Thursday, October 07, 2004

Confessions of an IT Cook: Ingredients for a Service Oriented Architecture

So we have all heard this over and over. Here's the magic receipe for SOA, as per the industry :

  • Multiple business applications, preferably residing in multiple databases
  • Lots of platforms
  • Lack of integrations, needing lots of manual scrubbing and redundant data entry
  • lack of standardized processes
  • Lack of common definition of fundamental constructs like Customer, Product, Item
Steps to a perfect solution:
  • First, identify all discrete business steps within every application
  • Create web services !
  • Now mix all the web services in all the apps together and shake well
  • Make sure that you pour this into a container that can hold all of them - call it the "orchestrator", "glue", "process integrator" or whatever
  • Now slap on various user interfaces on top of this and call them "consumers"
  • Just for fun, sprinkle some "business process tools" on top of this and let your end business users run amok with the combinations, just like a kid in a candy store who wants a mixture of jelly beans !
  • After this, all you need is to examine how great your culinary skills are - just use a "business activity monitor" and if the tempearture drops or rises above thresholds - just set off an "exception workflow" and you are all set!
So - what am I missing ?

IP Telephony and Web Callbacks

An interesting technology gap exists the innovators of IP technology and the adopters in the Enterprise - specifically referring to the multitude of contact centers and helpdesks. As Jon Udell mentions in this article there is a big difference between computer telephony and IP telephony - nowhere bigger is the contrast in the enterprise consumers. The big problem does boil down to the inability to articulate business benefits brought on by cool technology.

Why are vendors like Genesys, Cisco and Avaya still able to convince enterprises to cough up millions of dollars (no kidding there) and still charge the antiquated "per-seat" licensing ? What is per seat when there are virtual call centers, work-from-home agents and the like ? Isn't it only logical for these enterprises to seriously look for open source linux-based solutions that are fundamentally cheaper and more easily tailorable ? No - not yet! The main reasons are as old as the open source v/s commercial battle with one twist - the business leaders in the call center have not yet come to appreciate the benefits of technology.

The argument is that the demographic of future customers for customer support is changing - the workforce of the future will be younger, will demand more channels and more flexibility in how problems get resolved. Why wait on hold for 30 minutes when I can schedule a web callback ? How about IM'ing me when you have resolved a problem ? Send me a text message if you can find me online. These are very plausible scenarios that will not be addressed by enterprises with their wait and watch strategy.

In the meantime, the tech divide will continue to manifest as one of the most annoying problems in civilized society today - customers waiting for a human touch. With the over emphasis on automation in the form of self service, Natural language processing (NLP) etc it's ironic that the same crowd has placed an over-reliance of technology in one area to effectively alienate themselves from the very audience they were expected to serve - the customer.

Sunday, October 03, 2004

Now Amazon gets into Search...

I find it amazing that the equivalent of the OS battles of the past, and the browser wars is now being played out as search wars.

The undisputed king of search of course is Google, but now that they are public - the real question in management's mind would be how do you make this profitable?

One approach that they seem to be taking to become the operating system of the Internet. Which means that they solve the fundamental problem of providing the connectivity and plumbing to retrieve information - become a platform of some sort. All the other companies like Amazon with their latest spinoff A9 is planning to wrap content around the search OS.

And thus starts the grand convergence. Yahoo! Search takes on Google, Microsoft takes on Google and Yahoo!, and content again threatens to be king. Yahoo! has a great headstart from a Portal perspective ( is it just search in a different way?) and is what A9 trying to do here similar to Yahoo! ?

What is the competitive entry barrier for A9 ? In my mind - it will be access to content. Amazon has the proprietary technology that lets you search "Inside the Book" that Yahoo! does not have currently. Comparing apples-to-apples, the winner in this case would be determined by whether you are interested in searching for public domain (a.k.a free) information that is aggregated in a single place with some value added relevancy checks, or copyrighted content (such as books) where there is an e-commerce aspect to the right to view/consume it.

We'll be watching!
Any Email or Collaboration product managers reading this?

Harvard Business Review ran a recent article called Why Your E-mail Requests Are Ignored that claimed the degree of effectiveness of an email is inversely proportional to the number of recepients being addressed to. This means - you have a better chance of response when you send a note out to one person, than to four people.

How many times have you been faced with this situation: You need to send the same email to 10 people, with the same content with perhaps a change of the name, and still follow the effectiveness principle stated above?

Today there are a few options, in increasing degrees of complexity :
a. Copy and paste the name in Notes/Outlook
b. Use fancy mail merge in Word to generate 10 copies
c. Use a CRM app - like PeopleSoft, that has correspondence mgmnt and online marketing for mass email generation

I'm surprised that short of a BCC option which is really ugly, this is a simple mail requirement that has not been handled by any client that I have encountered so far. Have you ?