Thursday, May 15, 2008

Reputation, SaaS and Marketplaces ...

Open sourced standards provide a mechanism for portability and are hence a necessary part of any utility computing marketplace. A first step towards this portable world is Google's Open SDK, however to explore this theme more I'd like to take a look at Bungee Labs' service.

Bungee Labs provides a framework for its customers to develop and release applications into a computing "cloud" that is managed by Bungee Labs. It is provided as a service rather than as a product. The customer benefits by not having to worry about infrastructure and also from only paying for what they use. The customer's only concerns revolve around their data, the applications they've written and how locked-in to this environment they are. The latter is a fairly major issue to many potential consumers, who may be reluctant to use such services without choice in providers and second sourcing options.

Let's now hypothesize that Bungee Labs open sources their environment under GPLv3. Now rather than giving away their competitive advantage, this would instead be a bold yet risky move (all innovations are) to capture a much larger market.

As soon as they open source, then either it will be ignored or other providers could decide to offer this same service based upon this new would-be open sourced standard. We would now have the making of a marketplace with potential portability between providers, a much more attractive option to any potential consumer. Open source is also the fastest way to create an emerging standard by enabling adoption by others.

An important but not to be forgotten side effect of this is that it also allows businesses to become familiar with the service running on their own machines, until such time as they are comfortable to shift into the cloud or, in this case, into the market.

By using an open source approach, a company would hope to gain a small piece of a larger pie based upon the open sourced standard which it has expertise in. If the approach succeeds then the company would in effect trade ownership of a closed locked-in environment for influence in a much bigger marketplace.

Q. How can you influence a marketplace based upon open source?
Q. How do you stop someone branching off in a marketplace based upon an open sourced standard?

With GPLv3 code you can't stop them doing this. However licensing is the wrong way to attempt to control such a market; it's a hangover from too much product focus. What you need to use are the two powerful mechanisms of reputation and trademarks. By establishing a trademark and offering trademarked images to all service providers who comply with the GPLv3 version of the open sourced standard through a testing or compatibility service (e.g. remote testing of compliance to a basic set of primitives which define the service) then in one swift move a company could:-
  1. Ensure compliance information is given to the end user (trademark)
  2. Create a method of enforcing compliance (testing suite and enforcement of trademark)
  3. Create an emerging utility computing market based upon portability and open source.
  4. Establish themselves as a compliance authority (along with all the lucrative service and other revenue streams available)
  5. Encourage operational competition of providers in the market as opposed to functional differentiation (SaaS loophole)
The trademark needs to recognised and trusted, hence reputation is key to this. It is critical that trademarks are enforced.

This idea was essential to Zimki and the formation of competitive markets of providers in this space. An open source system which allows for maximum competition (through operational improvements which don't have to returned to the core because of the SaaS loophole) but ensures interoperability by provision of trademarked images (Zimki Compatible) for providers complying to a testing service.

By open sourcing a platform (such as BungeeLabs'), combining this with a channel program to encourage other competitors to setup and establishing a trademark system for reputable providers, you could start to not only create but influence a marketplace in the Software as a Service world. This marketplace would offer portability and choice for the consumer, a compelling argument against the predominantly locked-in world of SaaS today.

Once we start getting into the nitty gritty of portability in the SaaS world and providers start to overcome their product mentality of the past, then I expect that the establishment of such trademarks and compliance authorities will become a major battleground in the next few years. Being that authority in a world which is heading towards computer exchanges and brokerages is of significant value.


Rev said...

I did not see a trackback from my response on this, so please remove this if it ends up a repeat.

Thanks for your insights on this. It's a matter of specific interest to me.

swardley said...

Hi Ted,

Thank you for your comment, I read it with much interest.

OK, to explain my view, we may as well begin this story at the beginning.

Back at EuroFoo 2004, I gave two talks - one on commoditisation of manufacturing processes (3D printing) and one on the commoditisation & worth of IT. I was already interested at that time on exploiting the commoditisation of IT and in 2005 an opportunity presented itself and we started building Zimki.

