The IT industry came along and started warehousing information, creating massive infrastructures and databases, telling businesses that this was the way information should be stored. The process was one of taking the information out of the operating system and sticking it into the database. But according to Desmond Seeley, general manager strategy & marketing at T-Systems South Africa, the data was not presented in the format businesses wanted to, or needed to see it in.
"Essentially businesses now had heaps of data lying in some structure somewhere that was actually serving no value. The concept of data warehousing is flawed from a theoretical point of view - it stems from IT companies having got on the bandwagon, scaring business people into thinking that they were not getting any valuable information because it all lay in the transactional layer of their business. This opened the door for the IT companies to sell something else that would accommodate and process all this information that lay within the transactional system," says Seeley.
Seeley goes on to explain that the cost of managing the data and the necessary disk space needed to accommodate this mass of information does not measure up to the value derived from the data. There are many data warehousing projects that have failed - so clearly the way they have been implemented is flawed.
The perception of data should not be an internal concept, but an external one, as most of the information managed by an organisation comes from external sources. Seeley suggests that the right question to ask is "how DO you store this information?"
"The warehousing and presentation of information into the required format is a personal issue; people have different ways of looking at things," adds Seeley.
If organisations need online, realtime information to make decisions, they cannot place information into a data warehouse, because of their decision criteria. "Organisations need to concentrate more on a concept of mining, obtaining data from the source and not from a structural database, where data has been placed. That is where the concept of the Internet and network programming technologies come in - they are all about collecting data," explains Seeley.
"Organisations should utilise online software agents to search and collect information from online data warehouses around the world, allowing them to sort through the information gathered, in their own time as and when they need it," adds Seeley.
Seeley goes on to explain that the only time data warehousing makes any sense is when you are interpreting 'historical' data or trends, requiring analysis to be done on the data stored - such as within the medial claims industry. Adds Seeley, "Using data warehousing as a business information tool does not make sense. It is costly and does not work. Data warehousing has not delivered the expected results or value that businesses perceived it would, or how the IT industry professed it would."
Making 'now' decisions
Businesses are being faced with the same problems in their business chain that are evident in a normal store warehouse in the logistic chain. The concept of warehousing data for online, realtime use to make business decisions on a daily basis does not make any sense. "Nine times out of 10, businesses are not getting the data in the structure or format they need it in," says Seeley. "Companies need to adopt a more proactive 'mining' approach to data, putting mechanisms in place that will take them directly to the origin of where the data was created - allowing them to make the ever important 'now' decision."
Seeley explains, "Businesses' expectations of what warehousing can do for them needs to be clearly managed. The biggest failure within the IT industry is when businesses apply the right technology for the wrong problem, ending up with the wrong solution to the problem, forcing businesses to blame the technology - meanwhile the technology was not applied correctly - such as in some cases of customer relationship management (CRM) software."
A clear understanding required
Businesses need to clearly understand the application they want, and then make sure they buy the appropriate product that is suited to the business' requirements. A case in point is Knowledge Management - it is not necessary to install a data warehouse as it can be achieved through the implementation of a workflow system.
The understanding of what data warehouses are has been clouded - what needs to be differentiated between is:
* Does an organisation need a data library, for which a data warehouse will be suited? or
* Does an organisation need a just-in-time (JIT) product that will provide information when it is needed?
"The cost of storing data in a warehouse is far too high in comparison to the value derived from the data," says Seeley. "You can abstract warehousing to any level you want to and the same goes for the concept of mining - if you want information, you do not have to warehouse it. You need the information now to make informed business decisions that will increase your competitive advantage."
Seeley adds that organisations need to determine the timing, relevance and required usage of the information, which in turn will determine the mechanism to be used to manage the data.
"Warehousing makes sense when you want to restructure a piece of data, do period reviews or analyse the data in a particular way. If the data you need changes regularly, you will need another mechanism to find the 'now' information," concludes Seeley. "Data warehousing fits into the archive space."