Software company Erudine has told Parliament of the "current crippling change request culture in Government IT".
In evidence to the Public Administration Select Committee, Erudine says many IT-enabled government service delivery programmes fail because of a flawed IT development approach that cannot accommodate unanticipated change.
"As a result, the more complex and long term the project, the more flawed the project and delivery becomes."
Erudine didn't give examples but there are many. A report by the National Audit Office in 2009 on the C-Nomis project for prisons and the probation service gives a clue on how change requests can be managed by central departments. C-Nomis was subjected to about 800 change requests in two years, 2005 and 2006. The project became unaffordable.
The National Audit Office said of the change requests on the C-Nomis project:
"No process for assessing the cumulative impact of individual change requests on project budget or delivery timetable was in place.
"Our review of the board minutes found large change requests being discussed and approved collectively by board members, but no assessment of the cumulative effect.
"In addition, until a change database was introduced in April 2006, there was inadequate recording of individual changes and their costs, such that internal audit could only estimate that there had been between 300 and 400 change requests in 2005.
"The extent to which these were properly costed before being actioned is not known. Since April 2006 there have been a further 400 change control requests.
"EDS found over one-third of change requests were complex and had a material impact on the scope of C-NOMIS requirements."
2) Prime IT contractors rely on exploiting post-contract change requests (which are inevitable in a multi-year programme). Also, being the gatekeepers for those projects, prime contractors can and do prevent deployment of innovation that can make subsequent change requests cheap or quick to do as they threaten their lucrative revenue streams.
3) This model assumes a non-changing, static world which makes it inadequate for the demands of large-scale IT enabled government service delivery. This is known generally as the "waterfall" approach and involves a sequential design process flowing steadily through various phases to completion (for an overview see: http://en.wikipedia.org/wiki/Waterfall_model). That fundamentally is the core reason why so many government reports (hopefully not this PASC one) have been written and not acted upon, and why current planned programmes, such as the Universal Credit system, looks set to go the same way.
4) The way forward (with spectacular demonstrable success led by the likes of Amazon, Google, e-Bay and others) is to use iterative agile approaches to development and process which embrace and accommodate change, break large systems into component re-usable applications, and focus on payment by usage or results. This "agile" approach requires a step-change in the core skill set from large scale technically oriented systems integration to hybrid technology/business service component assembly. Commercial companies are adopting this concept, making step change savings on traditional approaches.
5) Adopting agile approaches is against the interests of incumbent IT prime contractors to government who stand to be disintermediated (to their detriment but to the benefit of the taxpayer). There is strong rearguard action and lobbying from them fundamentally to maintain the status quo of the waterfall approach as long as possible.
6) The outgoing Government CIO, John Suffolk, recognised the need for adopting an agile approach to break the stranglehold of current incumbent suppliers and to dramatically reduce the cost of change requests. The Cabinet Office ICT Policy published January 2010 ((http://webarchive.nationalarchives.gov.uk/20100304041448/http://www.cabinetoffice.gov.uk/cio/ict.aspx) sets this out clearly but appears to be languishing in the doldrums, with continued emphasis on the "status quo" "waterfall" model.
7) Savings will still be possible within the old "change request" approach, and those in the Efficiency and Reform group are by background well placed to shave savings - possibly up to the 20% (see appendix 3) mooted in the Treasury’s Martin Read Report - but this will simply bring UK government up to par with France and Germany and will not solve the exploitation of the current crippling change request culture in Government IT.
8) The current model is flawed because it cannot accommodate change but is too embedded within government to change. The incumbent suppliers are too powerful, and the executive resolve to change appears too weak.
9) We are happy to answer questions PASC, and expand on the implications of the above in the context of your questions.
Erudine Responses to Specific PASC Questions
1. How well is technology policy co-ordinated across Government?
There appears to be no clear technology policy. The Cabinet Office ICT Policy published in Jan 2010 provided a constructive path forwards away from the current capital intensive "waterfall" approach (which cannot accommodate change) towards an agile, a componentised, open, pay-by-usage approach (which expects and embraces change). This policy document has not been revised at the time of writing and is currently on indefinite hold.
Co-ordination appears poor on four fronts.
a) Ministers are clear about the nature of the step change in delivery and efficiencies they require
b) The former Government CIO was in tune with those requirements and taking active steps to take government IT but has now departed and the progress halted.
c) The Cabinet Office Efficiency and Reform Group appears to be continuing the "business as usual" which cannot accommodate change. It also appears to lack teeth to enforce change within the departments
d) Departments look to be continuing the "waterfall" approach - also encouraged by their incumbent suppliers.
