Given the importance of transactions in business, companies would do well to learn from their successes and failures. Yet as simple as it may seem, learning from experience is not easy. Many companies seem to miss the lessons of past experience entirely, not only failing to learn from their own history but also ignoring the cautionary tales of competitors.
This article will look at how companies can learn from experience by focusing on three issues: clearly defining success and failure; avoiding judgmental biases; and guarding against the Halo Effect.
Defining success and failure
The first obstacle has to do with the definition of success and failure. Simply put, there is no common definition of success or failure. Rather, companies appear to use a number of difference definitions, which lead to very different conclusions.
Consider, for example, the abundance of studies about mergers and acquisitions. Many of them claim that most acquisitions fail, and this idea has become a mantra across the business world.
The exact rate of failure varies from study to study, ranging from a low of 60 per cent to as high as 75 per cent. Despite these warnings, hundreds of companies undertake acquisitions every year. Are these businesses foolish, or merely affected by hubris? Perhaps some are, but a more likely explanation concerns the various definitions of success and failure.
■ Does failure mean the acquiring company did not meet its original goals? Given how ambitious many acquisitions are, and the optimism that prevails at the time they are announced, according to this definition, it is not surprising that a majority are considered failures. Few acquisitions meet the rosy expectations that are stated when the deal is announced. But only the most severe definition could conclude an acquisition that did not meet its original goals should be called a failure.
■ Does failure mean the deal destroyed value, as measured by a decline in the share price? Many empirical studies show that the acquiring company suffers an immediate decline in share price, reflecting the capital market’s expectation that the purchase price is likely to be greater than the value that stands to be created. This may seem like a better definition because it is based on objective financial data, but this standard involves a broader set of variables, too.
First, over what length of time should we study share price movement in order to capture the true impact of an acquisition? Some acquisitions may initially be greeted negatively by the market, but may be found to add value over time. Second, what might have happened to the company’s market value had the deal not taken place? Sometimes, companies faced with an eroding market position are motivated to acquire because their situation would worsen in the absence of a merger. For example, Hewlett-Packard clearly did not capture anticipated value through its 2002 acquisition of Compaq, but it does not necessarily follow that HP would have been better off had it remained in its previous position.
■ Does failure mean the company would not make the deal if it had the chance to do it all over again? I have not seen a study of acquisitions that defines failure in quite this way, but my guess is that the fraction of acquisitions that would be considered this type of failure is less than half. Although the acquiring company’s share price may have dipped, it is not clear that more than half of acquisitions are considered a mistake. It is possible, however, that more than half of acquiring companies would still undertake the deal if they had the chance to do it all over again, which suggests, they should not be labelled as failures.
The lesson of these competing and plausible definitions is that, in order to learn from success or failure in transactions, managers need to have a clear definition of these terms – and be sceptical of studies that do not.
The delusion of control
Even if we have clear definitions, we might still resist learning from past results because of our judgmental biases. Research by cognitive psychologists has identified a number of biases that shape human thinking, several of which are particularly relevant to the measuring the performance of transactions.
One bias is overconfidence. We tend to see ourselves as above average and frequently assume too much about our knowledge and capabilities. One result is that many people disregard base rates, the prior experience of a relevant population. As Dan Lovallo of the University of Western Australia and Daniel Kahneman of Princeton have written, managers often believe their chances of success are not bound to the same probabilities and frequencies that have shaped prior actions. They see themselves as immune to the fate that has affected others. They may believe themselves to be more skilled, determined or capable than the norm and, hence, not subject to base rates. To some degree, of course, optimism is healthy and should be encouraged, and it may well be that the present situation is different and not bound by past odds. But such reasoning demands a sober explanation and ought not to be based on bravado.
The lesson for managers is to pay attention to base rates of relevant populations – previous transactions of a similar kind – and to temper natural optimism with past experience.
The Halo Effect
Even if managers follow the steps outlined above, they may face another problem when trying to learn from experience: the Halo Effect, the tendency to make specific inferences on the basis of a general impression. For example, if we believe a company to be successful, perhaps because its sales and profits are growing rapidly, it is natural to infer that it has a sound strategy, a visionary leader, capable employees, an efficient organisation, strong customer orientation and so on.
By the same token, when we believe a company is a poor performer, with slumping sales and shrinking profits, we are quick to infer that its strategy was wrong, its people were complacent, its organisation was inefficient and it ignored its customers. In fact, many everyday concepts in business, such as leadership, culture and organisation, are fuzzy concepts, neither explicitly defined nor objectively measured. As a result, many of the factors commonly held to determine company performance are actually attributions based on performance.
The Halo Effect undermines many studies about business performance and complicates our ability to learn from transactions. For instance, if we know that an investment or an IPO was successful, we may find that participants often make positive attributions about the leadership and execution involved. They may explain their success in terms of a clear vision, focused execution, strong values, perseverance through adversity, or simply patience and courage.
By contrast, if they know that the transaction turned out badly, they may explain that the decision was poorly implemented, that leadership was ineffective, that a poor process was followed. It is possible that these explanations are valid, but unless we have a way of measuring such qualities as leadership, execution and perseverance independent of the outcome, we really don’t have valid data.
Consider the explanations that are commonly given about success or failure in M&As. A common refrain when it comes to failed mergers is poor leadership. For example, look at the blame heaped on the executives of the AOL-Time Warner merger or on Jürgen Schremp, CEO of Daimler-Benz, in the wake of the company’s acquisition of Chrysler. It may be true that these executives could have done a better job in leading the acquisitions, but such a judgment must be made independent of outcome because we can always pick an example of failure and infer that the leader was somehow to blame.
Unless we have a way of measuring such things as leadership, execution or persistence in ways that are independent of outcome, these are more likely to be attributions made on the basis of performance than causes of performance. Many studies of business transactions suffer from this flaw in logic. They begin by identifying successful or failed transactions, and then try to identify their common features by looking backwards, often interviewing the managers involved.
Some common explanations for transaction success or failure suffer from this flaw. The resulting findings will seem reasonable – they will appeal to our common sense – but may divert our attention from factors that have a more significant impact on successful transactions.
Much of business can be understood as a series of transactions. By their nature, transactions lend themselves to study and offer the promise of learning from past actions and improving chances for the future. Some companies capture the benefits of learning, but many do not.
The best ones do three things. First, they define success and failure clearly, neither accepting definitions without examining them closely nor altering a definition to achieve a specific, desired result. Second, they pay attention to their past experience and the experience of others, and temper their natural optimism with a recognition of base rate probabilities. Third, they guard against the Halo Effect and try to assess the drivers of success or failure objectively, rather than relying on the easy explanations we often make when we know that a transaction turned out well or poorly.
Phil Rosenzweig is professor of strategy and management at IMD. He is the author of “The Halo Effect… and the Eight Other Business Delusions that Deceive Managers” (Free Press, 2007).