ERP disasters - Bet the company - and lose

It’s enough to send a shiver down the spine of the boldest FD. You’ve got customers, you’ve got orders – you’ve even got manufactured product sitting in warehouses that precisely matches what those customers want.

But you can’t ship it to them. Not because of external events – blockading French farmers, severe weather or somesuch – but simply because you’ve implemented a new IT system and it doesn’t work.

Of course, it’s in order to avoid such eventualities that sensible businesses insist on dealing with major software companies, and rely on Big Five implementation consultants to install the systems. Inevitably, systems from small software houses of unproven provenance will be a risky proposition – particularly when systems from a number of such vendors must be pieced together. Which is why, you’re saying, your company would have played safe and bought an integrated system from a vendor such as SAP.

Tell that to America’s Hershey Foods Corporation, which issued its second profits warning in as many months in the run-up to last Christmas.

The cause: massive distribution problems following a flawed implementation of SAP’s R/3 Enterprise Resource Planning (ERP) system, which saw many retailers’ shelves devoid of Hershey confectionery in the critical pre-Halloween selling period, and again at Christmas, another peak time. In a booming stock market, Hershey shares ended the year 26% down from their year’s high, with 1999 earnings-per-share estimated at $2.04, compared to 1998’s $2.34.

Nor was Hershey alone: US-based domestic appliance manufacturer Whirlpool announced similar problems due to its SAP R/3 implementation in November, as did parts and supplies distributor WW Grainger in early January. But while Hershey’s, Whirlpool’s and Grainger’s beleaguered managements struggle to rectify the problems, they have something to be thankful for: their system is at least in place, and can be fixed.

Two other US companies, Allied Waste Industries and Waste Management Inc, were forced to pull the plug on their SAP R/3 systems, leaving them stuck at the starting block – having incurred hefty expenditures. Nor were Hershey’s problems wholly crippling: the company’s management was doubtless glad that its SAP R/3 implementation difficulties didn’t bankrupt the business, as did problems at US pharmaceutical distributor FoxMeyer, whose receivers last year announced a $500m lawsuit against SAP and co-implementers Andersen Consulting.

Alarming though such stories are, the question is, just how worried should British FDs be? Are these tales isolated instances, conveniently far away on the other side of the Atlantic, that affect only SAP – or is there a threat to all British and European companies, implementing all makes of ERP system?

Well, there’s no point softening the blow: British FDs should be very worried indeed. Disastrous implementations, says analyst Roger Phillips of specialist investment bank Granville plc, “simply go with the ERP territory.” There are, he says, “no cultural or managerial foibles that make US ERP implementations any more predisposed to disasters than any other country’s implementations.”

Look no further than the BBC, he points out, which has recently attracted criticism after spending seven years on an SAP R/3 installation that still isn’t finished – and won’t be finished until mid-2001, according to a statement that the broadcaster put out in January, in which it claimed that the 2000 target date mentioned in the 1998 annual report was “an error”. Or Newcastle University’s £8m SAP R/3 implementation, over-budget by £1m, according to a critical auditor’s report. Or indeed, look at the troubles at Royal Doulton, which led to delivery delays of up to ten weeks, and a sales shortfall of up to £12m, according to a profits warning the firm issued in November. “It was very frustrating,” recalls tight-lipped spokesperson Valerie Baynton. Company chairman Hamish Grossart was more forthcoming: “It’s a cock-up,” he was reported as saying while he attempted to soothe investors’ worries.

Clearly, with consequences such as profit warnings, share-price collapses and insolvency, the financial consequences of a flawed implementation can be a nightmare. So, to what extent can legal redress through the courts recover the losses? Not much: for legal redress to be possible, victims would have to demonstrate that the software itself was at fault – and it usually isn’t.

“When it comes to ERP implementations, the soft stuff is really the hard stuff,” says Dr Byron Fiman, principal and co-founder of Implementation Management Associates, a Colorado-based consultancy. Fiman should know: his consultancy is recognised as one of America’s ERP disaster recovery specialists. Not that every problem implementation is a disaster of Hershey or Whirlpool proportions. Typically, Fiman is called in when either the schedule slips or – worse – the “go live” button has been pressed and the looked-for improvements don’t materialise. “The screens are up,” he says, “but nothing has changed: the cycle times are the same, customer satisfaction metrics don’t shift, and the costs remain the same.”

Why does this happen? Fiman doesn’t mince words: “Unsuccessful implementations are largely the result of companies failing to prepare for the human side of change. Companies often mistakenly regard SAP implementation as a purely technical issue. In fact, at least half the issues in disasters are not technical, but people-related and culture related.” Hence Fiman’s observation about the difficulties of “the soft stuff.”

David Duray, global partner responsible for the SAP implementation business at PricewaterhouseCoopers, concurs. ERP implementations, he points out, are very complex undertakings that involve new technology, new organisational structures, new policies and procedures, as well as a host of technical conversion and data-cleansing tasks – much of which is often outsourced.

But usually, he explains, no single outsourcing partner is responsible for the final integration and testing of all this work: that role is left to the client company.

“When an ERP project unravels or is seen not to perform well, one of the suppliers is usually chosen as the culprit,” he says. “In my experience, this is usually more of a political decision than a proper problem-source identification exercise – and SAP, over the last few years, has been a popular target.”

