Sunday 10 October 2010

Count the cost

A few weeks ago, I highlighted the case of a local government organization which had tried unsuccessfully to implement SAP. As it happens, one of the regular readers of this blog had let me have details of another case in California shortly before that - I have been trying to get some more information on the situation out there, but without much success. In both cases, they have spent a great deal on their implementation without seeing any real benefits - and it looks highly likely that they could cancel their projects having spent millions of public finance for no real value.

About a month ago, I was browsing thru SCN (SAP Community Network). I read an article that I found really interesting (can't remember who by, and can't seem to find it again). The author asked why there were so many stories of poor implementations, and so few stories of successful ones. If I were to be a bit facetious, I might say that it reflects the real situation - but I am actually not sure that is the case. I found the comments made in the item quite interesting, particularly when analyzing the two cases I've highlighted and the situation at the place where I work.

As a general rule, if people are satisfied, they tend not to say very much - they will only make the point publicly if they are extremely pleased. On the other hand, if they are not happy for any reason, they will let you know, often in no uncertain terms! This is just the way that people are - anyone working in customer service will be aware of this. It's usually seen as normal that you will get a lot more letters of complaint than you will of praise - and the complainers will often highlight relatively minor issues, whereas the praisers will normally highlight more major aspects.

One of the common complaints over SAP implementation is that the cost is so high. In our case, we have now spent over $2.5 million with the System Integrators, and it is still climbing. But that doesn't include the internal costs of staff time involved in the project. No exact figures have been kept, but it would seem that most people agree we have spent at least another $1.2 million, probably more, and there is more to come yet.

And we are not finished - we have to roll SAP out to sites overseas, and no-one is prepared to guess what the costs will be, and certainly not what the final tally will reach once they are all done. It also seems likely that once we complete the intial implementation, we could find that we are paying for further development work - it's possible that the project costs will continue to rise over the next couple of decades.

Now there are many that will say we haven't actually spent that much - after all, the implementations at the two organisations highlighted have both been in the 10s of millions of dollars. But they are both much bigger projects, with more people involved so they are bound to be more costly. They have greater revenue streams, so can afford bigger budgets. In our case, the amount committed is many times higher than was originally planned and it has eaten up our annual profits for the last couple of years. It makes it difficult to find cash to pump into other areas of the business that are in need of investment, and this could have a negative impact on our ability to continue to trade.

There are those that say a project involving an ERP implementation cannot just be judged on the same criteria that a smaller project would be assessed against. For example, cost and time overruns are less important than making sure the project does actually get finished and that the product should be judged on many more aspects. I can see the logic in this, and to an extent, I would accept that we do have to take much more of an overview for such a major operation. But equally, no-one should just continue to pour greenbacks into something if there is no indication that the results will justify the expenditure.

In our case, we were assured right back at the beginning, that the costs for the first five years of software, consultancy and support, would be around $1.3 million. Clearly that was a badly underestimated figure and in fact we will probably have spent more than 3 times that much, not including the costs of our own people by the 5 year mark. If that were to result in savings, extra sales or improved margins, then an argument could be made that it was all worthwhile. I would agree that we have seen some efficiencies, but I'm afraid not enough to offset the high cost of the project.

No-one believes that implementing SAP will actually help us sell more - it doesn't make the sales process any easier, and is not going to help us win new business. It may help us improve our margins, but that is more of a long term initiative and won't help us deal with more urgent concerns. I may be wrong, but I suspect that we may never see any return on our investment, even if we take a really long term view of 25 years.

So how do we then assess the value of our implementation? I wish I could answer that. Yes it is working and in that we have clearly been more successful than some others. But has the cost been too high? Only time will tell.

2 comments:

  1. Look into tools from Flexera Software. They can't reduce the implementation costs though can reduce the ongoing licence and maintenance costs.

    ReplyDelete
  2. This is hardly the first time I see an account of impossibility to get actual returns form an implementation of SAP.

    The risk of facing the issues you report are maximized in SAP because of its inflexibility for configuration changes. You need to write your business design in stone before doing any configuration otherwise you will face issues in the implementation road. Besides that many changes that may just be a configuration need for other ERP's it requires programming in SAP.

    Once the cost of implementing is incurred, the customer gets into a vicious circle, it has spent too much money to start again so it continues to spend more money.

    If you are still rolling down into other branches, may be you should evaluate if the branches really need SAP.

    Regards

    Ramiro

    ReplyDelete