Johannesburg, 18 Jun 2009
In the intelligence (BI) world today, are we creating thermometers or thermostats? This question has significant implications for those involved in BI, from customer organisations to vendors.
It is an important distinction, and one that when you think about it can have far-reaching implications for the BI market. Let`s see why:
* Thermometers have been in use for close on 400 years. The invention of the thermometer was one of the defining moments in medicine, and it has saved countless lives. It typically consists of a narrow, hermetically sealed glass tube marked with gradations, and having at one end a bulb containing mercury or alcohol that expands and contracts in the tube with heating and cooling. It tells you what is going on, and more expensive ones even gather statistics so you can compare data. It tells you when the temperature is going up, and when it is going down. It gives you information you can use, but it cannot recommend or institute any change. In this sense, the thermometer is a static, unidirectional device, with the ability to pass on information but not effect any change.
* A thermostat, on the other hand, senses the environment and automatically makes predetermined changes. It is bi-directional, being able to sense a change in temperature and act on it. Typical examples where thermostats are used are water boilers, air conditioners, central heating and incubators.
Neither thermostat nor thermometer is superior to the other; they are different tools designed to do different jobs. You certainly wouldn`t use a thermostat to see if your child has a temperature, and you wouldn`t apply a thermometer to your hot water cylinder. (Implicitly, every thermostat contains a thermometer.)
But a problem arises in BI when we use the discipline as a thermometer when we should be using it as a thermostat.
It is a source of some amazement and frustration that many companies in our experience are comfortable and happy to accept static, unidirectional, two-dimensional reporting, and they pass that off as MIS or BI. Just to define this kind of reporting, can consume the time and resources of IT to such an extent, and fail to satisfy the needs of users, that BI remains a fledgling, under-used business tool.
Many users, confronted with under-whelming content in BI reports, and seeing how difficult it is to generate a new report through IT, simply do without BI.
It is also important to note that thermometers deal only with review information, which in the information hierarchy has no active financial value. The only time the information gains any active financial value is when action arises from it - and that is the difference between a thermometer and a thermostat, as the thermostat gathers information and acts upon it.
What is needed, and is actually freely available to companies at large, is the thermostat approach.
Here`s how to build thermostat BI:
* Lay a solid data foundation. All of the conclusions users draw must be based on empirically sound and unquestionable data. This implies impeccable discipline with regard to data sources, data routines, data profiling, data integrity, and metadata.
* Lay a solid foundation for data interrogation through the creation of a centralised data warehouse, or operational data store. There are non-negotiable routines and discipline that must be observed here.
* Abstract line-specific data into either a data mart or a BI cube, which permits rapid and speed-of-thought interrogation; this allows users to look and dig deeper into the issues that affect them and the business.
* Give the users a front-end tool that permits them to interrogate their BI system easily and intuitively. This tool should ideally be Web-based, as such an approach tends to democratise BI. It would also be helpful if such a tool were oriented around the dashboard or scorecard approach, as this reduces the lag between an event and the business knowing about it; the dashboard also tracks progress towards or ground lost relative to critical KPIs.
* Train and empower the users. Many a BI implementation has failed simply because the users were not trained or oriented around the impact their use of the system could have on their own and the company`s performance.
* Finally, programme the BI system to make automatic decisions that would speed up the decision-making process for business managers.
So it`s up to management and IT as they look at the return they are achieving on the BI investment: are they happy to stay in thermometer mode, or would they like to move to thermostat mode? That decision could spell the difference between survival and prosperity in today`s market.
Share