Jonathan Schwartz - Case Study: Restructuring IT
Five years ago: Sun thought: "people want the cheapest storage devices, but highest storage reliability - with products that are not reliable."From this effort, out popped ZFS.
In 2005, Sun decided to give it away for free - something that had taken 100s of millions of dollars in development and Sun decided to give the architecture AND the source code away for FREE.After a year, ZFS has been ported into MacOS X (Sun Leopard), ports were done to Linux community - and has gone around the world - essentially teaching people how to use storage, not manage it.
Around 2006-2007, Sun now launched a piece of data storage tech - "Thumper" which had 24T of storage at a cost of about $1/Gb. In two quarters, they went from 0 to $100M in revenues - simply because the community was already prepped for the product offered by the introduction of ZFS to the community.So, there are two reasons to invent the product:
- Attract the customers - if the marginal cost of adding a user is zero, then you want to add as much customers as much as possible.
He then shows the following map: http://www.sysnet/sunwarp.net/maps
Downloaders of the ZFS source code have to "earn" registration - if a company mandates registration, many people will often not do it. But by earning access and agreeing to share the information, the customer gets something (access). in exchange for this registration, Sun gets a "dot".
He then shows a photo of 13M dots who have downloaded the ZFS system.
"Internet makes it easier to attract the customers across the world. Dots tend to cluster around population and academic centers. The dot tells how/where a customer is. After our acquisition of mySQL, we are adding approximately 100K new "dots" a day."Q/A:
Kwerb: terrific - sounds great - questionable on building a sustainable business. Like, isn't this simply another Java-like model?
JSchwartz: licencees of Java on small devices are already making a nine-figure revenue stream. With mySQL, there is already a $100M revenue stream upon purchasing - and we own the intellectual property. With ZFS, very different.
Kwerb: So, it is about distribution?
JSchwartz: Since it is appealing to the developers (being open source) - we learned that the ratio is 1:1000 for developers to users. And, as you drill down further, a small subset of users will become the paying customers. Thus, the commercial clients make money around mission critical support. As I said, with mySQL - prior to purchase, already had a $100M situation. With ZFS, we had a better chance to monetize.
JSchwartz: It is about our class of customer - users who have more time than money: they are the user of free software. In the situation of Facebook - they have much more money than time - thus, they are now a paying licensee. Gotta remember that if the product is not free, the customer will pick someone else. If free, you are in the conversation.
And, just because you are not charging for your product, you can not simply ignore the customer. Just like a media company, you make money from your customers (the advertisers/the corporates), as long as your content is interesting to your customers/viewers. So, by giving away the source code for free, we have been able to grow the customer base. Free and popular tend to go hand-in-hand, but not necessarily.Kwerb: how do you distinguish?
JSchwartz: Retells a sory about a meeting with a large group of corporate CIOs and asked about how many companies had downloaded mySQL. Almost all of the CIOs said "no, not our company". But JSchwartz checked his logs and found that every company had at least 100 downloads to their IPs.
Anyone who can use adoption to drive revenue will succeed in this model.Kwerb: So, how do you change the direction of the supertanker (Sun)?
JSchwartz: lesson learned: can not change immediately. It takes time - need to work on it slowly and methodically, and then suddenly the acceleration occurs.
Shannon Clark: So, you collected all of that information? How did you know what to collect?
JSchwartz: simply, collect everything - you never know when the data will be valuable. But make sure to be transparent - being a little sneaky can have terrible consequences (see Beacon on Facebook).
JSchwartz: We have 14K patents - and we do not sue or are in a legal mood. We have made sure that we own the IP of anything that is woven into the Sun products/OS code. In all of the OS communities - Sun owns all of the IP - and can change the terms at any time.
Some of the commercial companies need a commercial license - so Sun retains the flexibility and are able to control their own intellectual property.Kwerb: Sort of a kind of ju-jit-zsu?
JSchwartz: Simply, you need to control your own intellectual property to be able to have control of your destiny.
Kwerb: Thoughts on Microhoo?
JSchwartz: Well, if the marginal cost of acquisition of a customer in this space is going to zero, and your business is based on getting people to pay for licenses is threatened. It was incredibly difficult for Sun - but changed the company’s situation with customers.
I think we are in the second inning on the Internet services - the first was search, the next was news, and now infrastructure software. But there are plenty of things have not been impacted by network effect yet - but they will.For example: cars - not yet. Carriers have been - they offer the free phones to get you on a plan, banks offer free checks. Major consumer-based products have not been able to do this.
Imagine what would happen if a phone was given that would have a completely free calling plan. While, the consumer benefit is not readily apparent - companies will use analytics to make good money out of the data.