Our research shows that building a business case is one of the major barriers when putting together an MDM initiative. Some of this is due to people not knowing the correct information to present in order to make a convincing business case, and in other cases it is due to the difficulty in estimating costs and benefits. This is actually not rocket science.

Join me for a webinar this Wednesday (at 11 EST = 16:00 UK time) on how to build the business case for an MDM project. It contains lots of practical information to make a proper business case, including data from our latest research e.g. on how much MDM projects cost to maintain once live.

You can register for it here:


It is even free!

Price is in the eye of the beholder

I enjoyed an article today by Boris Evelson OF Forrester, discussing whether a 20% or discount for BI software was “good” or “bad”. As Boris so correctly points out, vendors go to great lengths to keep secret the list price of their software, as they prefer to assess each pricing negotiation separately. Indeed most vendor contracts have a non-disclosure clause that prevents companies from sharing any pricing information.

Given this, what sense does a 20% (or 50%) discount make? None at all. If the vendor was offering an enterprise deal at USD 2 million initially, and you haggle and get them down to USD 1 million, that may make you feel good, but what if they had been prepared to settle for half a million all along? Vendors are aware that many corporate procurement people have targets to achieve a certain percentage over list, and in most industries that is clear enough. After all, if you wanted to buy a new Mercedes then you can just look up the list price. Then if you get 10% or 20% off that from the dealer it is clear what you have saved.

But if the list price is shrouded in mystery, discounts are meaningless. You need to achieve the lowest price that you can, and the only way for that to happen is to be in a competitive situation. As soon as a vendor knows that they are the sure-fired winner or (even better) the only bidder, then they can basically make up any number that they want and stick to it. I have been on both sides of multi-million dollar software negotiations, and I can assure you that companies frequently pay well over the odds, sometimes millions more than they have to, simply through foolish procurement practices.

You need a well structured evaluation, with several competing vendors, bidding against each other on price as the evaluation is proceeding. In this case you will see who really wants to offer you a good deal. I saw one bid drop from USD 12 million to under half that in a single day once the vendor was convinced that they were genuinely in competition.

I have put quite a few procurement tips together in one module of this course.

MDM and Spaghetti

When looking at the business case for MDM it is normal to look at the kind of business initiatives that can be enabled by better master data. For example with higher quality, consistent customer data it is possible to run more efficient marketing campaigns, or by having a complete picture of a customer it is possible to cross-sell products effectively or better manage an account. However such things tend to rely on having MDM as a piece of infrastructure, so it is hard to claim all the benefits directly for an MDM project. Perhaps it is time to take a look at some of the more murky and less sexy areas that can benefit from MDM, specifically by lowering the cost of maintaining application interfaces.

Large companies have hundreds of applications, even after they have finished implementing an ERP system (and then re-implementing it to reduce the number of ERP instances). One company I work with owns up to 600 applications, another to 3,500. In many cases data needs to be shared across applications, and of course the very fact of having so many systems can cause master data issues to occur, since each application frequently generates and maintains at least some master data that it needs rather than being fed such data by a consistent enterprise-wide master data repository.

One key difference between MDM hubs and a data warehouse is that a warehouse needs to have clean, pure data; this is achieved by an extensive process of data cleansing and validation that is conducted outside the warehouse prior to data being loaded, perhaps through a mix of data quality tools and ETL processing. Indeed one major issue is that in order to come up with high quality data for the warehouse, business rules end up being embedded in sometimes complex ETL scripts, which are opaque and hard to get business people to engage with. A good master data hub should be able to take on much of this burden of strong and managing business rules, and may be a more productive place to carry this out. For example it may be more effective to use probabilistic techniques to help determine matches angst potential data sources (say, multiple product masters) rather than needing to hard-code business rules, as usually happens with ETL scripts. If this is the case then you may be able to get away with a much smaller set of business rules in an MDM hub than were typically necessary in ETL scripts. In turn this reduction of complexity may be able to cause a lot of the maintenance effort needed in maintaining such scripts to be go away.

