Diffusion Web: A new framework for market entry decisions

Posted on May 14, 2010

0


I recently worked with a team at MIT Sloan on analyzing the diffusion challenges faced by Ember Corporation’s wireless network technology.

As part of the project, we developed a new framework for cross-market diffusion. This framework is useful for innovators that have technologies with applications in multiple markets and need to decide in which order to enter the markets.

Background

The question the framework aims to answer is this: Given a general purpose technology (GPT) and a set of markets, how can we prioritize entry into each of the markets such that we achieve the highest aggregate diffusion and hence returns?

We propose that the following set of factors play a role in how readily a market will adopt a GPT and how attractive the market is for the innovator to target:

  • cost of co-invention[1]
  • cost and quality required of the GPT[2]
  • referenceability of existing markets
  • expected ROI

The Framework

Cost of Co-Invention

The cost of co-invention is the cost incurred in including the GPT in a market specific product. For example, given a new battery technology, how costly is it to (re)engineer a product based on this new technology? This cost has a first order impact on downstream manufacturers willingness to integrate the GPT into their product.

It is worth noting that by cost we refer to both the actual costs as well as perceived risks associated with basing the product on this new technology. For never, more exotic technologies, the perceived risk will be higher and hence the cost of co-invention will also.

The cost can be shared between the innovator of the GPT and the downstream manufacturer to overcome an imbalance in cost of co-invention and expected returns to the two parties.

In the framework the cost of co-invention, i, is described by

i = O*R                    where

O = relative cost of co-invention downstream

R = relative risk of failure

Cost and Quality Required of the GPT

Economies of scale and learning are assumed over time to lower the cost of the GPT. Initially the GPT will therefore only be attractive to markets that benefit sufficiently from the new GPT to accept the higher cost. As cost is reduced, new market opportunities will open up.

Similarly, the quality of the GPT may limit its applicability, but over time process improvements will increase quality and again open up new market opportunities.

In the framework the cost/quality requirements are described by

w = Q/C                   where

Q = relative quality requirement of the market

C = relative cost requirement of the market

Inter-Market Referenceability

The Diffusion of Innovations theory tells us that diffusion follows an S-curve. Some individuals and organizations are early adopters, some the early majority some laggards etc. One factor that influences the rate of diffusion across the various adopter groups is the availability of relevant references.

Similarly, some markets are early adopters, some early majority etc. and references from one market may either help or hinder adoption in another market. As an example, consider how innovations from Formula 1 racing easily could be diffused into road cars, whereas innovations from the road car market would find it harder to be accepted in Formula 1. In marketing we often refer to this as spill-over: Some market segments can help drive sales in other segments and this is usually not a bidirectional relationship.

Hence both the cumulative diffusion as well as direction of diffusion is important. In our framework we capture this referencability by means of assigning a market to a peer group

p = [1..10]            where

a smaller number = more likely to adopt without a reference and

markets regarding each other as peers have the same number

Expected ROI

The expected ROI should of course be taken into consideration when choosing markets. However, it is important to note that this variable is internal to the firm owning the GPT whereas the others are attributes of the markets. Also, although higher ROIs generally are better, higher ROIs often are associated with higher investments and higher risks, which may not suit the innovator. It is therefore informative to interpret the framework with and without this variable.

Applying the Framework

We found that to apply the framework we often had to combine solid data with estimates based on our knowledge of the market. E.g. ROI is usually straightforward to estimate based on pricing and market size, whereas referenceability often had to be estimated.

For each of the factors we calculate/estimate their values and then rescale them to the range 1 to 10. This allows the results to be plotted in a 4 dimensional ‘spider web’ plot. We also calculate the length of each market vector and use this to provide the recommendation for the order in which the markets should be entered.

Battery Technology Example

Our interpretation of the results for the battery technology market is that the cost of co-invention is smallest for Power Tools, as they already sell products with battery packs and it is a simple case of swapping out the technology inside the battery packs. For New Cars however, the cost of co-invention is relatively high as it requires redesigning e.g. the battery management system etc. in the car – or building an entirely new car around the battery!

Power Tools and Bicycles are relatively similar in terms of cost/quality constraints. Small cells and higher acceptance of failure means that these two markets are likely to pay higher prices and accept lower quality than the Van Retrofit and New Cars markets.

Because Power Tools already use batteries, they are the most likely to want to adopt without a reference. Bicycles would be relatively more likely than Van Retrofit or New Cars to accept a reference from Power Tools, but Van Retrofit and New Cars regard each other as peer markets and are therefore close in terms of referencability.

ROI is however more attractive for New Cars than other markets and bicycles likely to provide the lowest ROI as it is a very small market.

To get an absolute ranking, we calculate the length of the 4 dimensional vector for each market and normalize (1.00 being the most attractive market). From this we determine that Power Tools is the most attractive market to target first. For the second market, Van Retrofit and Bicycles are very close so the choice should be to go for Van Retrofit first, to get ‘closer’ to the New Cars market in fewer steps, assuming that is the aspiration.

Wireless Networking Technology Example

For Ember’s wireless networking technology, the ideal sequence in which to target the markets is shown above. Rather than following this sequence, Ember dabbled in most of the markets since 2003 and has only recently begun focusing on the Power / Utility Automation market. We believe this has resulted in a lower than possible increase in shareholder value, increase in competition and a suboptimal approach to protecting its IP.

Conclusion

There is no doubt that additional work needs to be done to refine and validate this new framework, for example by considering how each factor should be weighted. However, these early results indicate that it could provide a valuable tool for managers facing the market entry decision.

In addition, there is one key question the framework does not address: At what level of diffusion within a particular market are you in the best position to enter the next market? Must you have reached the early majority within the market? The late majority?

Thanks to Andre Hamman, William Palm, Josiah Seale and Vivek Raghunathan for feedback on the framework.


[1] T. Breshnahan, S. Greenstein “Technical Progress and Co-invention in Computing and the Uses of Computers”

[2] T. Bresnahan, M. Trajtenberg “General Purpose Technologies: “”Engines of Growth?””

About these ads
Posted in: innovation