Nick Carr, Neil Macheiter, the blogger swardley and others have commented on my recent Harvard Business Review article, the full text of which is available for free online through November. The article presents a model that non-techie business leaders can use to determine what new IT (if any) they need, and what they need to do to extract maximum value from the technologies they purchase.
Many commenters have pointed out that all models are simplifications, but that ones that are oversimplifications can be useless or harmful. I don’t think my categorization into function, enterprise, and network IT (FIT, EIT, and NIT, respectively) is an oversimplification. There remain pretty sharp differences between applications that facilitate tasks (FIT), and those that facilitate interactions (EIT and NIT). There are also pretty sharp differences between enterprise technologies that pre-define interactions (i.e. that impose new cross-functional business processes), and network technologies that try hard not to (i.e. that let people and groups find their own ways to work together over time.).
Space limitations in the article prevented me from spending time on category-spanning technologies. Like Carr, I think such tools are potentially very important and I’ve written about them in a few blog posts. Google Docs and Spreadsheets, its acquisition of JotSpot, and the recent Office 2.0 conference all show that the boundary between FIT and NIT is getting quite blurry in some areas. Generating a best practices document or a financial model is sometimes individual work and sometimes unstructured group work; new tools will let users choose between these two modes as appropriate.
As I said at the Office 2.0 conference, however, the non-boundary spanning technologies are also very important, and are going to remain so. In particular, ‘pure’ EIT and NIT are going to continue to be important; there will be times when business leaders want to use technology to impose lots of structure (for example, when they want to completely specify an end-to-end business process), and times when they want structure to emerge entirely on its own (for example, when they want to set up a prediction market).
Carr makes the excellent point that a lot of companies have stopped treating CRM as an enterprise technology and started treating it as a function technology that helps individual salespeople do their jobs better. In other words, they stopped using it to impose cross-functional interactions, and started using it to facilitate the work of a single group of employees, namely the salesforce.
I actually see this example as supporting my categorization, not undermining it. The label ‘CRM’ has become so broad as to be useless, because it now encompasses both task-facilitating technologies (the currently popular ones) and interaction-facilitating technologies (the historically less successful ones). So my point is that business leaders should ignore vendor- and media-defined technology labels, and focus instead on what they want IT to do for them — facilitate the work of a single group (FIT), impose structured interactions / business process across multiple groups (EIT), or let groups interact without imposing structure on them (NIT). I did obscure this point a bit, however, by blithely including ‘CRM’ as a clear-cut example of Enterprise IT in the table in my article. As Carr points out, it’s not nearly so clear cut any more.
As a few people have pointed out, the classification into FIT, EIT, and NIT is only part of the article. I tried hard to make sure that piece was more than what Carr calls a "dry scholarly exercise in categorization" (ow). The heart of it is a capabilities-complements-responsibilities framework intended to make IT more comprehensible to business leaders, and show them where and how they fit in to IT efforts and contribute to their success.
I stress that each IT category offers successful adopters a small and stable set of capabilities, and that these capabilities are difficult or impossible to acquire without IT. Capabilities are also non-overlapping; EIT does not do the same things that NIT and FIT do.
I highlight that what differentiates the three categories is not what goes on behind the screen (i.e. technical details of the systems) but instead what happens in front of the screen — how the technology changes companies as they adopt and exploit it. Drawing on previous research I identify four organizational complements of IT: new skills, greater interdependence, new workflows, and re-allocated decision rights, and show how each IT category has a different relationship with these complements.
I recommend that business leaders accept three IT-related responsibilities: selection (picking appropriate new tech investments), adoption (getting the technology ready for first use) and exploitation (maximizing value delivered once the system is up and running). Rules of thumb are:
- During selection, work ‘inside-out’ by first determining what IT-based capabilities are required, then looking out at the technology landscape. This is in contrast to an ‘outside-in’ approach that attempts to determine IT needs by examining available technologies and deciding which ones to bring in.
- During adoption and exploitation, put in place the organizational complements that will maximize the positive impact of the selected technology. With EIT, the bulk of this work takes place during adoption. With FIT and NIT, exploitation is the more important responsibility.
One of my main goals with the article was to move past the true-but-tired old saw that "IT-based change is hard." I wanted to show that not all technology efforts are equally difficult, or difficult in the same ways. Thinking in terms of complements helps uncover where the change management challenges will lie. Thinking in terms of capabilities helps determine which IT projects will be worth the managerial effort required for success, and which will not.
The starting point for all of this is thinking in terms of IT categories. I agree completely with Carr that corporate IT is exciting these days, offering lots of new possibilities. I disagree strongly, though, that current trends "get beyond old dysfunctional categories." I believe, in fact, that systematic category-based thinking about IT is the best way to understand which of the new developments are most important for your business, and how to harness them.
{ 1 comment… read it below or add one }
Although there are a few good ideas and concepts in this article, it reads like a text book. According to a blog by Nick Carr, Andrews usually “comes at it fresh”. That is true of all the online writings and speaks but not about this article. This article has the stench of all texts books from the back of the library. Just look at the examples to support his “new” ideas. Nothing about this article is “Fresh”.
The idea of three areas, FIT, NIT and EIT, is good but simplifies the realities of IT. Online CRMÂ’s such as salesforce.com fit into all there of these areas. First, FIT, it becomes a useful and standard application for Sales, Marketing and management. Second, it facilitates online collaboration with structure between employees and departments. Third, and finally, Salesforce.com allows EIT through standardization and collaboration across departments.
That text book smell – Many of the examples given here are from 2004 and earlier. That was a lifetime ago. In real-world terms today, what Cisco and CVS did is straight out of the ice age. For Cisco to take three years to reduce their ERP systems from 9 to 1 is a waste of time and money. Today, there are better technologies and solutions to help companies make better decisions. Where are those examples?
The biggest and simplest message that the article has to offer is the idea of knowing business goals and developing IT strategies that support these goals. That is true and should be exercised by every company in the world. Unfortunately, many companies do implement new technologies based on what other companies, analysts and journalists think. In the past year alone, hundreds of CIO’s have told me that the biggest reason for implementing SOA is that others have told them that they need to. Whether you call it outside-in or inside-out – that stench will linger for a long time.
{ 1 trackback }