Consider the statement:
The strategies generated by data driven cultures are likely to produce sustainable competitive advantages.
Data Driven Cultures
Carl Anderson, 2015 (Data Scientist at Warby Parker) defines a data driven culture:
- Is continuously testing;
- Has a continuous improvement mindset;
- Is involved in predictive modeling and model improvement;
- Chooses among actions using a suite of weighted variables;
- Has a culture where decision makers take notice of key findings, trust them, and act upon them;
- Uses data to help inform and influence strategy.
For the purposes of brevity, I’ll define a strategy as:
- An artifact;
- That enunciates choices selected from acknowledged tradeoffs;
- Which is rooted in a paradigm;
- That is actionable;
- With the intent of causing a sustainable competitive advantage in the future.
Artifact, Paradigm, tradeoffs, competitive advantage
An artifact can be a deck, a story, or word doc, it can be a poster. Really, it can be anything that people use to communicate intuition to each other. A superior strategy that is communicable and understandable is also so much more likely to be executable.
A paradigm is meant here in the scientific, public-policy, sense. It is a set of axioms, definitions, and models that a group of people accept as being the way a world works. An element of a paradigm may be the term ‘customer centric’, which, unto itself, has a dozen or so definitions, ideas, and concepts.
There are tradeoffs. It is in the assembly of tradeoffs, in the context of a system, that reinforcing causes have a chance to add up into a sustainable competitive advantage. Ignoring tradeoffs, as though they don’t exist, is negligent.
Sustainable competitive advantage is the whole point. Sure, a rotting carcass on the Savannah is enough for a single bird to eat, but pretty soon all sorts of other predators will show up for a taste of that delicious, rotting, market share. Industries disrupt. Companies almost never. Because imitation is predictably inevitable, what set of choices can you make to make it hard, if not impossible, for new entrants to imitate? (See Porter (1998) for more)
The sorts of strategies that a data driven culture might create
Data driven cultures are likely to create a web of reinforcing because statements that form tightly coiled, virtuous circles.
- Because we only hire the top 1% of those who apply, we have to pay total compensation in the upper 5% percentile.
- Because we compensate in the upper 5% percentile, the time of our people is precious.
- Because the time of our people is precious, we invest heavily in automating or avoiding low value activities.
- Because the time of our people is precious, we have to charge higher rates per hour.
- Because we invest heavily in automation and people, we deliver a far more satisfying experience to our customers.
- Because we deliver a far more satisfying experience to our customers, customers will pay higher rates.
- Because we deliver a far more satisfying experience to our clients, we can charge more and attract quality clients.
- Because we charge more and attract quality clients, we hire the top 1% of those who apply.
- Because our customers care only for monetary price, we have to compensate the talent in the lower 20% percentile of the labour market.
- Because we do not pay overtime, and pay below-median compensation, the time of our people is not precious.
- Because our employees subsidize our business with their spare time, we can afford to charge our customers less.
- Because their time is not precious, we do not have to invest in anything.
- Because we do not invest in anything, our margins are sustainable.
- Because our margins are sustainable, we can afford to compete on price alone.
The first example I described is a boutique. The second is a sweatshop. And, in part, because there’s always lot more talent in the bottom 20% than there is in the top 5% (by definition!), there’s always going to be people who compete on price (sustainably) and on talent (sustainably).
It is possible for a data driven culture to compete on price and convenience (Amazon) just as it is to compete on talent (McKinsey). There’s nothing quantitatively bad about running a sweatshop, just as there isn’t anything quantitatively bad about running an elite practice.
Both are sustainable.
The Continuously Improving Strategy
A Canadian saying is to always skate to where the puck is going, not where it is.
Predictive modelling is forward thinking.
Good strategy is too.
There’s a lot of good evidence that prospection about the future, and making optimal courses of action, are not evenly distributed in the population (See: Zhu et al. 2016 for the latest on this) This insight is at the root of a number of financial instruments and product innovations. It’s an axiom at this point in multiple branches of consumer choice theory and in strategic planning theory — see! paradigms really do matter!
The norm isn’t prospection. The norm is anchor and adjust. Dig out the strategy from four years ago. Did it work? Which parts are you unsatisfied with? Great, tweak those parts and lets get on with it. Predictive modelling can be trained using data from the past. It can also be used to make statements about the future. And moreover, it can be used to make predictions about which actions, taken at specific points, generate alternate futures.
This is where the core difference between the activity of optimization, versus the activity of strategy making, becomes clear.
To pursue a strategy of continuous improvement against a set of tactics is called optimization. The choices in actions have already been made. What’s left is optimizing the settings on the policy instruments prescribed by the strategy. For instance, the activity of adjusting the paid spend-keyword mix is optimizing the instrument of search advertising.
To pursue a strategy of continuously improving strategy is…strategic. A culture driven by ego and dead-reckoning would operate by someone delivering a deck, calling it a strategy, and making no adjustment. After all, if a strategy is rooted solely on judgement and will, how can it possibly be adjusted with the loss of face? If, on the other hand, a strategy is subject is assessment and iterative testing, it could be adjusted and reformulated without as much hurt.
If the stance of the culture is rooted in change and an awareness of the dynamism of a market, then it is better equipped, even from a toolkit perspective, to cause strategic change. Moreover, if the culture is amenable to change, as data driven cultures are, then there’s a higher probability that the organization can change in response to shifting strategy.
It is also far more likely that a data driven culture will surface truths about the institution and the market far faster than other cultures. In a culture where decision makers recognize key findings, trust them, and take action on them, key findings are far more likely to be discovered and communicated. Shifting from a culture driven by fear and firings, to one of hope and management, is likely to cause higher rates of learning about strategy.
It’s for these reasons that data driven cultures are likely to generate strategy that causes sustainable competitive advantage.
Posts in this series include: