Rob's Blog

24/7: $4.5m Oracle ERP Project dumped for Legacy IBM Lotus Notes/Domino and IBM UniVerse Systems

My background since I qualified as a Chemical Engineer back in the days when a Commodore 64 had plenty of RAM has been bespoke software development for Engineering Contracting Companies - Oil & Gas and more recently Water.  Over the years, I have had a hand in the development of most of the modules required to operate these companies apart from the core Financials modules.  The emphasis of the systems has always been projects.  The company tenders for projects and takes responsibility for building a water treatment works or whatever - not just designing it.

Much of the software has been developed in a DBMS called IBM UniVerse - or at least IBM acquired the company that acquired the company that acquired the company that owned Prime Information.  It is a DBMS that supports multi-values fields and variable length fields rather like a Lotus Notes database.  It is clearly not a relational database although it can operate like one.  So it is great for ad hoc SQL type queries unlike Notes which is great if you've got a pre-built view.

In 1996 we introduced Lotus Notes into the mix and introduced additional features to the systems like workflow.  Lotus Notes/Domino and UniVerse work very well together partly because of the similarities in their architecture.

The systems could be operated with any Finance Systems that had open interfaces.  Between 1992 and 2008 the company used SUN Accounts licenced from a company called Systems Union.

As mentioned here and here, I was on the team that migrated to Oracle eBusiness Suite ERP away from SUN Finance and our in-house developed systems.

Well Oracle EBS suite was not a good fit for this company.  It ended up like a millstone around their neck and they have reluctantly taken the decision to go back to the legacy systems I wrote.  The $4.5m cost of the migration has been wasted.  I hope they learn from the experience but it was an expensive lesson.  I hate to say "I told you so" but it's in those blog posts.

So I start next week on the return journey.  What we could have achieved enhancing those systems with just 10% of that cost doesn't bear thinking about.  It is a crying shame.

I am trying to come up with a project name for the migration back into the "legacy" apps.  So far I've come up with Project Karma and Project Phoenix.  Any better suggestions?

30/6: Busy Busy

Since returning from the ILUG at Dublin just over a month ago I have been really busy with one of my clients.  Their US parent decided last Christmas to implement Oracle's eBusiness suite ERP across the group and to start in the UK.  The scope of the projects includes the Financials modules (General Ledger, Accounts Payable, Accounts Receivable, Fixed Assets and Cash Management), Human Resources, Time & Labour, Procurement, Expenses and - last but not least - the Projects modules.

In the UK, this implementation will replace software written using IBM's UniVerse DBMS and Lotus Notes/Domino mostly written by me over the last twenty years plus a bought in Finance system called SunAccounts from Systems Union.  So I have mixed feelings about this.  They have a fully bespoke ERP written specifically for them and their US parent has decided to rip and replace it.  This is personal.

The methodology used to specify the system by the Oracle Contractors is totally alien to me and not something I'm likely to use in the future.  However, I am learning quite a bit about how the modules all interface together and will be able to use that knowledge in the future.  I am travelling 60 miles to work and 60 miles back most days which I find tiring and especially difficult last week when the motorway was closed because of the flooding.

By 1st April 2008 (yes April Fool's Day) the company intends to go live with the new systems.  However successful the implementation project goes between now and then, I believe that the new software will ultimately fail to improve the company's business.  The investment cost of the project coupled with the ongoing additional costs in maintaining the infrastructure for it will never be returned by efficiency savings or increased revenue.  In addition, however, there are big risks involved in the data migration required to drive the Projects.

Data must be migrated out of the existing UniVerse databases into the Oracle tables.  The Oracle Projects suite works completely differently to the existing systems so the data migration is complex.  To migrate data to the Projects suite, a completely bespoke interface will need to be written using the custom APIs which are basically a large number of stored procedures.  There are likely to be around 700 projects that will need their data migrating.

Once the UK has its systems in place, the ERP will be implemented in Ireland, Mexico, Brazil, USA, Canada, Australia and a few smaller installations elsewhere.  However, I expect that there'll be some additional work to be done in the UK before these projects get going.

This may be one of my last projects for this client.  I feel like a turkey voting for Christmas.

19/4: Implementing Oracle eBusiness Suite

One of my clients has decided to implement Oracle eBusiness Suite throughout their business.  Their UK company has bespoke systems written using IBM's UniVerse DBMS and well as in Lotus Notes/Domino.  They have other companies based in the US, Canada, Mexico, Ireland and Australia amongst others.  The intention is to use global best practices yet adopt Oracle with minimal customisation.

This is quite a challenge especially as the intention is to go live with all systems at the same time  in the UK and Ireland later this year.  Other countries will then follow.

I have personally written many of the systems that are being replaced both in UniVerse and in Lotus Notes and my company will be helping with the migration.

I will be taking some training for both the Human Resources and Timesheets modules and liaising with the Oracle Consultants who actually configure the applications.  I am also responsible for the data migration strategy and for actually making data available from the existing systems.

I am quite looking forward to the experience.  I will learn a lot of new technology and will be able to apply it in the future with new clients.  I will also be able to adapt some of my existing products to integrate better with the Oracle ERP.

Do I see a quick return on investment?  Absolutely not.