Anyone who knows me knows I love strategy. I have railed incessantly about our overreliance on tactical execution and our overlooking of the strategy that should guide said execution. So imagine my discomfort this past week when, in the midst of my following up on the McLuhan theme of my last column, I ran into a tidbit from Ray Rivera, via Forbes, that speculated that strategic management might becoming obsolescent.
Here’s an excerpt: As amounts of data approaching entire populations become available, models become less predictive and more descriptive. As inference becomes obsolete, management methods that rely on it will likely be affected. A likely casualty is strategic management, which attempts to map out the best course of action while factoring in constraints. Classic business strategy (e.g., the five forces) is especially vulnerable to losing the relevance it accumulated over several decades.
The crux of this is the obsolescence of inference. Humans have historically needed to infer to compensate for imperfect information. We couldn’t know everything with certainty, so we had to draw conclusions from the information we did have. The bigger the gap, the greater the need for inference. And, like most things that define us, the ability to infer was sprinkled through our population in a bell-curved standard distribution. We all have the ability to fill in the gaps through inference, but some of us are much better at it than others.
The author of this post speculates that as we get better and more complete information, it will become less important to fill in the gaps to set a path for the future -- and more important to act quickly on what we know, correcting our course in real time: With access to comprehensive data sets and an ability to leave no stone unturned, execution becomes the most troublesome business uncertainty. Successful adaptation to changing conditions will drive competitive advantage more than superior planning.
Now, just in case you’re wondering, I don’t agree with the premise, but there is considerable merit to Rivera’s hypothesis, so let’s consider it using a fairly accessible analogy: the driving of a car. If we’re driving to a destination where we’ve never been before, and we don’t know what we’ll encounter en route, we need a strategy. We need to know the general direction, we need a high-level understanding of the available routes, we need to know what an acceptable period of time would be to reach our destination, and we need some basic strategic guidelines to deal with the unexpected – for example, if a primary route is clogged with traffic, we will find an alternative route using secondary roads. These are all tools we use to help us infer what the best way to get from point A to B might be.
But what if we have a GPS that has access to real-time traffic information and can automatically plot the best available route? Given the analogous scenario, this is as close to perfect information at we’re likely to get. We no longer need a strategy. All we need to do is follow the provided directions and drive. No inference is required. The gaps are filled by the data we have available to us.
So far, so good. But here is the primary reason why I believe strategic thinking is in no danger of expiring anytime soon. If strategy was only about inference, I might agree with Rivera’s take (by he way, he’s from SAP, so he may have a vested interest in promoting the wonders of Big Data).
However, I believe that interpretation and synthesis are much more important outcome of strategy. The drawback of data is that it needs to be put into a context to make it useful. Unlike traffic jams and roadways, which tend to be pretty concrete concepts (stop and go, left or right -- and yes, I used the pun intentionally), business is a much more abstract beast. One can measure performance indicators ad nauseam, but there should be some framework to give them meaning. We can’t just count trees (or, in the era of Big Data, the number of leaves per limb per tree). We need to recognize a forest when we see one.
Interpretation is one advantage, but synthesis is the true gold that strategic thinking yields. Data tends to live in silos. Metrics tend to be analyzed in homogenous segments (for example, Web stats, productivity yields, efficiency KPIs). True strategy can bring disparate threads together and create opportunities where none existed before. Here, strategy is not about filling the gaps in the information you have, it’s about using that information in new ways to create something remarkable.
I disagree most vehemently with Rivera when he says: While not disappearing altogether, strategy is likely to combine with execution to become a single business function.
I’ve been working in this business for going on three decades now. In all that time, I have rarely seen strategy and execution combine successfully in a single function (or, for that matter, a single person). They are two totally different ways of thinking, relying on two different skill sets. They are both required, but I don’t believe they can be combined.
Strategy is that intimately and essentially human place where business is not simply science, but becomes art. It is driven by intuition and vision. And I, for one, am not looking forward to the day where it becomes obsolescent.
Gord, you titillate with the headline, but you know strategy won't die. Big data informs a strategy and aids in execution, but alone can also lead a team to over-focus on KPIs or fail to recognize that the KPIs given them by management were in fact not aligned with business goals.
