Everything is Obvious 8/10

Parts 5, 6, 7.

  • Emergent strategies (like Zara’s 2-week design turnaround) are not about predicting trends, but about the quick reaction of the things that occur here and now. Fast prototyping and fast removal of non-selling items is an equivalent of the OODA loop in retail. In online this process is referred to the “A/B testing” (or “bucket testing” or many similar variations of the term). The whole idea is to break the “predict and control” approach and turn it into “measure and react” – if the business model and assets involved allow for this. It’s also a psychological shift when people actively refuse to predict the future.

  • [MK: the book makes a case for “crowdsourcing” that used to be all the rage in 2006+, but time has demonstrated that users may be better off with the “curated” (i.e. specifically selected) content instead.]

  • [MK: the book makes another case for the “Mechanical Turk” by Amazon, giving people access to a huge army of “turkers” doing mundane tasks like sorting out the data, providing opinions, etc. It correctly states that the people who work for pennies may not be the most representative audience for any research or provide sufficient quality of judgement. There are ways to slightly improve the quality, but I’m personally still sceptical of the widespread use of this idea.]

  • [MK: then the book describes (in a shallow way) the advertising analysis and prediction mechanisms for not just targeting ads to maximize revenue, but also to predict the outcomes of things like box office revenues for an opening weekend, the place of a new song in the chart, and the like. The author can be forgiven for this shallowness as the industry keeps evolving very quickly and whatever ends up in a printed book is already obsolete to a large extent.]

  • The book uses a cliché that “half the money I spend on marketing is wasted – I just don’t know which half”, but then offers an interesting observation about brand marketing. We all know it’s close to impossible to reliably measure brand marketing efforts, so allocating this budget is a huge leap of faith (unless, of course, brand marketing budgets are allocated as a percentage of revenue or the entire marketing budget, etc.). But what if increasing brand marketing expenditure and observing an uplift in brand recognition are merely correlated, but the former doesn’t cause the latter? For instance, increasing ad spend during peak shopping season may have a disastrous ROMI, because people’s shopping interest also peaks during this time.

  • One could argue that the observed uplift is a result of many parallel activities the company performs and the decisions being made about the brand; companies deciding to increase their (for instance) brand spend also do things ranging from modifying communication strategies, changing product packaging (and sometimes – the price) and doing other things from the 7P marketing mix as a result of the huge investment.

  • Checking the effectiveness of brand ads is sometimes possible in new geographies (countries, states, cities) where there’s not many overlapping brand messages, so that the signal can be told from the noise. But still – this may work well for awareness, but not translate into the intent to buy.

  • There are only limited opportunities for experiments in business – many decisions are too resource-intensive, require a certain not easily substitutable workforce or are simply irreversible. So, Boards need to make such leaps of faith (at the end of the day, it’s the Boards’ job) – but need “local knowledge” (the book tries to make an analogy between warfare and business where in the latter case “local” means the strengths, knowledge and the resources of the company).

  • Governments can choose many policies to encourage businesses to be more effective in the critical areas (currently this is the “green” agenda):

    • “Cap and Trade” for emissions - risky because of derivatives

    • Taxation of the undesirable performance (e.g., emissions) – tax is usually passed through to consumers

    • Offering prizes for specific outcomes – creates incentives for self-selected companies. MK: the applicability of this approach seems limited as companies don’t operate this way all the time.

  • Bootstrapping is another popular idea widely described in books: go bottom-up, from a citizen to a community, from a business component to a division (Toyota just-in-time operations, anyone?), aggressively weed out minor inefficiencies in one’s own yard. Tracing problems to root causes reduces the “technical debt” that’s present in any complex system. And looking around for comparable outside solutions to local problems increases the long-term flexibility of the system.

  • MK: This works really well in theory, but bootstrapping has natural limits, not just confined to the culture, but also the author underplays the micro-macro problem: citizens don’t have enough energy and resources to create an alternative to the government.

  • Effective bootstrapping assumes that the solutions exist at least partially, and the owners of such solutions don’t mind sharing them with the others; the job of a planner is not to develop solutions, but rather to scout for and promote them. The logic here is that local solutions have a higher adoption rate with a higher enthusiasm than the centrally prescribed plans. Plans fail not because they ignore common sense, but because planners rely on their own common sense when dealing with the people who are different from them.

  • It’s easy to criticize any kinds of policies because “they won’t work”, but it’s more of an intuitive judgement, most often – not backed by the data. Of course, data can be manipulated to highlight the preferred option to a policymaker.

Part 9.