Most of the discussion around decision-making in business can be easily grouped into three core branches: a) Strategy and target-setting issues; b) Problem-solving frameworks and tools; c) Digital and data-driven transformation. All three make perfect sense: you do want to have a winning strategy and smart objectives; you do want to use efficient methods and tools in your problem-solving; and you, of course, want to follow the trend and move into digital. These three are surely very needed discussions, however, there is one practical problem: each of them is focusing on its own narrow view of decision-making. Let me explain this in more detail below:
a) Strategy and target setting: defining the problem (or defining strategic goals) is a huge field of exploration for scholars and business thinkers. Here you get all sorts of approaches on how your optimal decision-making should look like (SWOT, PEST, Porter`s 5 forces, Rumelt`s Criteria, etc.). But this is where this branch ends. It is not going beyond strategy and not getting into problem-solving.
Example: Oracles` style: “We are Strategy Team, we strategize big, we feel markets with our guts, and we drive this company. We are not wasting time on feasibility analysis or potential practical implications”.
b) Problem Solving: Huge pile of studies on problem-solving is available already and it keeps growing partially being heated up by digital paradigm and advances in behavioral economics. Problem-solvers are pushing our understanding of decision-making under limited rationality (biases) and uncertainty (expected utility theory; business probability fallacy; as well as so-called known unknowns, black swans etc.). But unfortunately, they do not go beyond that and tend to forget the existence of strategic constraints in organisations.
Example: Doers` style: “Yes, I’m an irrational and completely biased manager trying to solve my business problem being surrounded by black swans… But (surprisingly!) despite all this darkness around, I do have guidelines defined for me by OKRs, SLAs and strategic scorecards”.
c) Digital and data-driven: The data science branch supplies business with incredible instruments and tools for estimation and quantification. But it often cannot resist its natural tendency to float away from the business context they are tasked with in favor of adapting even more sophisticated mathematical apparatus to play with. So that their algorithms, created to automate decision-making, are becoming even darker black boxes, and the stories they tell back to business become more of a science fiction than operational advice.
Example: Nerds` style: “We see through data. We live and breathe Digital. We master algorithms which only we understand. We reside in the clouds (of data) and dive deep into our data-lakes, we correlate anything to everything and only we can tell a proven story of where this company should be heading to”.
You may look at these three branches as a sort of three centers of mass where, most of the time, decision-making discussion tends to gravitate to one of these centers. As it cannot exist in reality (because three centers of mass do not make sense, ask any physicist), the same is valid for optimum decision-making: if we want to improve our decisions, we should also want to combine these three separate branches in our decision-making process.
This is what I call a full cycle of decision-making: you start with problem analysis and target setting (strategy), you propagate it into operational guidelines (OKRs and scorecards which streamline tactical operational problem-solving), you run the operation and make decisions (to make sure it runs), then you measure and reflect on how effective your decisions were (versus the targets set at the start of the cycle). So, we now get a full cycle with a feedback loop. And it is natural for a person (decision maker) to be curious about how effective in the end his or her actions were (until of course you are a total fan of the behavioral economics principle of human irrationality). Some would call this ‘retrospective analysis’ (similar to Agile and SCRUM methodologies), I call it simply a feedback loop on your decision-making.
If such retrospective analysis is so practical and obvious for a person, then why is it mostly skipped out of consideration in the organisations? — obviously because, at first glance, it is perceived as dangerous. It can come out that the decision was ‘bad’ (not optimum). And this is a pitfall to avoid here, as there are several key elements to keep in mind to get the maximum out of full cycle decision-making concept:
- The goal of the retrospective analysis in decision-making is to fight organisational biases.
For example: the natural tendency to estimate probabilities based on assumptions (but not on your historical data) or over-optimism (especially not supported by probabilities).
- Main focus of the feedback exercise is the process of getting to the decision. Not the decision itself. You, at first, want to identify key elements which are adding value to this process, as well as if there were any gaps in the analysis.
For example: it all was based on some ‘hockey stick’ graph, but did we look at how relevant this was to our organisation context?
- The outcome of the feedback process is not ‘binary’. It is not ‘good’ or ‘bad’ lists. Output is the list of improvements to the way management is taking the decisions in your organisation (where ‘your organisation’ means a tailored, context-specific to your business, not that ‘common’ one from a management book).
For example: reinforcing ex-ante market research; or mandating scenario-planning in at least three flavors (starting with ‘very negative’, then ‘negative’ and ‘realistic).
- Full-cycle decision-making does require two critical building blocks:
a) business analysis unit (spread through departments or centralised);
b) data collection and retention capability (can be tasked to an analytical unit or fully delegated to IT).
Both building blocks rarely appear out of the blue. They have to be sponsored and pushed strategically (top-down innovation sort of).
