Several excellent threads on this week’s discussion, good job.   There are  three concepts that I think are worthy of highlighting:

• EA is about both business and technology.  It aims to drive the technology decisions from its business findings but its end result are things like application and infrastructure standards.
• EA can become very bureaucratic.  If you are spending time charting and documenting things without providing the output needed to guide the enterprise, its likely a waste of time.  EA is  difficult to do well.
• Best-of-breed vs packaged ERP is a classic EA question. A well reasoned case for either, communicated throughout the company, would be an excellent outcome for an EA project.

I hope you see the difference between the topics of the first two week and EA.  Week 2 and 3 were about defining the IT organization, its mission to produce value for the enterprise, and the internal administrative controls needed to run it effectively and efficiently.  All good, necessary stuff but kind of generic.  You could apply any of it to any organization.

EA, at least its output, is different. This is the first time we are talking about what the IT organization should do, what it might focus on to add that value we talked about generically earlier.  It starts with understanding the business and its processes.  What is important? What isn’t?  A manufacturing firm will have very different needs than a consulting firm.  Just because some type of system is very effective in one industry, doesn’t mean it will work for the other.  EA goes beyond this, however, to look at a future state that has all the right applications and infrastructure in place to ensure the continued success of the company.  You might sum it up as saying EA strives to make the right IT decisions today, with an eye always on tomorrow.

Considering best-of-breed vs ERP type package.  My view is: it depends.  If I am a traditional manufacturer I am apt to go ERP.  My requirements are very well known and expertly supported by several great software companies.  Their tools have been proven to make manufacturers much more efficient.  Efficiency is king in my organization (manufacturers must keep SAR costs under control to be profitable) so this all sounds very good to me.  EA for such a firm will probably set some standards around infrastructure but say for applications “Use SAP (or whatever) first or explain why not.”

On the other hand, if I am an internet investment company, everything I do is IT.  I reach out to my customers through the web, I take their orders through the web, I deliver my work product to them via the web.  IT is my company’s life.  Here I may not want anyone else calling the shots. I want to bring in whatever will do the best job possible for my most important needs and I will hook it all together.  Thats what I do anyway, hook stuff together, so its no big deal.  Here EA is apt to be all about infrastructure and middleware.  So long as an application can live on our servers and communicate with our middleware, we will be fine.

So each firm will have a different EA for their organization and both will be correct in what they choose.

Reading Questions

1. What is the goal of having an enterprise architecture?
2. If a firm decides to add a new line of business, how might it affect its enterprise architecture? Explain?
3. Explain five possible ways that an enterprise architecture effort could fail?

The Strategic IT Transformation at Accenture Case

Think about the following questions as you prepare for Tuesday’s discussion of this case:

1. What is Accenture’s core IT philosophy?
2. Identify three key IT projects from the 2001 – 2208 period and explain how  each strengthened Accenture’s enterprise architecture?
3. What measures of success did Accenture use for this effort? Why?