In a recent post, Harvard Business School professor Andrew McAfee discussed findings published earlier in Andy Dornan's Business Week article about the perceived value of Web 2.0 technologies within enterprises. I've already commented on Dornan's article, and won't do so again, but I do want to comment on McAfee's insights relating to Business Week's findings.
Those of you who have already read McAfee's post might wonder what his article has to do with mashups. McAfee's comments relate specifically to legacy Web 2.0 (yes, my tongue is well into my cheek) capabilities such as wikis and blogs. He also extends his insights to social networking sites. (SNS) Although he doesn't specifically discuss mashups, I think what he has to say applies equally well to them. Mashups intrude on application development's turf, and app dev is one of the most isolated and siloed organization within IT. In McAfee's experience, people believe "...that corporate IT departments consciously exclude outsiders and outside influences, and are concerned primarily with expanding themselves." That's true in spades for applications development.
Here's some more of what he has to say.
Enterprise 2.0 tools have no inherent respect for organizational boundaries, hierarchies, or job titles. They facilitate self-organization and emergent rather than imposed structure...They require [grizzled R&D types] to...“practice the philosophy of making it easy to correct mistakes, rather than making it difficult to make them”...They require, in short, the re-examination and often the reversal of many longstanding assumptions and practices...
Do Mashups and mashers respect organizational boundaries, hierarchies or job titles? Nope. In fact, mashers don't even respect IP. It's the essence of a mashup that the masher pulls information from where she wants in a way she wants to create the application she wants. With in-the-cloud deployment models and screen scraping tools, she's not going to care whether IT likes what she's doing. She's not going to go through a project approval cycle. She isn't going to get permission. She's just going to do it.
Do mashups facilitate emergent rather than imposed structure? You bet. I like to say that blogging, combined with Google search, lets anyone become a pundit. You don't need to work for Gartner or Forrester, or even eWeek. Eventually the bloggers with something interesting to say will get found and heard. The same is true for mashups. With business mashup technologies putting the means of production into the hands of subject matter experts, anyone can be a masher. The good mashups will be found and used, turning mashers into valued experts. These mashers don't need to work for IBM, BEA or even your IT department.
Do mashups promote the philosophy of making it easier to correct mistakes rather than harder to make mistakes? They sure do, and this is the very issue that drives IT nuts. For years IT Ops has been focused on keeping the production environment secure, even at the expense of business agility. It's a long and difficult process getting any change into the production environment. IT isn't being obstructive for no good reason. IT has a long history of having their butts firmly kicked for allowing problems to creep into the production environment. Their quarterly bonuses are probably tied to system availability and reliability, not to innovation. App dev organizations have a similar focus. We read about business disasters caused by faulty software. Software defects, found in production, are 100x more expensive to fix than when these defects are found in the requirements or design phase. Application development organizations are told to make sure the software is bullet-proof , defect-free and completely stable before it goes out the door.
Yet with mashups we are asking the business to embrace the notion that software will almost certainly be flawed, but can always be fixed. In IT's mind this philosophy is no different than the bad-old hacker days when cowboy programmers working on the enterprise's 'big apps' could bring down an entire business unit's operations. No wonder application development organizations believe end-user created mashups are a bad idea. What app dev isn't quite yet understanding is that a mashup can be fixed in minutes or hours, whereas 'big apps' take months or perhaps years to fix. Changing the 'big app' mindset will definitely require "the re-examination and often the reversal of many longstanding assumptions and practices."
McAfee concludes that IT had better wake up and start delivering innovative solutions for the business, or they will find themselves taken out by Web 2.0. I'm not quite ready to predict that level of gloom and doom. I will certainly agree that IT needs to evolve. They aren't going to block adoption of Web 2.0 capabilities, to include mashups. What they can do, as I've discussed before, is provide a secure infrastructre and behind-the-scenes mentoring so these new technologies won't harm the business.