Posted 2011-09-29 17:34 under system dynamics, interactive, strategy
Here is a live on-line simulation model of a hypothetical consumer electronics company, with a dashboard to let you play with it. The model aims to explore the trade-offs between short-term performance and the drivers of longer-term success. One way of looking at these dynamics is that an organization needs to build up and maintain stocks of resources in order to be healthy, a perspective for which system dynamics is very useful.
The framework we are using here for examining the trade-offs between short-term performance and its longer-term drivers is the resource-based view of the firm, which views strategy as the process of building and maintaining a system of resources, in particular capabilities. The key implication is that key resources which take a long time to build are hard to copy, thus providing a sustainable source of competitive advantage. The seminal article we read on this during my studies at the MIT Sloan School was The Resource-Based View of the Firm by Birger Wernerfelt (from the Strategic Management Journal, 1984; Wernerfelt was and still is on the faculty at Sloan). But the framework has been much treated in the literature, elements of it going back to Coase in the 1930s.
This framework is of course very well suited to system dynamics, which models the world in terms of stocks and flows. Stocks can be used to represent strategic resources such as staff skills, installed base, customer loyalty, a strong coalition of partners, etc. And flows can represent the results of actions that build up those stocks, or forces which deplete them over time.
The model on this page, and the dashboard in front of it, are an attempt to construct a template that illustrates how the stock and flow approach could be a useful tool for formulating longer-term strategies based on the resource-based view.
Using the Dashboard
Note that this dashboard works best with recent versions of Firefox, Chrome, Safari, or Opera. It is currently rather sluggish with Internet Explorer, and does not appear to work with Camino, so please use one of the other browsers if you can. This is because IE does not yet directly support the modern web standards for doing graphics (such as canvas and SVG), and thus requires an intermediate layer to work (thanks, Google).
The left side of the dashboard has eight sliders, on which you can click or drag to change the underlying model parameter. When you change a slider, the underlying model is instantly resimulated for 20 quarters, and all graphs are updated.
The right side of the dashboard consists of six panels, selectable by tabs. Each panel is a separate view on top of the same, underlying model, and shows different graphs or gauges related to a particular perspective on the business.
You cannot yet save the value of your sliders (let us know in the comments if this is something you would like us to add). You can restore the starting values for all sliders at any time by clicking the refresh button on your browser.
How the Model Works
Tying it all Together
The various tabs on our dashboard (described above) describe different aspects of our company's operations and strategy. But these do not act in isolation, but rather work together to drive health and performance. This diagram shows the model underneath our dashboard represents these connections:
The connections are annotated in the diagram, but note the following:
- The diagram explains how the model works, not how the world works
- What connects the company's decisions (and this its health) with performance is market share, which is represented in the model by a stock. This stock does not change unless a change in competitiveness causes it to change
- There are separate links to and from Performance and each of the three main sectors of the model (People, Innovation, and Brand). The link from Performance drives the ability of the company to fund these areas, and the link to performance drives costs
- The link from People to Innovation reflects the model's assumption that the product pipeline will be more effective if staff are more skilled and motivated
- As explained above, the Brand part of the model is weak, and responds only to brute marketing spend; it is being refined to incorporate the effects of innovation and market competitiveness
This model is very simple, and is not calibrated to the data of a real company, but plausibly represents the dynamics behind company health and performance:
- You can increase performance in the short-term, at the expense of long-term health
- Conversely, investing in the short term can depress short-term results, but build resources that underpin stronger long-term performance
- Health is an aggregate of several metrics, and weakness in any one metric affects health as a whole (although perhaps not as extremely in real life as in the model -- discussion welcome)
- Health drivers are a useful leading indicator of longer-term performance, but the reverse is not necessarily true (???)
- There is a lot of feedback in the system: cutting staff affects not only costs, but quality, which in the long term will make it more difficult to attract and retain good people (there are many more instances of feedback in the model, perhaps these deserve a separate list)
Taking it Further
This first cut is a simple but illustrative, and invites several avenues for development:
- It could be customized to different sectors, and the health-vs-performance differences between sectors could be articulated
- It could be applied to a specific company, with a view to exploring strategies that balance short-term performance with investing for the longer term
- It can be refined to incorporate best practices into some admittedly crude elements, e.g., the assumptions behind brand equity
- Competitive behaviour could be introduced (the model now assumes that competitor behaviour is static
- It could be turned into a training game, where players make quarterly decisions in order to maximize value
I'm planning to refine sections of this model going forward, and to post entries as things happen. In the meantime, comments are welcome.