You think ERP projects are long and messy? Try data center and application consolidations. HP is a poster child.
See this WSJ article which describes HP's consolidation project from its start in 2005. Now read this Mark Hurd earnings call transcript this week. "We closed the quarter with 2,500 applications running Hewlett-Packard. We’ve got about 65% of our data centers now closed that we will close. We believe by the end of Q3 we’ll be about 90% to 95% of the way through and finish by the fiscal year."
Yes, a big enough project that he gets asked about it every quarterly call, and he often presents about it in industry events. No ERP project, even failures, ever got this much exposure.
Now let it sink in that this just covers HP's internal processes and data centers, not what Mark calls "trade" - all the data centers which run customer stuff - outsourced, hosted, ASP, cloud etc. And let it sink even further that for the average company one of the biggest hurdles to consolidation is the fact that many data centers and applications are outsourced. That the outsourcers will expect significant "early termination" fees - and in many cases just plain resist their stuff being consolidated.
The payback from such consolidation can be huge. Savings from removing redundant applications, easier integration of acquisitions, wall-to-wall integration etc, etc.
But where have we heard that same pitch before? ERP sales pitches starting in mid 90s.
And where did we go wrong? We relied too heavily on expensive external consultants to deliver ERP magic. Years worth. Billions worth.
Outsourcers are lining up to help company boards launch their own data center and application consolidation projects. Even in HP's case only a few of its consultants have ever done one before (most of its consolidation have been done by its internal IT group) . But it did not stop consultants from selling their ERP expertise - and will not stop them from selling their consolidation "credentials".
Caveat emptor. ERP is a rounding error compared to the effort in reversing years of "technology sprawl". It deserves to be done, but it has to be done by an ruthless, empowered, internal IT team. With top level support. Like HP is doing. With an internal team.


Hi,
I'm not working for HP, but had an opportunity to see few large scale consolidations (and ERP projects). You are absolutely right about the scale - consolidation is a big project. On the other hand, even if can appear to be a single project from the outside, from inside it in fact consists of several inter-related projects:
1. Infrastructure migration - move to a few strategic datacentres - this very much reduces costs related to physical datacentre space and the need to provide off-site support. It is very much focused on hardware (selectively virtualising), and common datacentre services - storage, systems management etc, leaving the operating system and apps as-is
2. Running in parallel (before or after) - review business applications, get rid of duplicates. Brave souls standardise on single set of global systems to provide back-office functions, de-facto standardising key processes and reporting of various acquired companies, reducing the application sprawl
3. Standardise software stacks (OS, middleware, DB), implement fancy support and DC automation tools - makes things run better, be more responsive etc, guarantee SLAs
This gets you to the brave new world you described so well on these pages, but in several (withdrawn and painful) steps, rather than at once. I agree that to do this, you absolutely do need top level commitment.
Regards
Jiri
Posted by: Jiri Ludvik | May 21, 2008 at 02:01 PM
Jiri, good point. But you know the whole program office is complex. Each location being consolidated from a DC perspective has an inventory, move/replace, new interfaces, new help desk processes etc. Apps are even more complex. Yes you can cookie cutter may of the migrations after the first few - a cheaper to do with internal staff or contract labor than with outsourcers - but it is complex overall to manage
Posted by: vinnie mirchandani | May 21, 2008 at 02:21 PM