IT management is undergoing a tectonic shift. The increasing pace of business, the emergence of cloud computing, the spread of social media, and the general public’s exposure to commercial application stores (Android and Apple) are all converging to a crisis point.

Less obvious but perhaps more fundamental forces include

  1. the increasing influence of Lean theory on IT management,
  2. the continuing strength of the Agile movement (most recently in the form of DevOps and software Kanban), and
  3. the shift in IT finance from capex to opex

My good friend Rob Stroud of CA digs into the evolving IT financial model in a post I much enjoyed. However, his statement that “we will be transitioning IT to the new model of service aggregator or broker” (emphasis added) crystallized some thoughts. My concern is that IT management will think that all IT need do is aggregate services, when in fact it needs to integrate services.

As an enterprise architect in large enterprises, systems integration was always a primary concern. Value is not achieved in a large IT context (or any business context) through simply aggregating functionality, whether internally or externally sourced. The data must flow correctly from system to system, accompanying some end to end business value flow. (See Lean Integration by Informatica’s John Schmidt, another good friend.)

This is one of my biggest concerns in the move to cloud and especially SaaS computing, and some of the loose talk recently around the app store delivery model. How can an application purchased using a iPhone paradigm be expected to integrate with another? My time-wasting copy of Plants vs. Zombies on my iPhone has no clue about ToodleDo. But what if my business model required them to talk to each other?

I suppose I could work with the manufacturers of each to build an adapter. But that would take a long time and (in a small but significant percent of cases) what if such an integration offered a competitive advantage that I don’t want the vendors to commoditize?

Plants vs. Zombies is obviously a silly example, but integrating cloud services is increasingly the kind of problem that IT architects increasingly encounter, for example in the relationship of SalesForce.com to their internal ERP systems. Or between ServiceNow and NetCool.

In short, will a wholesale adoption of SaaS by your firm effectively be a regression to high walled functional silos? And all the redundancy, inefficiency, and ultimately poor service that silos drive? There are so many combinations of integrations between the SaaS offerings, there is no way that the vendors will be able to step up and offer every possibility. And the quest for competitive advantage will keep the business looking exactly for those kinds of integrations that are not commoditized.

In general, the SaaS vendors have credible integration at a technical level, and vendors like MuleSoft are already proposing iPaaS (Integration Platform as a Service). But again, if senior IT and business management believe that all such concerns are miraculously now behind them… they need to be reminded that integrating services in novel ways will be one important way that enterprises compete.

Beyond integration, I think there is danger in the “specify and source” model – just write up the requirements and put them out to the market to bid on. Specifications are always incomplete, unless for uninteresting commodities like a payroll system, or generic infrastructure that in and of itself provides no direct business value. Interesting, competitive functionality must be defined iteratively, in tight partnership between the customer and the provider, and value is in the system as a whole.

The idea that a sufficient, large scale specification need only be well crafted and perfect services will manifest through market magic is misguided, at least in the most business visible parts of IT. Instead, success is found through iterative development, focusing on smaller increments of functionality delivered in a more continuous flow. And as Toyota has demonstrated in manufacturing, this requires long-term, high-trust supplier relationships.

What does this mean, concretely? As businesses move more and and more into the cloud, I think there is little doubt that traditional systems administrators (for example) will increasingly be aggregated into large IaaS and PaaS providers. But if some money is saved by doing this, what will the business do with that money?

Some may go back to shareholders, or to non-IT purposes. But in significant part, I believe it will be used to purchase new kinds of business intelligence visual metaphorIT skills and services needed for competition: in particular, to integrate the component cloud services in support of end to end value flow. Services to continuously enhance core components and adapt them to business needs with agility. Products and services to monitor end to end business performance across multiple components (cloud and internal). And services to integrate data from all of those services together for competitive business intelligence.

Some will say, “integration services and consulting on IT-led transformation can be purchased externally just like component services – in fact, we’ve been doing this for decades.” And this is true. But those will need to be long term relationships in order to be at all valuable – systems integration is hard. And how do you even know you need these integration services unless you’ve got someone skilled enough internally to recognize the need? (Note that these are IT professional services, not computing services.) And finally, do you want those integrators taking what they’ve learned in your organization, over to your competitor? These are old questions that don’t go away.

Contrarian investing is often wise, if you’ve got the courage. While everyone else is rushing to the cloud, perhaps now is the perfect time to strengthen certain key internal IT capabilities you’re going to need for the long haul. Architecture and integration, agile delivery and devops, vendor management, cloud risk management, demand and resource management, and business intelligence would come to mind as the top contenders for truly transformative IT.

Enhanced by Zemanta