Everything is Obvious 7/10

Parts 4, 5, 6.

7/ The Best-Laid Plans

  • There are two kinds of events that arise in complex social systems: the ones conforming to some stable historical pattern (we can make predictions about them), and the ones that do not (we can’t make predictions).

  • Understanding the pattern requires collecting a history of prior outcomes (the more – the better) and as a result gives a probability for a future outcome. This is where credit scoring and internet targeting kick in. [MK: this also explains that it’s incredibly hard to predict someone’s specific travel behaviour as the frequency of transactions is low and there’s not enough data to be meaningful.]

  • Some decisions have to be made based on the predictions of what will happen in 1-2 years (strategic plans, book publishing, etc.), and as long as the markets and societies are stable, one could look up to a comparable initiative / business to draw analogies and predict outcomes. [MK: with that in mind, I believe that comparing companies is of limited use, because companies compete with their org structures and business setups as much as their products and services. So, choosing a proper company to compare yours with is not such a straightforward task.]

  • Prediction markets: people bet real money on specific outcomes. While a single individual can be mistaken, for a crowd many mistakes are cancelling each other out (the wisdom of the crowd), so the group prediction is in theory better than an expert’s prediction (see the past chapters), but no better than a statistical model (!). The core assumption is that rational traders (i.e., those making bets) won’t deliberately lose money.

  • But these markets can be manipulated to make a certain outcome look more likely (and influence other traders), even if this means deliberately losing money for the sake of swaying public opinion (say, in politics the stakes are high and the amounts that can be sacrificed are also high).

  • Building sophisticated models is maybe only applicable to cases like high-frequency trading or internet advertising where a 0.01% improvement can mean serious money; in most other cases there’s simply not enough data for feed a sophisticated model as most businesses make dozens or hundreds of predictions a year at most. Improving a simple model is simply not worth it.

  • Relying on someone’s opinion is a risky venture: they may know the relevant factors for a prediction but won’t be able to estimate, which factors are more and less important (i.e., how to rank them). Companies tend to go to one expert at a time [MK: or if they engage several experts – the opinion of the expert who supports the CEO’s opinion the strongest usually wins – and they don’t average the opinions]. A single expert’s ranking will most likely be worse than the ranking made by the crowd.

  • Of course, the further the prediction looks into the future – the lower the probability of the good prediction and the larger the error will be. And it’s essential to keep track of past predictions to better understand the margin of error and to update the rankings.

  • It’s easier to make accurate predictions about events with the same average frequency (e.g., credit card default rates – will a person make the monthly payment on time or will be late or will default?). The things that matter the most don’t happen at regular times. Think wars, financial crises, or epidemics. This is the case where the wisdom of the crowds is helpless.

  • As anticipating black swans is impossible, it’s close to impossible to anticipate the outcomes of corporate strategy shifts better than assigning a probability somewhere in the decision tree. There’s no second chance (at least, without incurring very high and maybe prohibitively high costs) to get back to the “go/no go” decision fork. Using statistics or crowd wisdom for making strategic decisions can be deadly.

  • The strategy paradox: the companies seemingly employing the best practices in strategy planning can be most vulnerable to planning errors. The main cause of the strategic failure is NOT the bad strategy, it’s the great strategy that just happens to be wrong. It’s the clarity of the initial vision that matters, and the bet on the uncertain future, which is impossible to know. Think Sony Minidisk vs Apple iPod – both were executed brilliantly [MK: I stuck to my Minidisk player for far too long before giving up], but the timing for the former was completely off.

  • The solution to the strategy paradox is quote inconvenient [MK: especially for ego-driven CEOs who attribute all good decisions to themselves]: learning to integrate strategic uncertainty into the planning process. [MK: Strategic uncertainty has 3 features: the nature of things is uncertain, the consequences and the extent of events or circumstances are unpredictable, and credible probabilities to possible outcomes can’t be assigned. Isn’t it a wonderful environment to make decisions in?] Simplistically this can be equalled to scenario planning, where firms make a range of possible futures and the sequences of events and decisions that lead to them.

  • On a practical level, such scenarios have common (core) elements and scenario-specific (contingent) elements. Strategic flexibility is about building strategies around the core elements and hedging contingent elements through investments in several strategic options (i.e., having the action plans to the “what if” kinds of scenarios). Instead of picking one possible (and desired) future, the company should perfect the core elements and be on a lookout for the facts and observations giving rise to some of the contingent elements.

  • Anyone involved in strategic planning can agree that building strategic flexibility is a hugely time-consuming process with no obvious link to the short-to-medium-term compensation, hence the executives’ / Board members time is better spent on perfecting operational capabilities. Back to the org structures – perhaps, the Board and the C-Suite need to delegate more operational responsibilities to division heads and focus on strategy instead.

  • Since this requires a dramatic shift in responsibilities, compensation and investor relations, the chance of implementing true strategic flexibility is slim. [MK: Part of the problem is that this has a lot to do with corporate sustainability (and I’m not talking about tree hugging), which is more important to the company itself than its investors who can diversify the strategic risks away by investing into competitors.]

  • Oh yes, how about another well-known fallacy: the longer it takes to build a model / set of scenarios, the more invested in it people become? This is particularly dangerous when the scenarios can be invalidated by black swans [MK: in our experience, this is what COVID has done to the travel industry]. But at the end of the day, all committed strategies are bordering prophecies.

  • No matter how flexible the strategy is, if a marketplace shift occurs (in terms of new technologies or consumer taste shifts), there’s nothing a firm can do other than make a post-mortem and write off a lot of money (and probably give the poor CEO a boot).

Part 8.