3. Have past lessons from NAO and OGC reviews about unsuccessful IT programmes been learnt and applied?
· They cannot be applied and learnt because subsequent programmes have retained the flawed model of capital intensive "waterfall" IT enabled service delivery, whereas the solution is a componentised, payment by results G-Cloud approach - the approach recognised within Cabinet Office and enshrined in the Cabinet Office ICT Policy: (http://webarchive.nationalarchives.gov.uk/20100304041448/http://www.cabinetoffice.gov.uk/cio/ict.aspx)
10. How well does Government take advantage of new technological developments and external expertise?
Government, through its excellent R&D assistance, invests in British innovation and helps see them through technical due diligence. However Government is structurally unable to deploy the innovation it sponsors as the gatekeepers for adopting that innovation are often those whose revenue streams are most threatened by that innovation.
The more money innovation will save the less likely it will be adopted.
Our view is that British innovators are wasting their energies and resources trying to get traction with government and should focus us elsewhere.
Where government departments latch on to a new technological approach, for example "agile computing", the framework for success appears to be poorly understood. Ambitions to try to shoe-horn, for example agile computing development, into the traditional approach which a 2-year procurement cycle (which assumes a static environment and cannot foresee change) is a recipe for failure
· Erudine Ltd is a Yorkshire-based software development company which has done more than most companies in demonstrating significant potential savings within government service delivery. Unusually for any SME, the company engaged closely pro bono for over a year with the Cabinet Office Efficiency Programme to give views on the most efficient and effective ways of moving from the current capital intensive "waterfall" IT enabled service delivery to a componentised, payment by results G-Cloud approach.
· It also recently set up two round tables of CIOs from major commercial organisations who had delivered step changes:
o one to examine the fundamental changes government IT needed to make for successful delivery; (www.erudine.com/downloads/Erudine_Govt_IT_report.pdf)
o and the other (with City University) to demonstrate practically how 25% of government IT internal expenditure could effectively be placed with SMEs (reports in attached pdf files for your reference). (www.erudine.com/downloads/City_University_Erudine_Policy_Paper_Final.pdf)
· Our experiences with government IT have made us deeply pessimistic for the future of core government systems, and sceptical that anything positive can be achieved from this report. However, we will happily speak to your Committee.
Reference 1: The five key changes to ensure efficient 21st century Government IT-Enabled Service delivery
Reference 2: Policy Challenge: How can Government implement the Coalition policy objective of placing 25% of Government IT External Expenditure to SMEs?
Reference 3: The Theoretical Limits to Savings from the Current "Waterfall" approach to IT-Enabled government Service Delivery
(From Martin Read’s Report for HM Treasury on its Operational Efficiency Programme, Back Office Operations & IT, May 2009] . The failure of the ICT industry to deliver to the citizen means that even a 20% savings in ICT spend does not take the UK up to German and French levels of efficiency. There is no foundation for claiming that the current approaches and development models currently used will deliver significant extra value to the UK government when put in this European context.
Extract from: Operational Efficiency Final Report: HM Treasury: April 2009 ( http://www.bis.gov.uk/assets/biscore/shex/files/oep_final_report_210409_pu728.pdf )
4.36 In summary, whilst it is difficult to make directly comparable estimates of international public sector IT spend, and even more difficult to assess the benefits derived from this spend , the above analysis strongly suggests that the UK public sector’s IT spend is much more than other similar countries and that the UK does not get a proportionate return from this much higher spend . This analysis is supported by recent data from Kable22 which shows public sector IT spending in the UK is 22 per cent higher than in France and 37.5 per cent higher than Germany .
4.37 Even if an assumption is made that the estimate of UK public sector IT spend using Gartner data is too high and the estimate of £18.4 billion is reduced by 13 per cent, (i.e. bringing it in line with the OEP’s £16 billion estimate) and that estimates for the average of France and Germany’s IT spend are too low, and these estimates are increased by the same proportion (13 per cent), the difference in spend would still be 88 per cent. Put another way, even assuming a 13 per cent over-estimate of UK IT spend and a 13 per cent under-estimate of French and German IT spend, the UK still spends nearly twice as much as the average of France and Germany . Even allowing for the inaccuracies of the data collected, it is clear that significant savings should be possible.
4.38 Further support for this conclusion comes from a separate survey conducted by Kable. This shows, for example, that reducing public sector IT expenditure in the UK to the average of France and Germany would involve a 23 per cent reduction in UK public sector IT spend.
4.39 Based on these two pieces of analysis, a 20 per cent saving on the estimated £16 billion spend (equivalent to £3.2 billion) appears to be achievable.