DfT shows how not to do large-scale IT change projects

Keith Rodgers | 23 Dec 2008 | 11:12

If you’re thinking of embarking on any kind of consolidation initiative in HR or payroll, last week’s damning government report on the Department for Transport’s (DfT) shared services fiasco provides a helpful guide to what not to do.

The DfT was accused of “stupendous incompetence” by the House of Commons Public Accounts Committee for its efforts to pull HR, payroll and finance administration into one in-house shared services centre. And for good reason. In the course of a project that converted £57 million worth of anticipated benefits into an £81 millon net cost to the taxpayer, the department seems to have broken pretty much every rule in the book.

The project, which kicked off in April 2005, provides half-a-dozen good lessons for any organisation looking to embark on a big IT-related change project.

1. Don’t skimp on the planning Despite being aware of the risks, DfT put together an aggressive timetable for its shared services project that called for two agencies (the Driver and Vehicle Licensing Agency and the Driving Standards Agency) to go live within a year, believing that there was “no advantage in planning for a longer detailed design and later start”. This compressed timetable was at the root of many of its problems – not least its decision to save time by building on an existing IT system rather than going out to tender. Anyone who’s been involved in tendering knows it can be a time-consuming and tedious process, requiring liaison with multiple departments – but it has the big advantage of forcing you to define your requirements closely and analyse risks.

2. Get buy-in Centralised shared services projects are all about standardising on best practices – but the committee reported that the DfT couldn’t fully resolve “the inherent tensions in securing the agreement of seven separate agencies to a single set of processes”. After going live, meanwhile, it seems DfT wasn’t overly concerned that users lacked confidence in the system, believing this was normal for a large change programme.

3. Know what you’re capable of According to the committee’s report, the department didn’t have enough skilled or experienced project management staff. Likewise, its various agencies “were unable to provide appropriately skilled staff with detailed knowledge of their individual business processes to work at the shared service centre”. This kind of skills gap isn’t uncommon in a shared services project, so you need to be sure you can track down outside help if you need it and budget accordingly

4. Monitor progress and act on the evidence The department realised within two months of kicking off the project that its original timing assumptions were wrong, but pushed on regardless. Not surprisingly, the committee recommended that future projects adopt a milestone approach to assess progress and check the feasibility of the budget and timetable

5. Use meaningful metrics More than three-and-a-half years on from the project launch, DfT collects data on only 14 out of 18 performance indicators, which implies less than total commitment to the principles of performance management. Worse, at the time of the committee hearing, it had hit only four of them. In its defence, DfT argued that its performance lagged averages in the private sector because it has less economies of scale and government reporting requirements add complexity. Both claims may well be true – but, if they are, the department needs to develop more meaningful metrics that take these factors into account.

6. Test, test, test The harsh reality of software is that it usually doesn’t work the way you expect first time round, so you have to test extensively before you go live. DfT took the opposite approach, cutting testing from two months to two weeks to make up lost ground – and, not surprisingly, ended up with an unstable system.

The DfT initiative isn’t the first time that these kinds of mistakes have been made, of course – although it’s rare for one organisation to tick off quite so many boxes on the blunder checklist. The question is, will the rest of us learn from its experiences?

Electronic absence management, one step at a time

Keith Rodgers | 5 Dec 2008 | 12:55

It was great to see the government going digital last week when it unveiled plans to replace paper ‘sick notes’ with electronic ‘fit notes’. But are employers ready to match this modest step towards automation in absence management? Chances are that the first organisation to receive one of the new electronic medical certificates will marvel at the technology – then print it out on a piece of paper and lock it in a filing cabinet.

The government's commitment to a ‘fit note’ was one of a raft of proposalsin its formal response to Dame Carol Black's review of the health of Britain's workforce. Championed by the CIPD, the important part of this initiative is that it shifts the focus of GPs’ advice to the work employees are still able to do, rather than the things they can’t do. It’s all part of an effort to cut the cost of absence, which Dame Carol Black estimated at around £100 billion annually – or according to the CIPD, an average of £666 per employee per year.

As well as shifting the emphasis from sickness to wellness, the government plans to introduce an electronic medical certificate, following Black’s suggestion that with patients’ permission, fit notes could be passed electronically from GPs to employers.

Getting rid of paper is always a good thing, of course – but in this case, we’re probably talking about postponing the start of the paper trail rather than eliminating it entirely. Absence management hasn’t typically been a top priority in HR IT investment, and many organisations still make do with a motley collection of paper forms, standalone spreadsheets, emails, and sticky notes on the boss’s computer. As a result, collecting absence information from each department and getting it into a central HR system is often a haphazard process, and information routinely gets lost.

Worse, it’s expensive as well as inefficient. If you rely on line managers submitting paper forms to HR, you probably also pay a small army of administrators to key in the data – each of whom, incidentally, is costing you £666 each year whenever they’re too sick to make it into the office.

Absence management software comes in different guises, but the idea is to digitise information as quickly as possible, get it to the people who need it, store it centrally – and act on it. Of course, automating the way you manage absence doesn’t cure sickness – my company might have the most sophisticated monitoring system in the world, but when I’ve got the flu, I’ve got the flu. But what a system can do is reduce your outlay, particularly for mid-sized and larger organisations.

For one thing, it gives you better information about long-term absence instances and allows you to make faster, more effective interventions. Likewise, if you collect accurate absence data, you can start to analyse trends. As Black argued, better recording and analysis of certification would enable employers “to identify patterns of absence within particular departments or roles and so deal with possible health problems in the workplace”.

Lastly, don’t forget the people who abuse the system. Without good records, it’s hard to spot when someone’s repeatedly throwing a sickie – and harder still to make a disciplinary case. Conversely, if I log onto my employee portal every morning and see a bright graphic showing the number of days lost to sickness across my organisation, I know my attendance is being monitored. That alone might act as a deterrent the next time I’m tempted to take another day off work with an imaginary dose of food poisoning…