BPO in UK Central Government Sector: “Vendor Beware”
Plenty has been written about the shortcomings of large IT outsourcing contracts and complex IT projects in the UK central government sector. Many have ended up costing way beyond initial budgets, and/or have taken years longer than envisaged, or indeed have turned out to be abject failures. The contributory factors have been thoroughly analysed. Instead, I am going to focus on central government BPO contracts specifically, what went wrong, and the learnings to be seen from them.
In the last few years, we have seen some BPO “failures” that have commanded a lot of media (including social media) attention. Three that immediately have come to mind include:
- Problems in the Work Programme
- Atos and medical assessments for disability allowance
- The Compass programme (housing for asylum seekers)
All of them involve challenging work with vulnerable and/or stressed individuals, and work that often includes being the messenger conveying negative news in carrying out government policy.
The Work Programme, part of the government’s drive to help people back to work and off benefits, was introduced in June 2011 and intended to be a linchpin of the government’s radical welfare reforms. It consolidated previous welfare-to-work schemes, targeting the issue of long-term unemployment, and remains the biggest single “payment by results employment” programme ever introduced. The programme was introduced as being built around the needs of individual jobseekers. In Q1 2014, 18 private, public and voluntary organisations were selected as service providers: each of these had sub-contractors (including around 300 voluntary sector organisations).
The commercial model included the kind of innovation that BPO vendors keep telling us they are so fond of, with:
- Outcome-based pricing, that is focussed on the client’s business agenda
- Service providers being given the freedom to design their own services, based on the needs of jobseekers and local labour markets, as long as they delivered policy objectives – the government did not prescribe the approach
- The cost to the client (the DWP) to be met through savings generated from its legacy £148bn benefits payments
Potentially, these were huge contracts: total value at the time was projected at £3bn-£5bn over seven years, based on the number of people expected to go through the service. Providers can earn between £3.7k and £13.7k per person helped into work (depending how hard it is to give support to them), with an initial payment of between £400 and £600.
The announcement of the preferred supplier list on April Fool’s Day was, perhaps, an ill omen. So how has the Work Programme progressed, in terms of meeting its objectives, and also for the service providers? In year one, none of the service providers met their minimum targets. In the second year of the scheme, nearly half of the providers met or exceeded their targets… but none of the minimum targets in the different payment groups were met overall, and for those people on ESA (who are harder to place than people on JSA), the achievement was 5.3% against a target of 16.5%.
There was a significant improvement in the third year of the Programme, when all contracts for JSA 18-24, JSA 25+ and ESA new customers) met their minimum performance levels.
The approach to the Work Programme appears to have missed some basic aspects of how a buyer should behave in a business process outsource. To list a few:
- Thoroughness of preparation and appropriate speed. The Work Programme was introduced in haste, and, to quote the PAC report: “the Programme was not piloted, the design and development phases overlapped and the business case was devised after the decision to go ahead was taken. The Programme was launched before the IT system designed to support it was operational. At the time of the launch, the IT system could not carry out automated checks on whether the people the prime contractors said they had placed in employment had actually stopped claiming benefits…”
- The approach to payment by results…which works when the relationship between buyer and supplier is one of a partnership. Welfare-to-work schemes are, by their very nature, challenging. Transferring financial risk to a supplier is one thing, but where the buyer mentality is essentially to throw a business (or in this case societal) problem over the fence to the supplier and having no involvement in designing the solution, the benefits sought are not likely to be achieved.
- The dangers of extensive sub-contracting (which caused so many headaches in some early generation large scale HRO deals). In this instance, the accusation of “creaming and parking” (creaming the easier claimants, parking the harder-to-help ones – which was a predictable consequence) meant that some sub-contractors, especially charities, were only handed the most difficult cases.
- Due diligence, especially (but not only) with SMEs… Standards of service have not been consistent.
In March this year, the DWP terminated the worst performing primary provider (Newcastle College Group) and placed other poor performing primary providers in the scheme on an enhanced performance management regime. The DWP is now looking for a replacement provider, expected to be a year-long process.
