Making strategic decisions under uncertainty: The case for non-predictive strategy

The goal of strategy is to decide what to do in a given situation to achieve a given objective.  Basically, strategic decisions comes down to the question “what to do next?”. In environments characterized by uncertainty (defined as objective lack of information), this is no simple question, and several approaches are possible to address it.  Two dimensions characterize these possible approaches: prediction and control.

Prediction asks  to what extent does my approach rely on a forecast of the future environment. Strong prediction corresponds to either a planning-type approach – I create a detailed prediction of the future before initiating action – or a vision type:  I imagine the future and I strive to make this vision a reality.  Low prediction corresponds to a more adaptive approach:  I do not try to predict the future environment, but instead I move on and I adapt to changes along the way.
Control asks how I can control the evolution of my environment.  The over-arching assumption of classic strategy is that the firm has little influence on its environment, which is for the most part given (or “exogenous”).  All a firm can do is to find a place in this environment (planning /positioning) or adapt when it changes (adaptation).  Hence the importance of the notion of “fit” that the field insists upon (e.g. Michael Porter in 1996).  On the opposite side of the spectrum, the field of entrepreneurship observes that a firm can change its environment in profound ways, sometimes from an ex ante defined vision, or through the logic of future-agnostic gradual transformation of the environment.  There are many examples of entrepreneurs starting with odds apparently stacked against them and completely transforming their environments:  Michael Dell, Richard Branson, Sam Walton, to name just a few.

So there are four possible approaches, illustrated in the figure below:

What to do next? Prediction vs Control (Wiltbank et al, 2006)

Let’s review these approaches.

Strong prediction, strong control (top right):  in this configuration, I am a visionary, I have a strong vision of the future environment and I am committed to making this vision a reality through my actions because I am able, or I think I am able, to change the environment.  This configuration is assumed by strategy practices that take a vision and mission as a starting point.

Strong prediction, low control (top left):  In this configuration, I develop a strong vision of the future market, but I am basically unable to influence it significantly.  The paradigm is that corresponding to classical strategy, planning, based primarily on the discovery of a possible acceptable position (fit) for my firm in an environment over which I have little influence.  Therefore, the quality of my prediction is essential, and most of the strategic work is devoted to it (this phase is called strategic analysis).  If I get my prediction right, I succeed.  If I get it wrong, then I am in trouble; hence this strategy is also fragile, especially in uncertain and turbulent environments.  Note that for spectacular success, such strategies depend on a unique ability to make an accurate prediction, or you’re simply one more trend-follower.

Low prediction, low control (bottom left):  As in the previous configuration, I am a “taker” of the environment, about which I do not develop a vision or a prediction.  But I can not exert any influence on it either.  It is part of classic strategy, but upgraded to handle the case of industries affected by turbulence, such as high-tech sector.  The paradigm is that of adjustment or trial and error.  The key to success in this configuration is flexibility, i.e. the ability to adapt to a new situation quickly and at low-cost.  While this approach is popular (especially in today’s seemingly unpredictable world), its limitation is that adaptation by definition reactive, i.e. taking the risk of always being late.  This is because most reliable indicators are lagging, and because it takes time to react, however nimble one may be.  Also, being purely reactive means taking the risk of not having the right assets (knowledge, experience) at the right time. As such, adaptation is important, but it cannot be a firm’s sole approach in the long run.

Low prediction, strong control (bottom right):  Most interesting is the final configuration, in which I do not develop vision or prediction of the future environment, but I nonetheless seek to exercise a strong control on its evolution.  This can be done through partnerships, coalitions, influence on standards, etc.  The approach is said to be transformative, because in it I transform (at least in part) my environment, if only to a limited extent.  More specifically, I co-create it with selected stakeholders.  The approach – which has much in common with “integrated strategy” – corresponds to Effectuation, a theory of entrepreneurship.  Effectuation describes how entrepreneurs create new markets when faced with a situation of strong uncertainty.  It is this non-predictive approach to strategy that today offers the most prospects for strategy development, and it also works well for large companies:  they can learn from entrepreneurs who, after all, are the experts in dealing with uncertainty.  We will develop the concept of non-predictive strategy in future posts.

One source for this post is the following article, which I highly recommend: Wiltbank, R., Dew, N., Read, S. and Sarasvathy, S. D. (2006) “What to do next? The case for non-predictive strategy,” Strategic Management Journal n°27: pp981-998.

