Your main competitors also have vertical offerings. Do you agree that CRM's functional wars are now over and the main differentiator for end users is integration?
I would not agree that there's functional parity. The difference is that if you don't dig three inches down, everything can look the same from an architecture [point of view]. I have to give customers a lot more credit. These deep industry processes are very important. I do agree with you that integration is a key point of the equation. But very few customers are comfortable putting all of their eggs in one basket and saying 'I want one gigantic database.' However, [they say] 'I don't want a thousand points of light, either. I want fewer systems, not more. I don't want one, and I don't want a thousand. I want to be able to have integration that's intelligent between the systems.' Does Siebel have any plans to issue vertical flavors of its hosted software?
I would say to stay tuned on that point. The one thing I would tell you is we're fortunate that we built on a common architecture. If and when we choose to verticalize the hosted product, it's basically a matter of turning off and on certain capabilities. Do companies run a risk with vertical software of relying too heavily on the software to drive business processes instead of carving out a niche or specialty in their market?
There are core processes that if you're in say wireless, banking or pharma that everyone does the same way. These core processes you have to verticalize. From company to company, they're basically the same. On top of that, though, there is a whole range of different processes, some are vertical, some are not, but they tie in very closely with strategy. For example, look at churn management as a process outside of communications. It's well known to the wireless industry, but banks do not manage this well. So the ability to support the process is out there -- and built into the technology -- but it's how you use the technology and if you use the technology that's the secret sauce. Remember, Siebel is not a magic bullet. We can support your ability to make an offer to a customer, but we don't design the offer for you.
Does the 80/20 rule apply to vertical software? Or is an end user's level of customization much lower than 20%?
It can be a lower number. I think the biggest challenge that companies face when they implement our software -- and probably others as well -- is because of their build heritage, they'll take an app and immediately want to start building stuff. Companies that take the time to investigate what is shipping out of the box find that most of what they need is already there, and that 20% is reduced. It's a given, though, that you'll never get to 100% because companies have unique business processes. What have you learned about companies' vertical needs that you didn't know when you first started offering industry-specific software?
The first thing we didn't expect is that industries are more similar than they'd like to admit. One of the things we were fortunate in doing -- this was a little bit of luck versus perfect foresight -- is we built all of our vertical software on the same architecture and meta data repository. This became fortunate [because] trends that we didn't expect overlapped.
For instance, in the automotive industry we went in starting with things like customer roadside assistance and how the OEM interacts with the dealer channel. But captive finance became white hot in 2002 and 2003. Because we have a financial services offering many of the processes around loans, we were able to pull from our banking apps and move them to our automotive apps. Some best practices of one vertical can be applied to another industry and be a brand new, revolutionary concept. Take us into the labs for a minute. Is Siebel's vertical software developed from the ground up or is it built on top of the core Siebel product?
It's like two ends meeting in the middle. The first thing we do is spend time understanding the customer driver. This is something we're continuing to perfect. What is the customer's problem? What are the metrics of goodness that occur if we solve that problem? Reduction of churn? Increase of revenue? Reduction of costs? What are the processes that need to be automated to accomplish those metrics? Then we see what we have already constructed that we can leverage in that case. If we see a particular interest -- say in loyalty features from the airline industry -- we ask 'is this something that can be more broadly applied in other industries?'