Big Data is wonderful. We've (Didit) been able to do amazing things with "Big Data" in SEM by using mashups of first and third-party data to successfully predict audience based on geography (and then turn that process into a technology). But if the strategy doesn't properly define the audience the campaign efficiencies are not nearly as great.
As always, well said. Like Kevin, I've been involved in Big Data long enough to know "mostly" what it can and what it can't do. WSe certainly haven't exhausted it's capabilities, but regardless of how good it gets, there will never be a replacement for the human brain in the business world. We are attempting to automate the elements that can be automated and create process around the mundane tasks, but as you say, this is both science AND art. Art requires imagination and looking for the things that we still don't know. Big Data can't predict the most predictable of all things...things will change. Trends, Fads, Catastrophes, Breakthroughs, Sea Changes... I would also assert that over time, we're likely to see less data rather than more. Privacy issues are already beginning to impact our access to some data that we once enjoyed having. "Not Provided" is only one example directly related to search. There will be more case where we'll need to extrapolate more not less.
Love your thinking, Gord. And I, too, read that article with interest. Big data is a tool, not a replacement for thinking. What I do think is particularly interesting about where we are today with data is marketers' ability to use data to be adept, move quickly, make changes and adaptations based on what the data around your campaigns shows you. Some marketers can do that and many can't. And some businesses have a culture that makes that possible, and many more do not. And it might well present a competitive advantage. But all the data in the world without someone who can analyze it and plot a strategy that makes sense is about as valuable as the social media ninja you hire to manage your social media accounts who has no knowledge of marketing.
Rivera's article is very thought provoking and brings up interesting points particularly MacLuhan's tetrad to analyze the effects of media. Ironically, the 2x2 matrix is classic strategic analysis. Like you and the other commenters, we disagree with Rivera that strategy would become obsolete - and I believe that he hurt his credibility suggesting it. People and computers can't know the future, even with perfect data, predicting the future is an inexact science (Dennis makes some great points). Secondly, using computer algorithms to drive a company's strategy is a recipe for failure (IMHO), just like people have figured out how to game Google's algorithm, competitors could do the same thing. Finally, I am skeptical that we will ever reach the point where his utopian vision of data will ever be achieved. (" . . . data approaching entire populations become available, models become less predictive and more descriptive. As inference becomes obsolete, management methods that rely on it will likely be affected." ). There will always be unknowns, uncertainties, gaps and time lags.
Without strategy the data would be just that, data. Any data analysis logarithm requires strategic thinking to set up as well as monitoring to maintain its relevancy. I am with Gord on this.
Strategic thinking is a given. It won't go away.
What has changed is the role of data mining.
In the past, the relational database was the repository with which data scientists tried to glean patterns and trends to support the business roadmap.
With big data, RDBMS has a time lag to produce datasets that analysts can mine. For example, trying to determine optimum delivery of electrical energy needs real time data capture and analysis. RDBMS provides a snapshot but too late for real time response.
What this means is that the focus has shifted towards data in motion (analytics on transactional events) rather than data at rest (data repositories).
As a result, big data is not the problem as it makes no sense to flow all the information into a relational system. Rather, in memory SQL and non-SQL analysis on the data flow will become more important.
It makes little sense to capture granular events for analytics due to inherent noise in the data and think that studying patterns will impact strategic thinking.
Statisticians have always known that sampling events gives enough of a picture to allow planners to make decisions.
Big data is not the problem.
Absolutely. Its simpler than all that though. Planning would be unnecessary if we could react perfectly to any current decision. Agility trumps planning every time.
It isn't "big data" that wins, its the actions that result from "right now" data combined with "normative" data ie. historical data. As another poster suggests, bigger isn't better, but actionability makes money. A slightly better action is way better than a perfect decision that results in no action. The winners have been those who take imperfect information and use it to make better decisions. That, of course, has always been the truth about business information.
Yes, but not yet. Currently Big Data systems can beat people in repetitive tasks such as targeting marketing, or where making decisions where extreme speed is of the essence such as spotting arbitrage opportunities. But strategic decisions with a timescale of months or years are science fiction.
Here we go again, what is this current obsession with data, there's absolutely nothing innovative a\bout this, smart marketers have been using data of all kinds from all sources and devices successfully for years, so what's new?
Thanks all for the comments. There's significant fodder here for at least another column (or two).