BI is beautiful, but its extra work


With this modified quote by Valentin I would like to contradict with all clarity, all claims that the fruits of a successful BI-project are just a few mouse clicks away because modern tools support in a smart and intelligent way, the design process of a BI-solution so that a) you do not have to do much anymore and b) you can hardly do anything wrong. Comment (a) is complete nonsense and (b) is also complete nonsense.

If you read something somewhere along that line, forget it, regardless where you read it.

BI is a wide field and a more exact definition of what could be meant with it, is nearly impossible. Roughly speaking, it is about a system for the support of decisions on all levels of the operational and strategic management. To achieve that, data from internal and external sources are conveniently analysed and prepared in such a way, that they fulfill the purpose optimally. No one should seriously believe that that can be achieved with just a few mouse clicks.

When one works a little bit with literature talking about that topic, one will find to a large extent, rather technical topics, master-data-management, data-mining, technical data models and so on. One can wonderfully go on about these topics and truly design grand and complex architectures for the management of data. In the case of BI, one has at most just made one small step because BI is first and foremost a business topic and only then a technological topic. Without at least one tangible business case, BI is a pointless venture.

I have read somewhere a phrase saying: ‘…for a BI-project becoming successful, mathematicians have to talk with the business people.’ One can expect that this is for both sides an irritating idea. But unfortunately, there is some truth in that. BI-projects require a continuous dialogue between humans that under normal circumstances hardly talk to each other and on top of that hardly understand what the other one means or says.

Why mathematicians (or related folks) anyway? Is that necessary?

Yes, it is necessary. The design of suitable models for the analysis of complex, heterogenic data requires specialized skills and these skills are found in the brains of these talked about groups. These groups have then to understand what it is all about and what the expected final outcome is. You can imagine that this is rather strenuous for both sides but highly productive once it works.

For the first step in this direction, it is best to use a very familiar tool: your own head. This approach can, in comparison with some recommendations   of the consulting fraternity, be seen as novel.

Those responsible for BI are not spared, to firstly do some heavy intellectual work and to think intensely about interactions and system structures.


Myths and Legends –   and the dispelling of same

I say: be careful of nerds who maintain, that in future one does not have to think any more about such antiquated things like cause and effect and that this approach can be replaced through the search of correlations in big-data-clouds.

We let a few smart programmes roam within a huge data collection and like magic, the most spectacular knowledge arises. We then would not know anymore why something happens, but only when and with which probability and in which sequence. But that seems to be completely sufficient according to the nerds.

Until we reach that point we are still forced to gain knowledge and assumptions about links in business. We pass on these assumptions to the already mentioned mathematicians that they can think about how to figure out whether we are right. Because that is what BI-projects are essentially about: one establishes a presumption about quantitative and qualitative aspects of business-scenarios and examine internal and external data, to then verify the assumptions and if necessary modify these. Ideally you will obtain usable information that can be integrated in a suitable manner into decision processes.

It is not as if in the last decades there was a lack of processes to derivate and interpret KPIs. It is also not so that the problems arising out of the operative areas had desperately waited for someone to finally invent BI to solve these same problems. This initial situation suggests that the ideas and concepts that are offered today under the label ‘BI’ are not really soooo new.

And where is the Innovation?

Theoretically and practically we are today able to include much more data in analyses, to apply considerably more complex procedures and to do this much faster and hence more frequently – even in real time, if necessary. We should ignore for now that “more” of everything often results in more confusion.

And I guess no one will adjust for logical reasons strategic decisions with a high frequency, only because fresh data out of the operative business are offered in real time.

It is a different story in the operative area. A well designed BI project offers without doubt, the possibility to adjust own processes more exactly and faster to market requirements and/or recognize problems earlier.

As already said a well designed BI project requires a lot of work … see above.

Do you share my opinion? I look forward to a conversation in the commentary field.

 – by Mario Lütkebohle, consultant, intelligence AG-

Benzer mesajlar

Daha fazla bilgi
Daha fazla bilgi
SAP HANA Enterprise Cloud
Daha fazla bilgi
Daha fazla bilgi
Digital Transformation
Daha fazla bilgi
Daha fazla bilgi
Bizi takip edin: