According to Isaac Asimov, the most important phrase in science is not ‘Eureka’ but ‘That’s Odd’. He was alluding to the fact that, in research, observations of the unexpected often lead to important discoveries; Fleming’s observation of a zone of inhibition around penicillin in a petri dish is the most famous example
of this phenomenon. And Asimov’s remark came to mind recently as I was studying the structure of the largest medtech firms, although the observations seem to apply to many large pharma companies too. The most remarkable thing about these firms – Medtronic, J&J, Becton Dickinson, Stryker and others – is that they are all very strongly divisionalised, to the point of being more like medtech conglomerates of very different businesses than simply units within the same company. That’s very odd, for reasons I’ll explain with a diversion into the theory before I come back to what practically useful lessons we can draw.
If nature abhors a vacuum, traditional theory has it that investors hate conglomerates; markets tend to value aggregates of different businesses at less than the sum of their parts, which leads activist investors to call for their break-up or to demand spin-offs. The reasons for this persistent and well-recognised hatred are complicated. Mostly, they stem from the belief that any synergies between the diverse businesses are worth less than the negative implications of conglomeration. Investors claim that the latter arise because boards can’t fully understand, and therefore run, a set of dissimilar businesses. Unless there is some exceptional reason, therefore, conglomerates should dissolve into purer, more focused businesses.
And yet, in most of the larger life sciences firms, we see a group of very diverse businesses operating in an essentially autonomous way under a larger corporate umbrella, as in the above examples. This is odd, in the Asimov sense. And what makes it more so is that, as each business evolves, it often becomes more diverse, making the traditional argument for break-up even stronger. As Fleming must have asked: What’s going on here?
To quote Theodosius Dobzhansky, nothing makes sense, except in the light of evolution. The persistence and growth of the conglomerate structure, housing multiple very different businesses, predicts that
these assemblies must have some adaptive advantage compared to purer alternative models. And that advantage must outweigh any disadvantages of being in a conglomerate managed by a board that can’t possibly understand all its varied subsidiaries.
But what could those advantages be? As I researched this question, my first guess was – not usually – wrong. I thought that conglomerates selling very different products but to the same customer might bundle their offerings to create an advantage. But this doesn’t seem to happen much. It gets too messy and, generally, the separate business units operate with a high degree of independence.
The independence of business units leads to lots of different approaches to solving similar management problems in different markets
My second guess was also wrong, but slightly less so. I thought that there might be synergies created by the internal transfer of technologies from one business to another. I saw some hints of that, but again not enough to explain the phenomenon. Technologies differ too much and transfer is just too difficult, generally.
But the failures of my first two hypotheses do seem to be helping me in the direction of a better explanation for conglomerates in the life sciences sector. The independence of business units, revealed by testing my first hypothesis, leads to the generation of lots of different approaches to solving similar management problems in different markets. For example, how to justify expensive, innovative products, or how to understand the customers’ value chain or how to segment complex, multi- stakeholder markets. And the difficulty of transferring scientific technology, revealed by the testing of my second hypothesis, coincidentally revealed that it is much easier to transfer management technology, namely the processes for solving difficult business problems like those mentioned above. Management techniques flow between diverse subsidiaries more easily than scientific techniques do, it seems.
So, my current working hypothesis is that life sciences conglomerates thrive when, firstly, their diversity generates new management techniques and, secondly, when they have the capability to transfer those management techniques between subsidiaries. But this implies something else that is important about such conglomerates. Both generation and transfer of ideas imply interesting capabilities. The first suggests that the leadership of conglomerates can let their subsidiaries be independent enough to experiment but controlled enough to not run wild. The second implies a capability for identifying management technologies that might have wider application. As readers of my work will recall, those leadership capabilities must be the expression of organisational routines, the firm’s equivalent of genes. My work is hinting at what those routine are – for example, the routine for switching leaders between subsidiaries and for identifying shared issues – but I think I’ve got a lot more to uncover.
Read a textbook and it will likely tell you that conglomerates shouldn’t exist. Watch the life sciences industry and you will see that they thrive. That’s odd, as Asimov would say. That’s evolution, as Darwin would say.