Understanding cooperative innovation David Clark MIT CFP

advertisement
Understanding
cooperative innovation
David Clark
MIT CFP
November, 2012
The limits of disruption
• Small companies do not usually change the
world.
• Usually, small companies must take the
world as they find it, identify a niche, and
move in.
– Of course, there are exceptions—the true
success stories.
• Google, Facebook (perhaps), Twitter (perhaps)
– And then they are no longer new entrants. They
become the new incumbents.
How do big firms innovate?
• One answer: buy small firms.
– Let them take the risk, and buy the successful ones.
– But this only gets you “small company” innovation.
• Not enough if you dominate the market.
– If you own the pie, to grow you have to make the pie
bigger.
– This means change the world.
• Cisco’s Chambers, several years ago, said we can
no longer grow by M&A.
Making new opportunities
• Sometimes one firm can shift the landscape.
– Hypothesis: this is not often true.
• Multiple actors must align to create new
business opportunities (and new social
value).
• We are calling this cooperative innovation.
– Cooperation among competitors and along the
value chain to produce new functionality, new
capabilities, new opportunities.
Observations and examples
• Example
– The creation of the secure Web (SSL)
• Required modification of client and server software,
creation of CA system and merchant certificates, the
upgrading of many Web servers, and so on.
• Created e-commerce.
• Observation
– The distinction between entrant and incumbent
may not be very helpful in understanding the
dynamics of cooperative innovation.
Consider some failures
•
•
•
•
Curing spam.
QoS on the open Internet.
Migration to IPv6.
Better security generally.
Spam
• The rise and fall of Goodmail.
– Goodmail proposed a scheme:
• Bulk mailers register with their type of service: opt-in, opt-out,
spam, etc.
• Mail from registered mailers would be forwarded without being
filtered, so long as they comply with their service class.
• Other bulk mail would be filtered and perhaps modified.
• To pay for this, bulk mailers would pay a per-email fee.
– Bulk mailers turned on them.
• MoveOn, in particular.
• Google said the power to filter should be in the hands of the
user.
– Email providers abandoned them.
– They went out of business.
Lesson from Goodmail
• Large email providers have market power.
– The higher-level equivalent of the telco
terminating monopoly.
• A scheme that additionally empowers actors
with such market power will be rejected by
the ecosystem.
QoS
• Tools for QoS do work (technically) but have not been
deployed on the public Internet.
– No model for charging or allocation of revenues.
– No agreement as to what the actual service definition
would be.
• Continued disagreement as to whether QoS is needed.
– Necessity of increased information sharing among
competitors was worrying.
– Sense that partial deployment would be ineffective in
the market.
• (We held several meetings to resolve “some of” these issues.)
– Now, rejection of the idea in favor of managed services
that compete.
Lessons from QoS
• Disagreement about the value of an idea will almost
certainly kill it.
• Necessity for a lot of actors to move together is a
formidable barrier.
– No way to do a “proof of concept” to resolve first issue.
• Difficulty of negotiation with competitors is a killer.
– Business, technical, legal
• Failure of designers to deal with “money routing” is a
killer.
– Need to propose a business model, not just a technical
model.
IPv6
• Perhaps not an innovation, just a necessity.
• Slow progress. Very slow.
• Many ISPs, most OS and service platforms
are capable.
• Next barrier is upgrading all the Web servers.
– Perhaps 200M of them?
• Less than 1% upgraded so far.
Lessons
• Lots of web servers upgraded to SSL, but
very few upgrade to IPv6.
– Not surprising at all. SSL brings them benefit.
– IPv6 is a classic example of an externality.
• I expend, somebody else benefits.
– Too many actors to come up with an easy
scheme to internalize the externality.
• At least, I have not seen the scheme yet.
• ISPs could charge a premium to services that are not
IPv6-ready, if they themselves had the incentive…
Generalities
• Need a leader.
– Should not be the actor with market power.
– Need to hunt for the right actor who is in the right position.
• Need a financial model.
– Need to argue that all actors have some incentive to act.
– Shifting revenues along the value chain to balance incentives
may not be practical.
• Massive transaction costs.
• Need balance of power and control along the value chain.
• Need an approach that allows incremental deployment
and proof of concept.
Next
• We want to identify other case studies and
try to elaborate and better understand
general lessons.
• Can we solicit examples from you, either
historical or forward-looking?
Download