Jim Shepherd, senior vice president of research at influential industry analyst Boston-based AMR Research, agrees that very rarely are there instances when it’s the ERP system itself – the actual software – that fails. Indeed, he points out, both SAP and Hershey have acknowledged in public that in the Hershey installation the software does what it is supposed to, and that no bug fixes or patches are in the works. It’s the implementation of the software that is at fault, Shepherd says, not the software itself: seemingly-mundane people-related tasks such as training, planning, process mapping, change management and so on are the problem.

In Shepherd’s view, companies at risk of an ERP implementation disaster often share two important characteristics: senior management’s attitude to the systems, and a communications “disconnect” between senior management and the partners and project staff implementing the system. Perhaps as a result of a lack of direct exposure to areas of the business that will be affected by ERP, “management rarely understands how extraordinarily dependent the business is upon these systems.” As a result, their expectations may be unrealistic and critical judgement calls may be faulty.

Then there’s that communications problem: although the importance of proper user education and change management is by now well-known, both are costly and time-consuming to do properly, and rarely do the parties involved in the implementation – software vendors, consultants and so on – have an incentive to be the bearers of unwelcome tidings. “Senior managers often don’t want to be told that there’s a high level of risk and that there’s a great deal of expenditure involved if you’re going to minimise it,” Shepherd argues. “It’s just not a message that they want to hear.”

The result, he says, is an invidious conspiracy of silence. “The people with the message don’t want to tell it, and the people to whom it should be told don’t want to hear it,” sums up Shepherd. And the result is disaster.

Certainly, this is an explanation that seems plausible in the Whirlpool debacle, although the company itself is not saying much. Tests in the run-up to going live showed anomalies and so-called “red flags”: warnings of things that should have been corrected before the go-live button was finally pressed. But project managers chose to ignore these warnings, perhaps so as not to incur the wrath of senior executives. This was a gamble that failed – with consequences that will haunt the manufacturer for years.

There are echoes of change management problems, too, in the circumstances surrounding the Hershey implementation. The boom in ERP sales in the mid-to-late 1990s has been well-documented: ERP had become a fashion, particularly on the back of the fad for Business Process Re-engineering that characterised the early 1990s. But another impetus that took hold as the 1990s progressed was the urge to avoid Year 2000 millennium bug problems. Replace legacy systems with brand-new integrated ERP systems, ran the logic, and you’ll neatly capture significant business benefits while avoiding the chore of de-bugging obsolete systems.

Fine, in theory. That was certainly the logic at Hershey – the problem lay in leaving the decision to adopt ERP until it was much, much too late.

1 January 2000 was an immovable deadline – and one that the company knew it had to implement comfortably ahead of in order to allow for slippage and post-implementation bedding-in. Consequently, what was originally envisioned as a four-year project was crammed into 30 hectic months, with the system going live in July, just as peak season orders came in for Halloween.

This, as PricewaterhouseCoopers’ Duray observes, is hardly recommended practice. “SAP is performing effectively in 20,000 installations around the world,” he says. “It does, however, need to be tested and tuned to each situation: implement too early without sufficient testing, and the problems are encountered in a production environment where remediation is much more difficult.”

David Dobrin, a respected analyst at Boston-based consultant and analyst Benchmarking Partners, puts it more succinctly. “Hershey,” he says, “was simply out of its mind to go live with a big bang at the start of the peak season.” But the really interesting thing, he observes, is to look at why companies continue to make such massive mistakes – not from the point of view of attributing blame, but in order to understand root causes.

And when you ask such questions, the same smoking gun emerges time and time again: training and education. In the light of recommendations from analysts such as IDC, Gartner and Benchmarking Partners itself, a consensus has emerged that between 10% and 20% of a typical ERP project’s cost should be spent on training and education – of which about a quarter should be spent on IT training, and three-quarters spent on training the system’s users. Not just in the screens that they will be using, but – more importantly – in the integrated thinking behind them.

It is this “joined-up thinking” that is a common weak spot, says David Beresford, executive director of Diagonal, the UK-based SAP implementation house that regards itself as the UK’s premier SAP “lifeboat” rescue service.

“The classic example is the sales order process,” he says. “In the past, people could put in the wrong price, the wrong customer, the wrong delivery point – and somehow, the business coped and money came in. Now, surprise, surprise, they don’t get their invoice paid. And it’s even worse if the wrong product is entered: what people need to understand is that when ERP comes in the sales orders are automatically linked to manufacturing and accounting functions.”

As this message has been absorbed, companies implementing ERP systems have been shocked into action. The level of training and education is not up to those recommended levels just yet, but it is rising rapidly, says Patrick Newton, CEO and president of training consultants DACG. “As recently as two years ago, training and education amounted to around 5% of project cost on average,” he notes. But, spurred on by examples such as Hershey and Whirlpool, “companies have begun to wake up to the fact that user education is a key requirement. Now training spend has more than doubled.”

This change of tack cannot be over-emphasised. Historically, the big ERP vendors have focused primarily on training IT people to operate the nuts-and-bolts of the new systems. Indeed, there may even have been a deliberate attempt to underplay user education: “Every SAP implementation partner says that training is important,” observes Diagonal’s Beresford, “but it’s often one of the first things to go when the talks about pricing get tough.” Now, with memories of disasters such as Hershey still fresh, businesses will be less tempted to take such shortcuts. As DACG’s Newton sums up: “You can have the best software in the world – and still fail if people don’t understand how to use it.”