I had a great two weeks away from all things online, so I’m jumping into this debate rather late.
Part of my sporadic academic work of late has involved re-reading quite a lot of impenetrable stuff about the limits of empirical research in social sciences. Karl Popper etc. Do not fear, I will not attempt to force them on you here. In fact, I’m going to ask for the opposite here right now, I want a big solid pasta plate full of empirical research.
The blogosphere is full of folks dashing off opinions. Some are very erudite, others not. 99% of the blogs I read are opinions. That’s cool, there is nothing wrong with an opinion. I have a whole archive of them here on this very blog. It is just that an opinion, is well, just an opinion.
Cynthia Rettig recently published a short piece in the SMR called The Trouble with Enterprise Software Part of the piece is a rehash of ERP is complex, and therefore doomed argument. It is like the stuff Bobby Cameron did about SAP in the mid 1990s. The article also links to research about how CEO’s view CIOs. (similar results to how they view other C-level execs, I thought), and some commentary on large system complexity. It was strung together neatly enough, but it wasn’t based on any real research, at least in a recent enterprise software context.
The article’s core argument,
At present, however, corporations see in software’s seductive invisibility and seemingly open-ended flexibility a never-ending frontier of promise, where hope triumphs over reality and the search for the next new thing trumps addressing difficult existing problems.
Is tantamount to calling corporations stupid. Very few are, either that or us vendors have a conspiracy going that would defeat even Jason Bourne.
Some of my fellow enterprise software bloggers jumped to praise and expand on the article ERP is in a mess, and a donkey so they say.
Sensibly, Andrew McAfee points to some peer reviewed empirical research to counter this doomsday view of enterprise software.
The sober and understated language of this paper’s abstract contains a vital insight for people who question the overall value delivered to companies by their information technologies: if IT were not delivering value, rational decision makers would not keep investing in it. Rettig’s argument falls into a long line of pessimistic writing about the value of corporate IT. Much of this writing takes the implicit, and at times explicit, view that the executives who make technology decisions are dupes, perennially falling for a “triumphant vision” of software. These executives are presumably swayed by vendors’ sales pitches and the consistent message from IT’s ‘helper industries’—an ecosystem of analysts, journalists, consultants, and (yes) academics—that everything’s different now, so investments must be made.
10 months ago I read this paper Scale without Mass:Business Process Replication and Industry Dynamics (I’ve blogged about it a number of times.and sadly I even brushed up on some statistics to understand the methodology)
I wonder why the much enterprise blogosphere has largely ignored one of the most significant empirical research projects on the link between IT and Productivity, yet hypes up a neat little polemic that offers nothing new? Perhaps because the empirical evidence jarrs with those dearly held opinions? After all, where is the techmeme spike for the post “SAP and Oracle worked fine in 10,000’s of business again today”?
Over on the Enterprise Irregulars board Dennis Howlett dissed McAfee’s post . Dennis said…
This from Andrew McAfee: struck me as being incredibly simplistic and avoiding the reality that’s out
Odd that. I’m out ‘there’, and McAfee’s post describes exactly the ‘reality’ of much of my day job. The other week I was talking to a bank that is merging yet another take over and a huge utility dealing with deregulation. I spent most of a day recently with a very large chemical company that had gone from 47 core HR systems to one. The chap who normally sits opposite me is closely involved with a project at the largest food company in the world. Next week I’m at a leading high tech company helping work out what bits to run on the ERP, and what bits to run on niche applications, building a joint roadmap to reduce integration complexity. No, it is not all triumphant vision, but this stuff works. What all these companies have in common is that they are simplifying their businesses with an ERP system. The core ERP business is healthier than I’ve seen it since the late 1990’s.
If you ask P&G about how they absorb takeovers, their ERP system is a big part of this success.
“Staying the same will not work. The supply chain can hold your business model back, or propel you ahead,” says P&G CEO AG Lafley. “SAP allows us to get more in tune with the dynamics of the supply chain, with our consumers and with our customers.”
Or what the Nestle CEO says about the globe project, linking it to Nestle’s impressive business results.
…we’re fully benefiting from Globe, which now covers about 90 per cent of sales. That allows much better management of working capital and trade spend.
…”He puts Globe paramount among his enablers. “Globe allows us to have a much tighter management. A much greater level of transparency. Not only is the system now covering virtually all our businesses, we’re updating some of the earliest applications.”
But then, I’m just swopping anecdote for anecdote.
Digging a bit deeper, I read Hackett studies that say
by moving to a single ERP system for finance and at the same time implementing consistent data and technology standards, companies can cut the cost of finance operations by 23 percent, according to Hackett’s Book of Numbers™ Research. But companies that take either of these approaches independently may see little to no savings, or even a slight increase in finance operations costs, Hackett found.
World-class finance organizations rely on both of these approaches, which help them spend 31 percent less than their peers on finance, operate with nearly half the staff, and also complete their financial reporting cycle more quickly each month.
Hackett’s research found that world-class Fortune 500 companies run these functions at lower operational costs of $134 million/year ($7.1 million/billion of revenue) compared to typical companies, and process automation and IT enablement play a very significant role in realizing these lower non-IT back-office costs. In addition to this efficiency impact of IT, a direct correlation was found between performance of the IT function and effectiveness in finance, procurement, and HR
This stuff works.
Much of this blog is a rant against complexity, and if I look around here at SAP, I’d say fighting complexity is our biggest competitor. Sure we have much to learn about simplification, and we must get significantly better at reducing and managing complexity. But if there is one thing that I loathe more than unnecessary complexity it is the oversimplisitic. ERP is complex, so is the Belgian tax code. Many of those that damn SAP and Oracle for its complexity seem to suspend business reality when discussing the next great start up that will blow us away.
I’ll restate my rumplestilkin test.
Lock your new paradigm busting vendor in a room and only let them out when they have a compliant Polish payroll.
That sounds trite but most of what ERP applications do is complex ugly stuff. Companies do this stuff, not because they want to, but because a lawmaker, auditor, union or some regulatory authority demands it. I’ve argued before that putty and lego are not good metaphors for the software that helps run your business. If you can sprinkle some magic ceteris paribus dust on business, and create a neat guns or butter world, then enterprise software would be simple. Actually, you could do it on a napkin. But tell that to the folks who thought up the Norwegian (or was it Finnish) business travel per diem rules. (especially the ones about mileage rates for reindeer sled travel)
So keep the comments about ERP salespeople and Porsches rolling. But where is the big stonking empirically solid research that shows me that ERP is dying or that it doesn’t work?
As much as I hate to quote a competitor success story, herewith a small sprinkling of irony… Andrew’s Ducati wouldn’t be on the road today if it wasnt for enterprise software.
I’ll finish this long opinion piece by quoting a bit more Andrew McAfee.
I agree that it’s important not to naively accept anyone’s triumphant vision of corporate IT. But it’s also important not to make claims in the other direction that are too sweeping. Perhaps most fundamentally, it’s critical at some point to stop floating hypotheses about IT’s impact (or lack thereof), and to start testing them. We have enough history and enough data to permit more excellent studies like the one conducted by Aral, Brynjolffson, and Wu. Designing and executing resarch that is both rigorous and relevant is difficult, at times dismayingly so, but as these three show it’s well worth the effort.
Bring on the empiricists.