Which is exactly the situation it is in with the Work Capability Assessments being performed by Atos…
But Atos has longstanding experience of providing medical assessments on behalf of the government – in fact since 1998 (when the former Sema Group was provided a contract by the Benefits Agency Medical Service). So what went wrong in its work on WCAs? One major issue was massively increased volumes (the number of assessments it conducted increased from ~25k per month in 2010, to >100k per month in 2013), combined with the fact that the tests and the accompanying documentation took longer to complete than anticipated – which put huge pressure on resourcing.
And there appears to have been pressure on the individual healthcare professionals carrying out the assessments to meet “statistical norms”: any assessor whose assessments deviate from a narrow range of averages is audited.
In July 2013, the DWP announced it would be contracting new providers to conduct WCAs alongside Atos. In February 2014, Atos announced that it was seeking an early exit from the contract, citing that its employees had been threatened on a regular basis because of their assessments. Unsurprisingly, given the unpopularity of the WCA, a shortage of healthcare professionals has been a persistent challenge. Will things improve with a new supplier?
The fact that Atos continues to provide PIP assessments in two regions (Scotland, NE & NW England; London & S. England) and also provides medical assessment services to Northern Ireland’s Department for Social Development reflects that this remains a specific capability.
Moving on to yet another challenging activity that is likely by its nature to receive negative media attention: providing accommodation, support services and and transport for asylum seekers. The cost of providing these services in 2011-12 was £150m. In 2012 the Home Office sought to save around £140m over seven years (or 13.3%) through introducing new contractual arrangements, reducing the number of housing providers to just three, of which only one (Clearel) had prior experience of the asylum housing sector. Following a transition period from the former smaller suppliers, the contracts became fully operational in all regions by January 2013.
But there were problems, again reflecting undue haste, with Serco and G4S struggling to meet the original September 2012 deadline, and taking on housing stock from previous suppliers without having the time to carry out full inspections. A NOA report in January this year stated that the transition took place during a “demanding period” for the Home Office, also that “the providers believe the information provided to them by the department during procurement was inadequate …. and has resulted in some of the difficulties now faced in running the service. For example, historical information on demand and the service user population does not match the reality they are facing, with take-up of asylum accommodation higher than (the HO) predicted”.
The HO has recovered service credits from G4S and Serco. Serco recently revealed that its COMPASS contract has incurred an operating loss of ~£15m since it started in late 2012 and it expects the contract (which had an initial TCV of £175m) to be loss-making over its five-year life.
To summarise some of the attributes evident in these contracts, which have cost the suppliers very dearly (including loss of reputation):
- Undue haste in design and transition, often driven by a desire to introduce radical reforms at speed
- Lack of information from the buyer during tendering and transition
- Creative commercial models that may be inappropriate, particularly when outcomes cannot easily be measured in binary terms
- Lack of involvement by the buyer in service design, or, arguably, interest in the service user experience
- Inappropriate and/or inadequate service metrics
- Lack of partnership ethos
And add to this the dangers of policy uncertainty and of demand uncertainty (hallmarks of all three examples I have looked at)…. caveat vendor!
I am not defending any suppliers, nor commenting on government policy, but best practice for a buyer in approaching large-scale and complex BPO is not in high evidence.
The latest activity being outsourced in a “payment by results” model, is aspects of the National Probation Service (~70% of the work), which has the added complications of being a) an activity that service providers have no prior experience of delivering and b) of being widely unpopular. It would be nice to think the model will be successful.
As a footnote, the government is entering into some other commercially creative partnerships that are much more likely to be successful, for example where the gainshare is about revenue generation. One example, with Capita, is about commercialising government IP from its portfolio of project and IT management training tools, including PRINCE2. The target, to triple revenues (from ~£40m pa) by year 10, appears realistic. Being commercially creative in government outsourcing can work.
This article was first published in Outsource #37 (Autumn 2014)
About the Author