Besides the educational hurdles to be overcome, the main issues regarding Zimki were lock-in to an environment and how to grow it quickly whilst disrupting any large company play. This became paramount when shortly after Zimki was live, Amazon launched its S3 service.

So in 2006 I formulated a strategy designed to create a competitive utility computing market. The approach was to gain a small part of a large pie with strong influence over that market based upon our expertise. Open sourcing the platform was a critical part of this plan but we also had to make sure the technology allowed for portability.

Not everyone was comfortable with this idea, and this was an open source company. The biggest danger cited with this approach was branching and how do you ensure that portability works.

I examined the issues of licensing and at first thought the "Affero" would be the most suitable for the reasons you state. The licenses were still being finalised at that time.

However "Affero" enforced that changes are to be released back and this could possibly discourage use. It could also discourage companies creating operational improvements to the system, i.e. keeping to the primitives of the platform but improving the code efficiency.

What we wanted to do was to create a marketplace, based upon portability and encourage competition. So in fact we wanted companies to improve their version of the system, as long as portability between providers was guaranteed.

After a number of discussions in 2006/07 it became clear, that an alternative route was possible which had additional benefits.

By releasing the entire system GPLv3, we obviously would allow people to install, modify and improve the system without returning the code. There would be no barriers to adoption and competition.

However, I planned to also :-

1. Release a portal providing a list of service providers.

2. Implement a partner program to encourage new competitors to set-up with our open sourced system.

3. Introduce compliance monitoring of any competing service and comparison of the service to our open source standard.

4. Use a range of trademarked images to provide end consumers with information on compliance and portability.

Any competitor who matched all the primitives in the open sourced standard (which could be tested) was welcome to use our trademarked image of a "standard provider". We would also enforce and advertise the benefits of this trademark.

Pass our compliance test (i.e. change none of the primitives and how they functioned), allow portability and you could use the image.

For the consumer, it guaranteed that their code and data would work on any provider who had the trademarked image.

For the provider, they could still make operational service improvements in the code and keep that competitive advantage if they wished.

The "standard providers" would be listed in our portal and we would encourage and direct customers to use their service as much as our own. The portal was essential for exploiting network effects.

The intention back in 2006 / early 2007 was that once enough service providers were up and running, we would then withdraw from the utility sales business and focus on the compliance, exchange, switching and brokerage services.

That was my intention and why we were open sourcing in 2007. Unfortunately it never happened.

I wrestled with the "Affero" license for some time, but in the end it became clear that the GPLv3 and the use of the trademarks was a more powerful approach.

You can't control such a market through licensing and you shouldn't seek to do so.

The "SaaS loophole" in GPLv3 encourages competition in such a SaaS marketplace - it turns out to be a wonderful thing. That's why when I met Eben Moglen at OSBC in 2007, I urged him not to close the loophole.

However, that was back then. The danger of a company attempting to out bid you in the open source world today is higher but given that your company has the expertise in this field, the risk is still relatively small.

I would strongly encourage you to consider the use of GPLv3, Trademarks and a partner program to encourage competitors to yourself.

I would encourage you to go for the marketplace. There must be many ISPs today wondering how are they going to compete against the new giants of the service world.

Give them a market and show them how they can fight the big boys as a market. It is a far more attractive option to consumers than any single company can ever offer.

There is always risk, it is a bold move but the potential rewards are far greater than most people can imagine.

I don't like making predictions, and fortunately in this case I can just use history in both the electricity and banking world.

Whoever captures the overall compliance and exchange market in this utility computing world, is the next Google.


Simon W

Rev said...

While not hitting your core points on your reply, I have put up two more tangentally-related think-pieces:

Long Live the SaaS Loophole

Is AGPLv3 Too Radioactive?


swardley said...

Hi Ted,

Interesting posts. I'll put together a reply over the next few days.