By Nikki Baird, Managing Partner
May 11, 2010
Last week I attended Manhattan's Momentum user group conference. At the event, Manhattan unveiled the completion of their transition to an integrated platform, enabled by a service-oriented architecture. Paula touched on some of the immediate benefits coming to distributed order management and some of the inventory and transportation benefits from a platform approach to optimization in her article last week. Personally, I think it's fantastic. I think that anyone looking to buy software should be pushing vendors hard on how much progress they may or may not be making in modernizing their internal architecture - it's one thing to have to deal with the mess that your own internal systems may represent, but the last thing you need is to perpetuate it by investing in technology that can't play in a modern architecture.
Thus my frustration. Manhattan Associates, a moderately sized software vendor, spent six years making its transition to a platform. Along the way, the innovation did not stop - they came up with a world-class distributed order management solution, and developed a unified forecasting method that is a step-level improvement in how organizations should be thinking about the discipline, not to mention continued incremental improvements. Why can't retailers manage this?
Okay, granted, the economy made life hard for a good long while, and there have been other things that have delayed these investments - PCI and Y2K being the biggest money-holes in the last decade. I have heard retailers refer to discussions of architecture or platforms as "the t-word" - like it's a bad, four-letter word that will get their mouths washed out with soap (or get them fired, I guess, I don't know). What's the t-word? Transformation. One vendor told me that they talked to a customer who said that they would rather put their IT money towards discounts on products and just shut down the department. I would hate to be an investor in that company. Why? Because consumers are leading IT departments now.
Here's what I mean - let's just take store technology as but one example. There are others, I assure you. Stores were originally architected for distributed computing. Islands of automation that were not dependent on that "unreliable" network (or, what was actually the case originally - prohibitively expensive network). That is no longer the case. In the developed world, the network is cheap and pervasive. But customers want you to know them no matter where they are. And now that the shopping trip no longer begins and ends in the store, the store is no longer self-sufficient. It is dependent on information and insights that need to come from other channels and other organizations, and all of that needs to be delivered on-demand - when the consumer walks through the front door. A move from distributed autonomy to centrally managed, pervasively connected coordination and collaboration. All because consumers expect it. And if you don't have it, they will merely pull out their mobile phone and work around you - putting their potential purchase into play for competitors to steal.
Here's my challenge to retail: If a company the size and shape of Manhattan can completely modernize its architecture in six years, while at the same time innovating and moving its business forward, why can't retailers do that? I've met too many retail IT departments who avow that they are at least "as good" as a software company when it comes to developing solutions. Oh yeah? Then where's your platform?
Retailers are too locked into an old way of thinking when it comes to IT. Yes, there are exceptions. You are duly acknowledged. But this is a serious problem. I hear from vendors that even when they have the right architecture in place, retailers don't have the skill sets internally to take advantage of things like SOA, or their budget is so tight, or the timeline so ridiculous (because they waited until it was an acute problem) that expediency - and point integration - is the only way to get it done. But I challenge you to take a lesson from Manhattan and its cohorts. Making the move to a modern architecture may be "transformational" but if you've been paying attention, you'll have seen that transformational does not mean big bang. In fact, those modern architectures enable bite-size implementations over time that actually still yield a cohesive whole. They even enable "modifications" that are insulated extension which survive future upgrades.
If you need proof, just keep watching these guys. Here are two things I look for in any vendor that claims they have achieved a modern SOA-based architecture: rapid composite apps and a customer base that is largely current (for new entrants to SOA, obviously this takes awhile). Here's why. If you have a clean, modern architecture, pulling different pieces together in new ways becomes ridiculously easy - almost (ALMOST) drag-and-drop. It's just a call-out to a service in a library of services, after all. So there should be rapid innovation derived from bringing these pieces together in new apps. And, if they have a customer base that has transitioned to the new platform, then the clean architecture should mean that no one gets version-locked ever again - provided they have implemented it cleanly. That means modifications are extensions - new services, call outs to other applications, etc. - not hard changes made to source code, or point to point integrations that are easy to break in an upgrade. Implementing that upgrade should be as easy as hitting "install."
Can your systems do that?
|
