Zimbra had the good fortune to get tapped for Business 2.0’s Next Net 25 and last week had the chance to go hang out with the Business 2.0 gurus—Erick Schonfeld, Om Malik, and Michael Copeland—as well as with most of the our fellow entrepreneurs that made the cut.
While it was a tough group to get to work well together in one room—lots of opinionated talkers and not that many listeners (myself included :-)), with divergent agendas, half of us focused on the consumer and half focused on businesses, the discussion thread I had the most fun with was what will be the next, next big thing.
Here are my ideas on that topic. I’m afraid most of this handful of problems are somewhere between hard and intractible given the state of technology today. On the other hand (and as was pointed out at the meeting), I’m now a grizzled senior citizen relative to the next generation of high-tech entrepreneurs (Zimbra is start-up #3 for me, and I spent much more time than the average navigating college), and so could easily have become jaded.
(1) Customization, extension & integration for on-demand applications — As more software is delivered on-demand (via Software as a Service/SaaS), we are going to see increasing challenges to how that software can be customized, extended, and integrated with other applications (both on-premises and on-demand). Today, customization tends to be relatively superficial (reskinning for example). Extension can be accomplished by adding new data to existing tables, but then there is always the problem of not being able to host full-on business logic you didn’t write (at least in the general case, there is too much risk of errant or malicious code doing harm). Lastly, the secure and reliable integration of multiple applications across the Internet is not an easily solved problem, and the higher-end web services infrastructure (e.g., WS-Trust, WS-Federation, etc.) being proposed to make this easier is at risk of being too complex to ever achieve widescale adoption. More declarative/meta-data based programming and workflow models seem to be the likely answer here.
(2) Infrastructure for multi-sourcing and change management — As software projects incorporate code from multiple open source stacks as well as other internal re-use efforts, and talk to/interoperate with many other applications and systems (via XML, SOAP, REST, etc.), how is IT to deliver reliable, non-fragile deployments? Some interesting companies chasing this are Gauntlet Systems (WebLogic buddies recently acquired) and SourceLabs.
(3) Mobile applications 2.0 — Phones/PDAs still do not provide a sufficiently unified platform/profile for applications (whether web-oriented such as via Ajax or Flash, or native programmed client, such as via J2ME) to get us to critical mass. PCs are far more consistent devices for web applications at least, even when you take Windows vs. Mac vs. Linux (as well as IE vs. Firefox vs. Safari) into account. Today, you can develop a website for all PCs pretty straightforwardly (albeit that cross platform testing is still required), but we are simply not yet there for mobile applications. Which is too bad, because the network and performance of phones/PDAs is getting to the point that the technology is ready to do more.
(4) Alternative on-line advertising models — While I confess I didn’t suggest this one initially, it struck a chord with me. There is much renewed interest today (remember all those Web 1.0 ad’s before the advent of pop-up blockers?) in leveraging advertising as a means for monetizing on-demand/SaaS deployments (as an alternative to subscriptions). Much of this renewed interest in ad’s comes out of Google’s success I think, but we should keep in mind that Google has found a unique sweet spot—when I am searching for something, I am actually often as interested in the ad’s as I am the general search results. With virtually everything else I do on-line, ad’s are at least a nuisance, and may well interfere with whatever I’m trying to get done. So the key, I think, is to find ways to blend advertising with content that are both unobtrusive and occasionally helpful. For email, IM, calendaring, etc. Zimlets could provide such an alternative model.
There’s my pie-in-the-sky wish list. Love to hear if you have managed to nail one of these. Happy entrepreneuring.