Data-driven information analysis
The data-driven approach either relies on the data that originate from the information systems that support business operations or goes in search for available external sources that can fulfill the information needs. This approach operates as follows:
- Inventory of resources: the organization makes an inventory of applications and information systems that register and administer the organization’s data. These may include an ERP system, a tool for financial administration, the CRM system, a complaint system or a combination of those. If relevant, and not to forget affordable, the organization might also look for external data such as market information (supplied by a market research agency), demographic data, Land Registry (Cadastre), the patent register, a registry of brand names, weather forecasts, press bureaus such as AP and Reuter, data from CPB (Economic Policy Analysis) or the Chamber of commerce and finally websites of competitors.
- Structure inventory: most sources have their own specific storage structure. This inventory is intended to determine whether those structures are suitable for creating information based on the data. Some sources are not readily suitable for performance measurements because they lack structure or because they are in a format that does not match the IT infrastructure the organization currently uses. Many organizations still use so-called legacy systems, which are very difficult to unlock and even some ERP systems cannot downright be used to generate information.
- Determining counters and indicators: when we know how a source is organized or structured, we can begin to define various types of counters and indicators ranging in complexity. We could, for example, place a counter on the table containing the orders so to be able to measure the number of incoming or processed orders. Or we could multiply the number of items with their price on the table containing order-handling lines and calculate the total sum in order to get insight in the daily turnover figures per product. In this way, we could define all sorts of counters and indicators on different tables or data.
The data-driven approach works fairly easy and can be executed without too many hurdles, provided that resources are easily accessible and well documented. There is a major risk to this approach though: we can not be certain whether the indicators are also genuine KPIs that actually say something about the functioning of business operations in relation to the mission and the strategy. The risk thus is that the Business Intelligence system contains (too) many data but too little actual management information and will eventually not be very useful except maybe operationally.
An alternative: the market-driven approach
On the Business Intelligence market, various industry specific systems – so-called analytical applications- are available. These systems help organizations to complete a first draft of information needs relatively quickly and then to possibly produce a first version of a Business Intelligence system. The market-driven approach consists of two steps:
- Market research and tool selection: analytical applications like Power BI and Cognos that are available on the market are being reviewed based on a number of (technical) selection criteria. The main criterion regarding the mapping of information needs is to see whether both the application and its indicators sufficiently match the organization’s business operations and critical processes.
- Choosing and adjusting: once an organization decides on which industry specific application to use, the indicators often need adjusting as they should be tailored to organization-specific characteristics.
The market-driven approach can be costly but only when it turns out that the application does not fit well with business operations or when it proves to be rather difficult to adjust the indicators – and the required data linkages to the proprietary source systems. If this is the case, the organization will likely need to identify its information needs otherwise – using one of the aforementioned approaches – to then implement this in a customized Business Intelligence system. Another risk is that the organization becomes too fixated on the functionality the application offers and consequently rather blindly trusts the indicators without giving thought about how they should be used and implemented in the business processes. Ultimately, information needs to be turned into use in some way or another. It is therefore sensible to not blindly use analytical applications without adjusting them first, especially since competitors may steer on similar indicators already. It is thus important for an organization to distinguish itself from its competitors by ensuring that the application is very much in line with both the strategy and the key processes.