I have not seen any quantitative analysis out there of the relative productivity of MDM hubs v ETL processing for storing business rules, or the potential effect that this could have on the support effort needed to maintain interfaces. However it was always the case that a high proportion of overall support effort in an enterprise was associated with interfaces, so even a small effect here could have a significant saving in terms of IT costs. I think there is an opportunity here for someone to do some serious research into this area, getting hard data rather than making hand-waving benefits statements. If followed through, it would not surprise me to see this as an area where properly implemented MDM could have a significant effect on IT support costs. Given that so many people claim that making a business case for MDM is one of the biggest obstacles, this would seem to me a fruitful area of further research.

Master data: from jungle to garden in several not so easy steps

I very much liked a succinct article by the ever-reliable Colin White on MDM approaches. Companies still struggle to get to grips with what a roadmap for MDM is all about, with apparently competing (and incomplete and immature) MDM technologies and management consultants who are only a few pages ahead of the customers in the manual. This piece neatly sets out the end goal of MDM and the various approaches to getting there (via analytic MDM or operational MDM as a start). It would have been even better had it explained in more detail how the alternatives can be run in parallel, and going into more depth on the issues of each sequences of steps. However by clearly separating out operational and analytic MDM and showing how these are complementary he is already doing a significant service.

The issue he mentions with “approach 1” i.e. the “complexity of maintaining a complete historical record of master data” can be dealt with if you choose an analytic MDM technology which has built-in support for analysis over time. Colin points out that a key step is to end up with a low-latency master data store as the system of record for the enterprise, acting as a provider of golden copy master data to other sources, both transaction systems and analytical ones such as an enterprise data warehouse. If properly implemented, this will result in a change of the centre of gravity of master data, from the current situation where the system of record is ERP to a situation where the enterprise master data repository is actually the system of record, providing data through a published interface (and an enterprise service bus) through to all other systems, including ERP. This is a desirable end state, and is a key step to starting to unlock the monolithic ERP systems that companies use today into more manageable components.

I really hope that this paper gets the attention that it deserves. Getting most of the key messages into two page article is quite an achievement. I would like to see this developed further, and hopefully it will be.

The mythical software productivity miracle

We have got used to Moore’s Law, whereby hardware gets faster at a dizzying rate, though there ought to be a caveat to this that points out that software gets less and less efficient in tandem with this. A neat summary of this situation is “Intel giveth, and Microsoft taketh away”. However when it comes to software development, the situation is very different. Sure, things have become more productive for developers over the years. My first job as a systems programmer involved coding IBM job control language (JCL) decks, which entertainingly behaved pretty much as though they were still on punch cards, with all kinds of quirks (like cunningly ignoring you if you had a continuation of a line too far to the right, beyond a certain column). I just missed Assembler and started with PL/1, but anyone who coded IBM’s ADF will be familiar enough with Assembler. However it is not clear how much things have really moved on since then. In the 1980s “4GLs” were all the rage, but apart from not compiling and hence being slower to run, they were scarcely much advance on Cobol or PL/1. Then there were “CASE” tools like James Martin’s IEF, which promised to do away with coding altogether. Well, we all know what happened to those. Experienced programmers always knew that the key to productivity was to reuse bits of code that actually worked, long before object orientation came along and made this a little easier. Good programmers always had carefully structured code libraries to call on rather than repeating similar code by editing a copy and making minor changes, so I’m not convinced that productivity raced along that much due to OO either.

This is all anecdotal though – what about hard numbers? Software productivity can be measured in lines of code produced in a given time e.g. a day, though this measure has limitations e.g. is more code really better (maybe implying less reuse) and anyway how do we compare different programming languages? A more objective attempt was to measure the number of function points per day or month. This had the big advantage of being independent of programming language, and also works for packages – you can count the number of function points in SAP (if you had the patience). Unfortunately it requires some manual counting, and so has never really caught on widely beyond some diehards who worked in project office roles (like me). Well, we always used to reckon that 15-30 function points per man month was pretty much a good average for commercial programming, and when Shell actually measured such things back in the 1990s this turned out be pretty true, almost irrespective of whether you were using a 3GL or 4GL, or even a package. Shell Australia measured their SAP implementations carefully and found that the function points per man month was delivered was no better (indeed a little worse) than for custom code, which was an unpopular political message at the time but was inconveniently true. Hence, while 3GL productivity definitely was an advance on Assembler, just about every advance since then has had a fairly marginal effect i.e. programmer teams writing today are only slightly more productive than ones in 1986. By far the most important factor for productivity was size of project: big projects went slowly and small projects went quickly, and that was that.

