Saturday, February 02, 2013

Fully Costed Oracle

Who knew that's what FCO actually stood for?  All that time we've been thinking it was simply about diplomats in far flung locations enjoying "unimaginable luxury" and getting up to who knows what.

Late in January, the FCO announced, to predictably widespread criticism, that it was intending to launch a new framework:

...  supporting the Cabinet Office Shared Service strategy ... for the provision of Oracle Enterprise Resource Planning (ERP) development, delivery and support services ... 
The framework intends to have a limited number of vendors, for instance a number of Lots might be awarded to the same vendor. 
The scope intends to cover existing Oracle platforms in UK government departments ... and to include upgrades and implementations of new Oracle versions for these existing platforms. It also intends to cover any move of a Department from a non-Oracle platform to an Oracle platform


The value of the framework is suggested to be £250m to £750m.  The notice is silent on framework duration but others have suggested a minimum of three years with an extension of one year.

Time clearly being of the essence, the first meeting for suppliers is planned for the 11th February.  Attendance is expected to be restricted, such will be the crush of entrants. Book now to avoid disappointment.

Parsing government procurement announcements, particularly those for frameworks, is challenging.  But here are a few points:

- Framework values are always made up.  When a framework is launched, there's never any idea of what the take up will be (and it's rarely mandatory that frameworks be used - and, even if it were, there are many overlapping frameworks that would mean you could use a different one). But what's important is that the number is set as large as possible because that (a) ensures that the limit will never be breached, which would be terrible and (b) ensures that suppliers take notice and seek to bid.

- Frameworks offer you a chance to bid for future work, not a right to it.  So you compete, as a supplier, against generic requirements providing detailed pricing (that you can be held to) and get on the framework, and then you have to wait for business to arrive or you have to chase business which you will also have to compete for (against specific requirements). What's missing in this notice is the statement "and here are the departments who have already committed to using this framework and this is why we came up with the range £250m - £750m).  I'm not feeling the love.

- This framework, unusually, says it will seek to limit the number of vendors.  It's also unusual in that it says one supplier might win multiple lots.  Yah boo to the small business agenda one might say.  Other departments - the MoJ and FCO for instance - have sought to ensure diversity of supply by making it difficult (even impossible) for one supplier to win multiple lots in their ongoing IT procurements.  This framework seems to lessen competition and certainly takes an opposite view from G-Cloud's hugely successful "Come one, come all" approach.

- Existing departmental Oracle systems (or any other ERP system for that matter) are almost always wrapped up in their wider outsourcing agreement.  So IBM run Defra's services, Cap run those for HMRC and Logica runs the MoJ's (though the MoJ is more complicated than that with its multiple divisions). So this framework only 'works' when an existing contract comes up for renewal and a department wants to separate its ERP from its other IT. I don't see why a department would do that as its first choice - they're struggling already with managing the splits into a dozen towers, brought together by a SIAM.  Only direction (read force) will change that - the equivalent of gov.uk in ERP. 

- Separately, the Cabinet Office Shared Service Strategy which targets savings of £400m-600m/year with full delivery expected by 2014.  This document was only published in December 2012. It includes this paragraph:
  1. Single Oracle ERP Platform. A number of customers included in ISSC 2 require
    an upgrade of their Oracle Release 11 ERP solutions. It is felt that, rather than allow departments to upgrade separately, this situation provides a unique opportunity to consolidate platforms and provide standard processes across the major Oracle-based departments.

    A feasibility study will be commissioned to test whether the aspiration of government is realistic and the design will be based on a ‘prove why it cannot work for you’ approach rather than a ‘what would you like’ approach. This study will also look at Oracle departments who are not immediately in scope for ISSC 2 such as the Foreign and Commonwealth Office (FCO). This project will be managed as part of ISSC 2 until the completion of the feasibility study. 
I haven't seen the results of a feasibility study that says such a consolidation is possible but one assumes the issuance of the FCO's framework means that it's already been proven. Otherwise why go to market?   The project plan in the strategy shows the feasibility study completing in about mid-February 2013 and implementation completing in December 2013 (that would suggest to me that the solution is already known - I don't see anyone buying one, let alone building one by then otherwise)

-The scope includes "upgrades ... implementations ... moves from non-Oracle to an Oracle platform"?  Surely it should read "migrations to THE Oracle platform".  The strategy also doesn't say that the FCO will lead the delivery of a single Oracle platform (only that they will participate in the feasibility study) - though the notice does say that FCO are supporting the Cabinet Office.

- Elsewhere the strategy says that the current cost per head of Oracle services is £160, though the DWP achieve £89 (I have no insight as to whether these comparisons are truly like for like - the strategy notes comparing these things is challenging).  It goes on to say that one solution should save 40% and avoid £32m in upgrade costs to Oracle 12 (because there would only be one Oracle 12 in government).  It notes also that DWP have already completed their upgrade to Oracle 12.  So if DWP is the cheapest, and they have Oracle 12 already, are they not the obvious place to consolidate to?

- Cabinet Office recently conducted a review of existing and inflight frameworks. Some frameworks were kept (G-Cloud, PSN), some were stopped in their tracks (SIAM, G-Host).  Bill Crothers was quoted as saying: “This is a new approach to frameworks to procure ICT for central government. This approach will support goal of making it easier for all ICT suppliers, particularly with eye to SMEs, to do business with us."

This new framework is, then, confusing, inconsistent with the recent framework review, the overall ICT strategy, the Shared Services Strategy and common sense.  

But it does potentially provide a route to consolidate away from multiple Oracle solutions (that exist today) to a single Oracle solution (that hopefully exists today - because building another one to satisfy everyone is never going to happen, not for £750m nor in 750 years).  And they certainly built Versailles in less time than that - though probably not for less money (estimates for the palace vary wildly from £1.5bn to £200bn)

It is, though, very hard to see how a true cost save is achieved any time soon if there is a long line of departments waiting for their turn to migrate to the new single Oracle system, each one wanting a tweak or a change every 5 lines of code let alone once you factor in the cost of data transition, re-working of departmental accounting, retraining of staff (and possibly redundancies on the assumption that one systems needs fewer people to operate).

That said, why wouldn't you go to one system if you could?  Large banks rarely run their books on a country by country, business by business basis.  Cisco doesn't have to send couriers to every corner of the world to get its financial results.  National Grid doesn't have to ask each division to send a spreadsheet once a month with how much they've spent.  

The first question then, is which system?  The second is why consolidate to what you have, at a high and rising cost, rather than to something else at a lower and more stable cost?

- If the target per head cost of an Oracle-based system is the DWP's £89, then the best way to get that price is to configure a system that is identical to the DWP's and able to support other departments.  We could call it something like, oh, "the DWP Oracle ERP system".  Let's have a competition amongst suppliers to see who can look after the existing system (including all of the people and surrounding processes) and see if the cost can be brought down further.

- Getting from other, higher cost, Oracle solutions to DWP's will not be free of charge and will certainly not be pain free.  Every department with Oracle will have configured theirs to be "just so" and will happily die in several ditches (over and over again) to protect their unique and absolutely required configuration.  So let's be sure to add that cost in.

- And then let's see what the ground up cost per head of an alternative solution is given that the migration costs are going to be there in either case.  I'd be surprised, I think, if it turned out to be as high as £89.

And then you need a line of departments (whatever the solution) ready to adopt the new system so that the cost per head target can be achieved and further economies of scale sought.  The last thing you need is lots of suppliers competing to supply a similar thing as they can never achieve the same economies of scale.

It would, of course, not surprise me to see this framework be marked for, ummm, 'review' and for it to disappear from view before too long.

No comments:

Post a Comment