Wednesday, 2 July 2008
Failing from the word GO
At the Burton Catalyst conference last week, there was a fair amount of talk about failed provisioning and IdM projects. So much so, it prompted me to post an observation.
Many large projects fail to achieve their initial definition of success, not just IdM projects. But we appear to have more than our share of failures. If a customer doesn't take the time to understand the problem enough such that they can define a practically achievable solution, vendors will take advantage of them.
Many centralized strategies, nearly by virtue of their large and complex scope, fail to ever realize their initial definition of success. When is the last time someone defined success in a provision project to simply connecting two data-repositories? These centralized strategies are often not done until everything is centralized, and that's nearly impossible to achieve in today's large and dynamic enterprise environment.
And here is where the vendor community, especially the large suite vendors, combined with the integrator channel, exasperate the problem. If a vendors route to a customer is through an integrator who adds-value by unraveling complex products, someone who is responsible to make all the complexity work, then why would that vendor feel compelled to make their software easier to install, integrate and use? After all, they rely upon their channel partner to make it all work for them?
And if your sales organization is motivated to sell larger and larger deals, or worse yet, enterprise-wide licenses, then there is a motivation to cover more use-cases (instead of making existing ones easier to implement), and so vendors build larger and larger products, which of course are even more complex. It's all fine if someone else can make your complicated product work, but what if they can't? And when success in implementation is defined in years, what are the odds that the person responsible for starting a project is actually still around to see the project completed?
Maintaining quality and simplicity as use-cases grow is a real challenge for vendors. Maintaining achievable scope for customers is similarly challenging, but a requirement if expectations are to be met. Selecting an architectural approach which rewards quick, tactical wins on the way towards larger strategic objects can help, and that's one of the benefits of the federated identity (decentralized) approach. Customers can celebrate a win one connection at a time.