A new book “Dreaming in Code” by Scott Rosenberg is a timely reminder of why this is. Many of the issues of writing a moderately complex application are not to do with individual programmer productivity and everything to do with human issues like a clear vision, good team communication, teamwork etc. All the faster processors and latest programmer tools in the world can only optimise one part of the software development process. Sadly, the human issues are still there to haunt us, having moved on not one jot. Scott discusses the Chandler open source project and its woes, reminding us that software productivity is only a little bit about technology, and a great deal about human nature.

When I was doing technology planning at Shell I always had a rule: if a problem was constrained by hardware then it would be fixed quicker than you expect, but if the problem was a software issue it would always take longer than you would think. This book tells you why that is not a bad rule.

In the project jungle, your MDM initiative needs claws

Matthew Beyer makes a start on trying to come up with an approach to tackling master data initiatives.  Some of what he says makes good sense, as in “think strategically but act tactically”.  However I’d like to suggest a different approach to him in the way to prioritise.  The biggest problem with the issue of master data is one of scale.  Enterprises have a lot of systems and many types of master data, many far beyond the “10 systems” that is used as an illustration in the article.  Just one Shell subsidiary had 175 interfaces left AFTER they had implemented every module of SAP, to give a sense of the magnitude of the problem in a large company.  Hence an approach that says “just map all the master data in the enterprise and catalog which systems use each type of data” is going to be a severely lengthy process, which will probably get cancelled after a few months when little is to be shown for all the pretty diagrams.

I believe that a master data initiative needs to justify itself, just like any other project that is fighting for the enterprise’s scare resources and capital.  Hence I believe that a good approach is to start by identifying and costing problems that may be associated with master data, and putting a price tag on these problems.  For example, poor customer data could result in duplicate marketing costs, lower customer satisfaction, or misplaced deliveries.  Having an inability to get a view of supplier spend across the enterprise (as 68% of customers in one survey stated at a 2006 UK procurement conference) will have a cost in terms of not being able to get an optimal deal with suppliers, and in resulting in duplicate suppliers.  These things have real costs associated with them, and so, if fixed, have real hard dollar benefits.  Interviewing executives in marketing, procurement, finance, operations etc will soon start to tease out which operational issues are actually causing the business pain, and which have the greatest potential value if they could be fixed.  Business people may not be able to put a precise price tag on each problem, but they must be able to estimate at least a range.  If they cannot, then it is probably not that pressing a problem and you can move on to the next one. 

At the end of such an interview process you will have a series of business problems with estimates of potential savings, and can map this against the master data associated with these business processes.  Now you have a basis for priority.  If it turns out that there are tens of millions of dollars of savings to be gained from fixing problems with (say) supplier data, then that is a very good place to start your MDM pilot.

Such an approach assures you that you will be able to put a business case together for an MDM initiative, even if it has limited scope at first.  Such an initiative has a lot more chance or approval and ongoing survival that something that it perceived to be a purist or IT-led data modelling initiative. 

Provided that you adopt an architecture that can cope with master data in general and not just this one type specifically (i.e. try and avoid “hubs” that only address one type of master data) then you can build on the early success of a pilot project confident that the approach you have taken will be useful across the enterprise.  By getting an early quick win in this way you build the credibility for follow-on projects and can start to justify ongoing investment in protecting the integrity of master data in the future e.g. by setting up a business-led information assed competence centre where ownership of data is clearly defined. 

IT projects of any kind that fail to go through a rigorous cost-benefit case risk not being signed off, and then being cancelled part way through.  The race for funds and resources in a large company is a Darwinian one, so equip your MDM project with the ROI teeth and claws it needs to survive and justify itself.  When times turn sour and the CFO draws up a list of projects to “postpone”, a strong business-driven ROI case will go a long way to ensuring your MDM project claws its way to the top of the heap. 









