Search and the Poetic Giant Stinking Mess
November 21, 2012
I did not craft that elegant phrase “giant stinking mess.” I am not Shakespeare of software. Turn to “It’s Not Just HP And Autonomy, The Enterprise Software Space Is A Giant Stinking Mess.” The article is a good example of a meta-play. One takes a newsy item like Hewlett Packard’s realization that it may have overpaid for Autonomy, watched the founders exit, and then figured out that enterprise search is not quite what it seems. The idea is that enterprise software is going “social” and that the data science behind workflow is the future. The article told me:
Berkholz’s [an expert at RedMonk] post reflects how not all is rotten in the enterprise world. EMC is taking steps to adapt to the new collaborative market. It’s also evident at SAP, VMware and even in some quarters at HP. But the cult of sales still looms over these big companies. Breaking down that culture means creating a new dynamic that embraces modern social and collaborative practices.
Okay, I agree that looking at a particular issue from a different elevation is useful. Let’s assume that social and a collaborative market is the future.
I just wrapped up 13 or 14 reports for IDC. I focused on open source enterprise search. What I learned was that it is getting tough to figure out where an open source search company and an proprietary search company differ. The most successful of the open source vendors look quite a bit like traditional software vendors. One open source vendor in my report—IBM—is a proprietary outfit which uses open source search technology. More interesting is that IBM keeps its arms around its traditional business model. The “new” IBM is not much different from the “old” IBM. Open source software allows IBM to shave off some costs and deploy expensive engineers in what seems to be higher value work.
The question is, “Why do open source search vendors drift toward the traditional business models?” My opinion is that these business models produce revenue and yield margins when they work. What are the elements of a traditional software business model for the enterprise? Those which come to mind include:
- License fees for something—software, upgrades, support
- Variable fees for some other things—engineering services, specialized code widgets, access to previews at a slick looking lab
- Box office tickets for training, webinars, etc.
- Premier services so that the best engineers respond right away to a problem
- A surcharge for working with a pre-eminent firm
- Options like cloud services, appliances, remote optimization
A software company has to produce revenue and my hunch is that this line up of options for a business model exist because customers want these services. When a software company has to generate revenue, the traditional business model is something that investors and stakeholders understand.
One can pop up a level and invoke social, collaboration, and even open sourciness. At the half time news break, the talk turns to booking revenue. Deals will be crafted which meet the resources the client has available. Are these deals convoluted and opaque? Do accountants write haikus? Not too often. Enterprise software, even when delivered from something as wonderful as the Amazon cloud, can become hideously complex and fraught with Byzantine pricing.
I like popping up a level. Revenue generation has a way of bringing some of these viewpoints down to the ground zero.
Stephen E Arnold, November 21, 2012