Yesterday I had the opportunity to sit in on a conference call with the Irregulars - a group of bloggers who all write about technology for the enterprise. In my case, that would be for the Small and Medium Enterprise, the SME. The topic was sf.com's AppExcahnge. Here's what I learned and some ideas about what all this means for you, and me. First, what is AppExchange? Near as I can tell it encompasses several different technologies. For example, the ability to create custom records and fields in the application to meet the requirements of your specific business model. Also, it includes a published API (an API is an Application Programming Interface, a piece of software code that acts as a door to developers, enabling them to enter another database and find what they need and leave, or, alternatively, to enter and bring what they want inside - a new record perhaps - and place it in the database according to the house rules). The API is the really important piece because it allows other software applications to work with sf.com. For example, you can use it to build a bridge between the sales data and your order management system, since sf.com has no order management system (ok, that was the last dig, but I had to get one in; at any rate, how do salespeople live without order management?). There are some other more technical bits and pieces but essentially the AppExchange is these two things - Customization of sf.com and the ability to write an interface, from sf.com or to sf.com. Stepping back for just a moment, it seems like a lot of hoopla - AppExchange makes a lot of noise in the press - for such a small contribution to technology. Also, the way they have bundled all of these different things together tells me that at the end of the day AppExchange is finally a marketing construct more than anything else. NetSuite does exactly the same thing with its NetFlex. It is really just a bundle of system customization functionality and a published API that can be used for writing an interface with NetSuite. It does not receive nearly the press that sf.com's bundle does, but there is really no difference whatsoever. OK, so what does all of this mean for the average SME wondering how these new technologies are going to help their business? Let's put AppExchange and NetFlex into a technology context first. Many years ago large businesses built systems by purchasing parts and assembling them into a whole. For example, they purchased Cullinet for AP and General Ledger; they often wrote their own order management, billing and AR system because these are very business model centric, and they purchased an MRP system (there were many of these) and Manugistics for running operations. Then they cobbled these together, along with several other minor systems, like HR and Payroll. This approach was called Best of Breed. It was an expensive pain in the ass. Interfaces require a lot of effort, constantly. Think how hard it would be to create a translation tool from Spanish to English. As soon as the speaker rearranges the words in a slightly different way, or throws in a new slang word the translation fails. Same outcome with systems. Interfaces are broken more than 50% of the time. They still work, but they need a lot of hand holding. Then Oracle and SAP came up with the idea of putting the whole shebang into one system. Everything, they thought. And most of the world's large organizations, suffering along with their interfaces, bought into the idea wholeheartedly. Make no mistake, almost every large organization in our world today runs itself on an integrated suite of applications. So what the difference between integrated and interfaced. Simple. In an integrated system the vendor has created a single database that includes tables for every function in the Organization, from finance to payroll to operations to supply chain. All of your core business functions are covered. The benefit is that you will never need to write and maintain an interface between different systems. The vendor, for example, has included a single table that hold the main customer record, and everyone from finance to operations uses that table to process transactions. The cost is that you are limited to the functionality offered by a single vendor for all of your business functions. For example, Manufacturing and HR must both wait for the same vendor to roll out new functions that they need to run efficiently. Yeah, it could take a while. Still, no more interfaces. Both sf.com and NetSuite were born from people who worked at Oracle, like myself. And believe me, if there was one idea that everyone at Oracle knew on their first day it was INTEGRATION. No one championed the Integrated System more than Larry Ellison and Oracle. And for good reason. Where are Cullinet and Manugistics today? They have both been buried, as have 100s of other best of breed offerings. If you come from this background you know about the Integration vS Interface battle, and you know who won. So why does sf.com push AppExchange like it does? Well, it does not want to become another Manugistics, that's why. Unless it can create a huge body of interfaced third party applications, a critical mass really, then it may suffer the same fate. Its early success could turn out to be its demise. How's that? Consider this: sf.com was clever to go after the sales and marketing first because Software as a Service, SaaS, was an unknown entity and not at all accepted by the conservative folks over in finance and accounting. But for the sales group most companies understood the appeal, and prospect data, though important, does not compromise the company. But as SaaS gains more traction, and it will with Microsoft, Oracle, and SAP all moving in this direction, and NetSuite already there, companies eventually will run their entire enterprise on this model. They will get out of the Technology business and concentrate on the information: Taking the T out of IT is the strongest argument for SaaS. When they do, will they look for an integrated suite? Or a best of breed amalgamation? Some people whom I respect in this business look to the future as if the past does not exist. They recite all kinds of ideas about how the next iteration of the Interface will not include all the weaknesses of the past. They have new tools, like Web Services and Service Oriented Architecture, SOA, to back up their claims. For me the argument does not wash. I understand that some interfaces are always going to be necessary. No system will or should include every possible requirement of today's complex businesses. But for core functions like sales, order management, finance, hr and payroll, you need an integrated system. You do not want to fall down when operations cannot see the sales pipeline correctly because of a failed interface. Putting together a bunch of applications via web services and published APIs excites some folks because they see the $$ signs in technical work. Others like the idea of the 'mash-up' as it is called because it sounds like great fun. But for me, I know that I will see other interfaces in my life but I'll be damned if I want to celebrate them. Do them when they are the only option left on the table, not as a matter of course. Spend your hard fought earnings on something with a realizable rate of return. An interface is a money pit with little future, or present, value. If you would like to read other takes by the Enterprise Bloggers you can check out http://enterprise.crispynews.com for their views. |