To read more on the challenge of prediction and forecasting, you can read my post: We have met the enemy and he is, her, forecasting. On the topic of strategy itself, and to see why prevailing tools are stale, you can read Milo’s post: Start with Geostrategy, or call it tactics.

Finally, if you enjoyed this post, why not subscribe to our blog?

Advertisements

6 responses to “Making strategic decisions under uncertainty: The case for non-predictive strategy

  1. Excellent.

    And the transformational aspect is also essential for my company itself, since I need a lot of organisational fit-ness and auto-regulatory dynamics to create transformational impact on my environment.
    Which of course demands a new understanding of leadership.

    […link to follow…]
    Christian

  2. Fascinating! If you’re familiar with Myers-Briggs, you know people either tend to favor a Sensing approach, taking in information in the now, as it comes, or an Intuitive approach, in which they imagine, interpret, and forecast. I am a believer in the both/and approach. I would argue for a blend. I wouldn’t rule out using the power of prediction — intuition is faster, cheapter, easier, and solves problems before they start.

  3. Carlos Ruiz Gomez

    Interesting, thanks.
    The transformation cage is indeed interesting. But I also find interesting the dynamics of the framework, as in the BCG matrix, where it’s not just interesting to see what type of product / service does a company sell (or where a BU fits in), but even more interesting is i.e. the efforts that a company must do for a question mark to transform into a cash cow and not into a dog.
    This is a similar case. I.e. it would seem that, as long as you are in control, the prediction (or lack of) may not be that important. And as such, generally companies try to either prove that they are visionaries to their clients or that they are driving the transformation of their industries.
    But there are also some times when a company enters directly in a adaptation stage, a seemingly undesirable cage. But, if it is part of the company strategy to use that cage merely as an entry stage, to then shift into being more in control, then that is valid. This is the case of some of the cases depicted by Prof. Christensen and the disruptive innovation framework. Now, the question is, what is it more convenient to become, a visionary or the transformational industry force? I guess, it depends. A truely innovative industry should make companies in the adaptation guilder to become visionaries. While comoditised industries will head to the transformational stage; and it’s a terrain for oligopolies.

    Carlos Ruiz Gómez-Ibáñez

    Cell: 00 34 609 226 676
    E-mail: cruizgomez@gmail.com
    Blog: carlosrgomez.wordpress.com
    Profile: http://www.linkedin.com/in/carlosruiz
    Facebook: http://www.facebook.com/cruizgomez
    Twitter: @cruizgomez

  4. Of the four possible approaches, illustrated in the figure – What to do next? Prediction vs Control (Wiltbank et al, 2006), we can fill in some famous names as examples, to understand the inherent connotations involved.
    Top left – APPLE
    Top right – Man on moon mission
    Bottom left – Cheap imitation, software piracy.
    Bottom right – South Korea in the nineties.

    Any comments on this approach ?

    • Hello Sid
      Thanks for your comment. Giving example is an excellent idea. So on the top left you would have a large company such as GE planning for a new plane engine. On the top right, the moon mission is a good example; Salesforce.com is also the result of a “vision” approach to change the way we use/buy software. Bottom left is indeed copy cats, but not necessarily “cheap”. Bottom right is transformative entrepreneurship: Google, Facebook, Grameen, Edison, Sony in the 50s, etc.
      Please understand that none of these four approaches is bad or good. What is important is to understand when they are relevant, and what they mean.

      Best
      Philippe

  5. Sorry, a bbit of a late reply.

    I find that when dealing with the human dimension we will never be able to predict anyhting 100%. There are far too many variables. I find that the non-predictive aprroach towards strategy is good; however, would there be a preponderance towards pattern based anlaysis and relying on experts who, being human, may have specific biases that may be a negative influence towards your strategic decision.
    One method that I always lean towards in my work is to look at multiple hypothesis, chosen through other methods and based on a problem set, reverse engineering them, identifying indicators that would lead towards each hypothesis. Once that is developed you rank or weigh the indicators, primarily looking at its relevance and timliness (early warning) and build them into a time line of line of operation. As we begin to see certain relevant and timely indicators being met the analyst can begin adjusting assessments and predictivity in order to form sound advice towards a decision maker or to make the decision for her/himself.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s