Benefits of Enterprise Information Flow - Part Four: Data Flow
We’ve already explored some ways of approaching Enterprise Information Flow when managing information in big companies with big data warehouses.
Today’s topic is simple, yet complicated: data flow.
Proper description and analysis of data/information flow is the ultimate key to EIF. EIF needs to account for all three different data/information flows – technologically formalized data flow, semi-automatic data flow and also non-technical information flow. Take a look:
Your EIF needs to work with transformations on a structural level (data from one data set is transformed into totally different data to fit properly into another data set). That’s common knowledge and, unfortunately, an inadequate approach nowadays. EIF needs to be ready for today’s enterprise data structures – structures of highly integrated and normalized data. For example:
Contacts in company A are collected in one huge data file with a unified structure. All contact data coming into the company is stored in this file. The data comes from a lot of different systems (manual employee input, web apps, subsidiary companies’ systems, etc.) and is exported for everybody in the institution (anybody can search any contact). It is a relatively simple data flow; it is standardized and the data is easily available. But it has one big problem: Information about the source of data in the file is missing.
EIF needs to employ methods to recognize the source of data, even in situations like this. The simple solution is to identify the source in every record (which leads to a tremendous increase in the amount of data). Other possible solutions employ different methods based on better descriptions of transformations or more sophisticated transformational processes.
In short, you have to be ready to work with data that is actually being transformed, but also with data that is needed to transform the original data. This data is not actually part of the data flow, but it is absolutely necessary for the successful transformation of data (and subsequently for information provided to users).
Is your system ready for structural level transformations? Submit your comments in this Reddit, follow us on Twitter or get in touch via email.