Are SIs Poised for a Golden Age or Imminent Irrelevance?

On a recent plane ride, the fellow sitting next to me fired up a movie shortly after take off.  He was watching the first Matrix.  Agent Smith had just arrived at the apartment where Trinity had been detected.  A cop, dripping with arrogance, informs Smith that his men were bringing her down.  Smith replied, "No.  Your men are already dead."

Many pundits and the SIs themselves believe that the advances made in next generation enterprise architectures like SOA will usher in a golden age of prosperity for the integrator community (at the expense of vendors and other ecosystem players).  Nothing could be farther from the truth. 

• The barriers to application integration will lower.  The first time I ran an Oracle integration project was an intimidating experience.  We just needed add one piece of information and tweak the process in one module, but were faced with hundreds and tables and thousands of views that no one seemed to understand.  SIs have owned that world, sometimes having a better understanding of the ramifications of tweaking a table better than the vendor's own consulting group.  SOA is eroding that advantage.
• Open platforms like AppExchange and Netweaver are creating ecosystems of vendors around a dominant providers core offering.  These vendors are enhancing the value of the host's core offerings.  They are doing so at the expense of SIs.  It used to be the case that SIs made a great living creating niche process extensions around an enterprise software vendors applications.  This has been so lucrative, Accenture has often flirted with the applications business.  Increasingly, niche software providers have leveraged the opening up of platforms to create out of the box, packaged software for the cracks once filled by SI "customization" projects.
• The knowledge required to understand how to modify and repurpose workflows and business processes will be more easily available.  It's only a matter of time before users will create and integrate their own applications.  Imagine a Duet world in which a project that would have kept a team of Bearingpoint consultant's families fed for a year is reduced to adding a column to a custom view in Microsoft Outlook.

I had breakfast this week with an friend and enterprise software executive.  When I mentioned the future of SIs, he chuckled.  He had just finished a phone call with a long time friend and partner with a large SI firm.  The friend had just left his firm.  During a CRM engagement, they had completed a needs analysis for a F500 manufacturer.  It turned out that the best application for their requirements was Salesforce.com.  After heated internal debate, they decided to deploy Siebel instead.  If SIs weren't blinded by their myopia, they would look into the trenches and see that their men are already dead.

Who will kill Enterprise Software?

There have been a number of popular posts on the demise of enterprise software.  The posts are varied and well articulated.  Most make strong cases - and for good reason, anyone watching the consolidation and stagnation of enterprise software the last six years knows something is going wrong.

Most point to the commoditization of the stack by open source providers as the looming crisis.  While it is true that open source is crawling up the stack slowly but surely, there are some valid reasons why Open Source won't be the killer of SAP and Oracle.  Regardless of your thoughts on the probability of SugarCRM replacing SAP at Intel or IBM, this isn't the largest problem faced by  enterprise software vendors.    The is a much larger problem...one that has been fifteen years in the making vs. the five years (more or less) that Open Source has been a significant threat. 

This threat also isn't SaaS.  SaaS is absolutely changing the landscape and empowering the SME market with functionality that they could only read about in CIO magazine (while waiting for Microsoft Excel to load their integrated inventory and order management system).  Most enterprise software companies recognize that eventually some of their customers are going to want to move their cost structure to a more predictable model.  They are busily building out their SaaS offerings blunting the upward mobility of NetSuite and Salesforce.com in the process and banking on the painfully slow culture changes of F500 IT.

The real threat to enterprise software vendors is users...or lack thereof.  Enterprise software users don't use enterprise software.  I'd like to say I have come to this as a startling revelation.  There was no burning bush, however.    I think it would be more aptly put that I have slowly come out of denial.  Early in my career, I imagined that every time a large enterprise bought some small application from a startup, despite the fact that the startup's product was two bullets on the feature list of an app they had already purchased and had running for years, I would think, "Wow...that <fill in the blank> implementation must have gone horribly wrong for them to buy a whole new product for that small piece of the process."  However, I quickly realized that the problem wasn't the functionality, the implementation (though most had gone horribly wrong), or the ROI.  The core problem was that the application got in the way of the user doing his or her job.

While opening the 2glue corporate bank account, I had an interesting experience that serves as a perfect illustration.  I spent over an hour filling out an online application.  While this was a bit annoying, I was happy I wasn't in a branch office, with a paper form.  Once completed and submitted, I shuffled off to bed, happy to have another task checked off my overly long todo list.  The next morning, I had an email waiting for me from the business services group of the bank.  "Congratulations.  Your application has been accepted and processed.  Please fill out the attached acceptance form and fax it to xxx-xxx-xxxx."  I opened the "acceptance form" only to discover it was a near identical copy of the form I had just filled out the night before.  I grumpily completed the form and faxed it as requested.  Rachel, a cheerful banker newly assigned to my account gave me a call the next day to finalize the account opening.  During the course of the conversation, Rachel, sounding slightly perturbed and seemingly talking to herself as much as me, said, "please hold on a minute, I need to re-enter your application into the xxx system."  Intrigued by her frustration level being a close match for my own at the whole process, I inquired further.  "I hate 'that system'  I only use it when I need to open an xxxx account, I usually just use email otherwise."

Having been in enterprise software a dozen years, I seen the trend.  Users don't use enterprise software to get their job done.  They use productivity tool like email, IM, Office, and RSS Aggregators...the same tools they use at home.  It is usually a gray area; not sanctioned by IT, but not at the top of their list of problems with which to deal.  Sales guys copy their leads out of Siebel and into Outlook as fast as they can.  They call them from their cell phones in the airport lounge and don't update CRM until they get yelled at, or need their commission check.  Buyers keep their order and demand planning data in an excel spreadsheet, and don't enter it into SAP until they need a PO sent. 

Enterprise software is a system of record.  It's value in most organizations is the reports it generates.  The real threat to enterprise software is the users.  They're getting their job done...without the app.

Coming soon...2glue

It won't be long now.  2glue has started its private beta.  Soon there will be a new way for individuals to leverage their personal web based productivity tools and their company provided outsourced applications while still maintaining contact with their legacy "land locked" corporate data.

It will be convenient.

It will be empowering.

It won't require IT.

2glue.