Some rare common sense

Ad Stam and Mark Allin have written an excellent piece in DM Review this month covering data stewardship and master data management. They correctly point out that, with regards to business intelligence systems, that “change will occur, and time to react will decrease” and lay out a sensible architecture for dealing with this issue. I very much like the way they put emphasis on the need for a business unit to deal with data governance as a key building block. In the article they explain the key requirements of such a group and make the interesting analogy of logistics, which is usually sourced these days to a separate unit or even separate company. Similarly they believe that the management of master data should be managed by a non-IT business unit.

The article also correctly distinguishes between “golden copy” data held in the data warehouse and a master data management repository, which in addition will hold master data in all its stages. The master data repository should be linked to the data warehouse, but are not the same physical entity since the master data repository has to handle “unclean” data whereas the data warehouse should have only fully validates data stored in it.

It is a pleasant change to read such a sensible article on best practice in data management, but this is because Ad and Mark are real practitioners in serious enterprise-wide projects through their work at Atos Origin e.g. at customers like Philips. They are not people who spend their lives giving slick Powerpoint presentations at conferences but are close to the action in real-world implementations. I worry that there are too many people on the conference circuit who are eloquent speakers but haven’t actually seen a real live project for a long time. I have known Ad Stam for many years and can testify that his team at Atos are an extremely competent and experienced set of practitioners who focus on customer delivery rather than self-publicity. If you have a data warehouse or MDM project then you could do a lot worse than use Ad’s team.

Don’t prototype: iterate

Stephen Swoyer makes a case for using EII as a prototype for data warehouses in his recent Enterprise Systems article. As the article reflects, there are some dangers as well as benefits here e.g. the prototype may just be extended and a proper data warehouse system never built. This is a problem because, as I have argued elsewhere. EII is suitable only for a small subset of business intelligence needs. However the valid point is that business users do want to see prototypes, especially in an area like business intelligence where the requirements tend to be fluid and ill-defined. However there is an alternative to buying an EII tool, knocking up a prototype and then building the proper warehouse.

These days you do not have to build a data warehouse, since you can buy one. Packaged solutions can be deployed much more rapidly than data warehouses that have to be designed and built by hand, and if they are flexible enough then an iterative approach to the warehouse can be taken. A great example of this was at Owens Corning, who deployed a data warehouse package over a 90 day period, using weekly builds of the business model. Each week a new piece of the business problem was tackled, the business model was updated in the package and the results presented back to the users. Issues would arise, feedback taken, and the next week the model would be refined, and a new business area started. This highly iterative approach ensured that the business users were fully engaged with the project, and could see a visible improvement in what they were going to get week by week.

After a few weeks the problems became less technical and functional, and more business related e.g. issues of data quality and how certain business model issues were to be resolved. After 90 days the application was delivered, and this was no prototype: it was a fully working, deployed production warehouse. The insights this application gave saved Owens Corning a great deal of money, so much so that the project had a three week payback period. Indeed the project became an Infoworld 100 winner.

Data warehouse project leaders need to rid themselves of the notion that data warehouses have to be long, custom build projects. At present TDWI reckons they take 16 months to deliver on average. This is far too long if using a traditional waterfall methodology, and indeed needs a more iterative approach. But why build a throwaway prototype when you can implement the real thing via a data warehouse package?

Mergers and Measurement

Margaret Harvey points out in a recent article that the effort of integrating the IT systems of two merged companies can be a major constraint and affect the success of the merger. Certainly this is an area that is often neglected in the heat of the deal. But once the investment bankers have collected their fees and an acquisition or merger is done, what is the best approach to integrating IT systems? What is often missed is that, in addition to different systems e.g. one company might use SAP for ERP and the other Oracle, the immediate problem is that the two companies will have completely different coding systems and terminology for everything, from the chart of accounts, through to product and asset hierarchies, customer segmentation, procurement supplier structures and even HR classifications. Even if you have many systems from the same vendor, this will not help you much given that all the business rules and definitions will be different in the two systems.

