I recall a seminar at London Business School a few years back where I dissected a crisis of sales and profit decline at Marks & Spencer (M&S), the UK’s beloved retail giant. This iconic company had recently experienced a golden era of robust profit growth, but then found itself in a troubling predicament marked by an alarming fall in sales and profits.
I showed how the dwindling sales and profits could be traced to three pivotal factors: a slow-down in customer growth, the loss of long-standing loyal customers, and falling sales to the customers who remained.
But how had this come about?
The ‘Slash and Burn’ Approach. It became evident that the very tactics that had propelled the company’s short-lived success had sowed the seeds of its recent decline.
The Board had appointed a rock-star CEO. He had adopted a two-part strategy. First, launch lots of new products aimed at winning younger customers. Yes, that brought in new customers, but [1] those customers were fickle and soon bored of the me-too products that other stores did better, and [2] it turned off long-term customers, many of whom had shopped at M&S for decades.
So sales, revenue and gross profit soared, then stalled, then fell.
Then, to boost operating profit, our super-star turned to a “slash and burn” approach to cost control. But no amount of cost-cutting could make up for that loss of sales.
A Simple Model Told the Story. To help my audience grasp the issue, I showed a simple quantified working model of the main elements at play – the growth and retention of customers, both young and old, their purchase rates, unit-prices, margins and operating costs.
This could not be a full “digital-twin” model, because the company (like most others) didn’t publish details about changes in its customer base and purchasing behaviour. Nevertheless, plausible assumptions, backed by stock-analyst’s investigations gave a persuasive picture of how the falling operating profit had come about.
The Professor’s Challenge. I opened the floor to questions, eager to engage in a fruitful discussion about how the M&S business system worked and had been broken by the wrecking CEO..
But the very first comment from a senior professor floored me completely. “That’s interesting, Kim, but there are a million possible reasons why the company’s profits might have fallen,”
Now consider how most management research gets done.
You have an indicator – the dependent variable – you want to explain; often that is company profitability. Then you have a hypothesis regarding how some range of plausible explanatory factors – the independent variables – might explain the values of that dependent variable. You then put data from a large sample of cases through statistical analysis, to see which of the independent variables most strongly “explain” the dependent variable.
You have to set out your reasons for choosing those hypothetical drivers. BUT … there is no requirement to prove the pathways of causal connections back from the outcome to those hypothetical causes. So, thousands of researchers test hypotheses for what drives business results, competing to see who can get the best statistical fit with any of the many possible explanations. (To be fair, the same dysfunction applies to most research in the social sciences.)
I should have countered the Professor’s challenge with:
“No, there are not a million reasons for a fall in operating profit – there is just ONE reason … “
That one reason? “… The decline in operating profit is due to the net change in operating costs exceeding the net change in gross profit.“
I should then have gone on with “And there are only two possible explanations for the decline in gross profit – a decline in revenue and/or in gross margin on that revenue”.
We know there was a big fall in revenue so … “There are only two possible explanations for a fall in revenue – a fall in customer numbers and/or in their spending rate“. Both were known to be true.
Then … “There is only one possible explanation for a decline in customer numbers – more customers are being lost than are being won.” And we know there were problems with both the win-rate and loss-rate.
From there we go on to codify the causal relationships described in the story above.
Lessons for Strategic Management. The M&S saga demonstrates the importance of a robust approach to sustainable growth and profitability.
The story shows the perils of fixating [1] on sales growth, and not the quality of that growth and [2] on mindless cost reduction as the fix for profit growth. As more thoughtful CEOs have said “You can’t cut your way to success“.
The CEO who followed on from the wrecker had a really tough job, but gradually pulled the business around by rebuilding its reputation with its core customers, and gradually extending its appeal to new segments.
The story is also a warning about a certain breed of management – the quick-fix merchants who drive impressive short-term results, only to jump ship to the next victim, while leaving others to fix the damage they did. (Watch out for these folk inside companies too – be wary of any job offer to succeed any implausibly successful predecessor!)
Questions for Management Research. Management Academics; especially in Strategic Management; have for ever lamented the poor uptake of the findings that emerge from their research. But maybe there’s a reason for that. Their findings are not used because they are not useful or reliable (See my post on Why Theory Matters – if it’s General Useful and True).
Every new researcher is warned that “Correlation does not prove causation“. But then they are thrown into a system where everyone tries to do just that!
The causal pathways that make up the business systems we create are not especially complex, and are amenable to being visualised, codified, quantified and validated. Statistical coincidences are no substitute for a functioning model of how the business system actually works and performs – a digital twin. Indeed, for reasons I may try to explain another time, it is meaningless to use correlation to confirm causal explanations for accumulating factors (like customer numbers). It is therefore equally meaningless to use correlation to confirm causal explanations for anything that depends on accumulating factors – like sales and profits!
The method that gave this solid,, quantified explanation for the problems at MS is rigorous, yet do-able by anyone with reasonable analysis skills. See how it works, how it is done, and the power of its results at the Digital Twin Business Model Workshop.