Thursday, May 22, 2008

Cambridge Open Innovation Workshop - Session 3

The following are notes from the first session of the Open Innovation Workshop at Cambridge University in the UK May 22nd and 23rd of 2008.

Key questions posed for the third session of the Cambridge Open Innovation Workshop by Professor Peter Williamson of Cambridge’s Judge Business School were as follows:

Where does open source fit into the current and future business models of different firms?

-Reaching critical mass
-stimulating and leveraging investments by others
-speed up innovation by harnessing diverse capabilities and experience

Where do we find these profitable business models?
What are the tensions between open source, control, quality and value capture?

Value Capture (X-axis) vs Value Creation (Y-axis)

High Value Creation, Low Value Capture: Open standards – good for society, few rewards to innovators (other than fame). Examples: Linus Torvalds, Tim Berners Lee
High focus on value capture, low focus on value creation: Killing the Golden Goose – narrow focus on profitability restricts ecosystems
“Magic quadrant” - Ecosystem leadership – growing the ecosystem and securing profitability.

So how do we get to the magic quadrant, especially when open innovation (especially when not customer centric) often produces products that can fit oddly into the marketplace (think open source software).

Lorraine Morgan from the University of Limerick touched on a topic, Open Source business models that is a frequent topic of discussion in the IT and innovation communities. The main question Morgan deals with is this one: How does a firm create value for the customer while capturing some of the value for itself?

Nick Wainwright and I had an interesting conversation where he mentioned that supporting open source can be an effective way to subvert a powerful market leader. Look no further than IBM supporting Linux almost a decade after OS/2 bombed. While the benefits to IBM were great, the appeal of hurting Microsoft must have been nearly as attractive.

As Morgan discusses, understanding the components of the value chain is difficult in open source. She uses a model breaking it into value creation, value capture and the value network. However, many of the traditional frameworks fail for open source. For instance, traditional resource-based views are concerned with value appropriation, which is strongest when intellectual property is owned. Thus, they have little explanatory value about when open source software is a feasible strategic option. Morgan has a strong understanding of the literature around this issue and as a first year PhD she has posed a lot of interesting questions that many are seeking the answer to.

Dr. Christina Raasch from the Hamburg Technical University’s talk focused on using open source outside of the software business. Some models she is looking at are open content such as cultural goods, open science and bioinformatics, and open design such as hardware, transportation and consumer goods. I listen with interest, since I am working on an open content model startup right now and staring at pushing the stone uphill to get it to critical mass in terms of content. When I reflect on it, content is a substitute for credibility, which a startup doesn’t have much of.

Raasch’s model is more sophisticated than my thinking. She breaks it up into economic, social, technical and legal aspects. However, credibility goes a long way towards getting people to contribute to an open source project.

She mentions a “bazaar governance” structure for open source management. With open source projects to develop physical objects such as cars, network effects, access, and other issues of project governance are different than open source software.

Maha Shaikh, a post-doctoral researcher from the London School of Economics talked about a shift in focus from open source products to process.

Why open source? She started up by summarizing some reasons to open-source:

  • Reduce lock-in to commercial software
  • Company-wide or OS community access to expertise
  • Reduce R&D Costs
  • Speeds up time to market
  • Better code quality through scrutiny
  • Better quality product through visibility

Shaikh wants to do additional research to validate these general assumptions. They’re all good questions, but it sounds like a challenge as these are so product and situation specific.

Product focus open-sourcing

Charter-sourcing - Low innovation, relatively small projects where software development is outsourced.
Symbiotic Source –Get supporting code with version control for software projects (underlying source code for a bigger project). Example, GlueCode.
Coalition – Alliance of strategic convenience
Mature source – Use commercially mature open source software.

Processed Focused
Talent sourcing – Build relationships with external communities, requires companywide adoption of open source ideology.
Portal Source – Access to a large pool of talented developers exploiting distributed intelligence.
Open Outsourcing –Open source philosophy with open standards.
Inner Source – In-house research department with high transparency, sharing. Gary – Inner Source – Bell Labs, IMB, HP, Sun, Phillips Medical. Take open source and bringing inside the company.

After Shaikh’s talk, there were a few interesting questions.

Nick Wainwright, who spoke earlier in the day. He asked about addressing the nature of the license. Some have political intent, some reserve rights while some are almost completely open.

Raasch indicated that physical objects, licenses cause a lot of problems. Each project tries to develop its own license so it is problematic. A more general license procedure in the product realm similar to GPL and the like would make things easier.

