BCG, this is an inescapable problem with matrices of all kinds, and always leads to the same justified criticisms (too narrow, mechanistic, subjective etc.). But, as alluded to before, the false sense of simplicity is also why they appear so alluring in the first place; they allow users the perceived luxury of being able to concentrate on collecting a limited amount of very specific information.

Ultimately, what one does with said information determines how valuable the exercise is. Take it for what it is – a flawed version of reality – and it might yet provide something of worth. Take it as gospel and the results are likely to be damaging. As Zakon himself put it, the BCG matrix can be ‘a helpful tool, but it can also be misleading, or worse, a straitjacket’.

How much it helps today is up for debate, but I would not send a person who brings it up to the asylum. At least not until they have explained how they intend to use it, and how said use helps the organization reach its shared understanding of success. Of course, that means that the strategic ambition would already have been set, much as how the various companies in the corporate strategy would already have to exist.

It is for this reason that the matrix failed to make my original list. As likely or unlikely as it is to come into play, it will ever only do so later in the process.

Next week, we are going to do a special Christmas edition of the newsletter, followed by a recap of the year that has been. Once firmly in 2022, we will continue our discussions of strategic concepts with the notion of core competence, McKinsey’s 7S model and much, much more.

Until then, have a lovely weekend.

Onwards and upwards,
JP