To begin with the priority should be to understand business performance across the combined new entity, and this does not necessarily involve ripping out half the operational systems. When HBOS did their merger, both Halifax and Bank of Scotland had the same procurement system, but it was soon discovered that this helped little in taking a single view of suppliers across the new group given the different classification of suppliers in each system. To convert all the data from one system into the other was estimated to take well over a year, but instead they put a data warehouse system in which mapped the two supplier hierarchies together, enabling a single view to be taken even though the two underlying systems were still in place. This system was deployed in just three months, giving an immediate view of combined procurement and enabling large savings to be rapidly made. A similar appraoch was taken when Shell bought Pennzoil, and when Intelsat bought Loral.

It makes sense initially to follow this approach so that a picture of operating performance can quickly be made, but at some point you will want to rationalize the operational systems of the two companies, in order to reduce support costs and eliminate duplicated skill sets. It would be helpful to draw up an asset register of the IT systems of the two companies, but just listing the names and broad functional areas of the systems covered is only of limited use. You also need to know the depth of coverage of the systems, and the likely cost of replacement. Clearly, each company may have some systems in much better shape than others, so unless it is case of a whale swallowing a minnow, it is likely that some selection of systems from both sides will be in order. To be able to have a stab at estimating replacement costs, you could use a fairly old but useful technique to estimate application size: function points.

Function points are a measure of system “size” that does not depend on knowing about the underlying technology used to build the system, so applies equally to packages and custom-build systems. Once you know that a system is, say, 2000 function points in size, then there are well established metrics on how long it costs to replace such a system e.g. for transaction systems, a ballpark figure of 25-30 function points per man month can be delivered, which does not really seem to change much whether it is a package or in-house. Hence a 2000 function point transaction system will cost about 80 man-months to build or implement, as a first pass estimate. MIS systems are less demanding technically than transaction systems (as they are generally read only) and better productivity figures can be be achieved here. These industry averages turned to be about right when I was involved in a metrics program at Shell in the mid 1990s. At that time a number of Shell companies counted function points and discovered productivity of around 15 – 30 function points per man month delivered for medium sized transaction systems, irrespective of whether these were in-house systems or packages. Larger projects had lower productivity, smaller projects have higher productivity, so delivering a 20,000 function point system will be a lot worse than a 2,000 function point system in terms of productivity i.e. fewer function points per man month will be delivered on the larger system. Counting function points in full is tedious and indeed is the single factor that has relegated it to something of a geek niche, yet there are short cut estimating techniques that are fairly accurate and are vastly quicker to do that counting in full. By using these short-cut techniques a broadly accurate picture of an application inventory can be pulled together quite quickly, and this should be good enough for a first pass estimate.

There are a host of good books that discuss project metrics and productivity factors which you can read for more detailed guidance. The point here is that by constructing an inventory of the IT applications of both companies involved in a merger you can get a better feel for the likely cost of replacing those systems, and hence make a business case for doing this. In this way you can have a structured approach to deciding which systems to retire, and avoid the two parties on either side of the merger just defending their own systems without regard to functionality or cost of replacement. Knowing the true costs involved of systems integration should be part of the merger due diligence.

Further reading:

Software Engineering Economics
Controlling Software Projects
Function Points

Size still isn’t everything

Madan Sheina, who is one of the smarter analysts out there, has written an excellent piece in Computer Business Review on an old hobby horse of mine: data warehouses that are unnecessarily large. I won’t rehash the arguments that are made in the article here (in which Madan is kind enough to quote me) as you can read it for yourself but you can be sure that bigger is not necessarily better when it comes to making sense of your business peformance: indeed the opposite is usually true.

Giant data warehouses certainly benefit storage vendors, hardware vendors, consultants who build and tune them and DBAs, who love to discuss their largest database as if is was a proxy for their, er, masculinity (apologies to those female DBAs out there, but you know what I mean; it is good for your resume to have worked on very large databases). The trouble is that high volumes of data make it harder to quickly analyse data in a meaninfgul way, and in most cases this sort of data warehouse elephantitis can be avoided by careful consideration of the use cases,probably saving a lot of money to boot. Of course that would involve IT people actually talking to he business users, I won’t be holding my breath for this more thoughtful approach to take off as a trend. Well done Madan for another thoughtful article.