@Jana
Consolidating data silos rarely provides more than a short term benefit and requires a significant investment. Most efforts soon discover that there is not a single view on what the data should be, but a set of overlapping views which are each evolving at different rates. Trying to force all these dynamic views into a single static schema rarely works, and it hurts, a lot.

A more pragmatic approach is to mange the complexity, segregating the data into large pools of stability (core customer data on standard schema) with small pockets of instability (the evolving stuff) between the pools, with an eye to minimising overall cohesion. Then we can take an incremental approach to development/deployment because, as you point out, the schema is always evolving.

@Elmer
It seems that the SaaS vendors are trying to reuse the play books from the boxed software vendors. By trying to be the “one stop solution shop” they are forcing their customers to cram their data and processes into the Salesforce.com meat grinder.

This doesn’t make sense to me, but then I’ve always seen SaaS as an asset management tool, where I’m getting the SaaS vendor to manage my comoditized data asset; the stable pool of data from my comment above (which is why I put this comment second). The rapidly evolving, important data, is something I should maintain myself, perhaps in a private cloud[1], and the whole lot brought together with something like, I expect , Corizon.

More on this thoughts on a comment I left on Martijn’s blog[2] (he left a comment on this post on the Headshift blog[3])

@Lee I expect I should take all of the above, and the comment in Martijn’s post, and make it into a post of my own

r.

PEG
–
1. http://peter.evans-greenwood.com/2010/01/12/reducing-costs-is-not-the-only-benefit-of-cloud-computing-saas/
2. http://www.martijnlinssen.com/2010/01/it-utility-and-cloud-and-why.html
3. http://www.headshift.com/blog/2010/01/data-ubiquity-threatening-usef.php

meaning that companies are continuously evolving, so the requirements you make at the beginning of a three years project, can often became irrelevant after a year, so we could just accept that IT landscape and architecture from large corporations are just a part of how the company works. companies evolve, so do IT landscapes…

Do you think that this makes sense?

In the meantime the problem only gets worse. As time goes on, records accumulate, new products are launched, and more bandaids are applied to the systems. The end result is something far uglier than if you bite the bullet and just do it (even if it isn’t perfect).

We see the bad effects of some of the processes you discuss in the SalesForce-ification of sales and marketing. Everybody is basically using the same databases and running them through the same CRM systems with the unintended consequence of making it all worthless, hurting reputations and “hardening” the defenses of targets.

We posted on Twitter a good article on how targeted media is less productive since everyone is chasing the same niches. The “piling-on” phenomena is in effect big time. There are additional deep structural issues here.

We find brain and group information processing helpful in looking at what you describe.