I asked a question about free public license conflict and the complexity of managing a code base, specifically mentioning Black Duck Software, Doug Levin’s Waltham, MA company that is making a killing helping companies understand just what exactly is in their code base. The panel agreed that it was important to focus when deciding what software with what existing licenses what the end results you are going to use it for.

Satya Dash asked Raasch a question about how much project management actually works in a open source environment?. Raasch has seen different approaches, but stronger, proactive project management tends to result in better project, as opposed to decentralized, more informal. Central interfaces are very valuable. Raasch also suggested that timelines are valuable, especially when the project is funded.

Marshall Van Alstyne finished the day by presenting research about Innovation, Openness and Platform Control. His research is funded by the NSF, Cisco and Microsoft.

Some key questions addressed by Van Alstyne’s talk:

  • How do business actually make money off open innovation?
  • Can intellectual property regimes be used to promote downstream innovation?
  • How do sponsored platforms (as opposed) to standards create value?
  • Does competition threaten or enhance innovation?

Van Alstyne went into a number of interesting examples of how downstream enhancement can add value. Examples:

  • Google Maps – Paul Rademacher – combines Craigslist with Google Maps. Lawyers say sue. Engineers say hire him! It’s now a staple of Google Maps and central to their strategy of bringing local information to the individual.
  • Metallica originally sued. Radiohead allows re-mixing, but if you post them back, they own it. Marshall doesn’t think they got this right.
  • Linden Lab allowed people to actually own property in 2nd life. Applications allow downstream innovation by users which has been central to their success.
  • Facebook ‘s rapid growth once they open the standard.

What’s the economic model behind opening platforms up? An initial sponsor offers platform value of V. Developers build value by creating new layers of application. In the future, the sponsor can bundle developed work in with their application to increase value extraction. Meanwhile, if developers have been treated fairly, sequential rounds of innovation can occur.

According to Van Alstyne, sponsor and developers divide surplus based on Nash bargaining.

Economics can also account for value added by developers. The price is limited until the platform sponsor bundles application into the open pool. Output is Cobb-Douglas – the open resource pool is added to production.

So how should a firm manage the platform ecosystem? How open should it be? And when should new features become free parts of standard platform? In this case, it’s a spectrum, the more closed, the more you can charge upfront. The more open, the more others can add value. A firm should open a platform enough that the opportunity cost is proportional to the growth in value (elasticity of output). Addressing the latter question, firms should fold new features into the platform at the point time where the value of next generation output passes current generation.

Van Alstyne then asked: Who will win the next generation of the fight between Apple (iPhone) versus Google Android? Some advantages of open platforms versus closed:

  • Closed doesn’t require you to sacrifice platform project.
  • Closed allows you to selectively contract. To this end, private subcontracts dominate when developer pool is small. Open innovation dominates as network effects rise.
  • Open give you network effects from low cost experimentation, transparency, lack of hold-up.
  • Open allows users to see, modify or redistribute code which can lead to network effect benefits.

When does open beat closed? Originally, Apple was a lot more closed than Microsoft.
Microsoft opened up and a lot more software got developed. Van Alstyne points out that Android has already force Apple to open up some of its iPhone APIs. Another example is Cisco, which allows others to experiment on their platform. Cisco then buys the best ones (usually at market research).

Does competition deter innovation? Traditionally, you reduce innovation because you curb profits and if people can’t profit, they won’t enter. Google supporting Firefox is another example of subverting Microsoft.

Increasing developer competition decreases openness and innovation. Reason – if downstream profits fall, platform sponsor loses interest in subsidizing developers. With less open platform, developer output falls. One good question pointed out that this is non-linear, as companies have virtually no incentive to innovate when there is no innovation, but high competition results in inability to attract sufficient rents. That’s one reason to me why it is scary to be investing in energy right now with some of these models that seem relatively easy to invent around.

Increasing platform competition increases openness and innovation according to Van Alstyne. If a platform sponsor faces direct competition, the marginal value of downstream royalties rise relative to the marginal values of sales. Thus the platform opens! Look no further than the aforementioned Google Android versus iPhone example.

Prisoner’s dilemma keeps developers from cooperating naturally. Both want to develop on one another’s stuff. Everyone wants to use one another’s code to build on, but it becomes a “You give me access first” prisoner’s dilemma. This becomes a multi-party negotiation problem. In this case, you eed platform sponsor to give access to it (Think Microsoft bundling a browser, instant messenger, etc. into their platform). When that happens, the platform value grows greatly.

According to Van Alstyne, optimal choices substantially grow the platform, benefiting the sponsor, users and developers (lots of economic surplus to go around!) Overall, I thought Van Alstyne's talk was one of the best at the conference. He is bringing a real scientific framework to innovation and that is very much needed.

No comments: