What is Extended Relational Analysis (ERA)?

The reason Extended Relational Analysis (ERA) is so hard to encapsulate for the technically oriented is because ERA is trying to solve a problem that nothing else on the market solves – a problem few are even aware exists. Technical tools exist to solve technical problems. Database management systems exist to manage large quantities of data. Application development tools exist to build screens and reports. Schema management tools maintain data structures and diagrams. But no tool exists that does the most critical step in a project: communicating with the user in a structured manner that completely extracts the details of their situation and constructs a representation of that situation that can be both understood by the user and implemented with the technical tools at hand. That is something no software tool can do, because software cannot communicate with people. ERA is a translation technique, designed to extract the operational knowledge in the heads of the users and translate it – completely translate it – into a form that makes sense to the technical side of the house.

So what does ERA “look like”? It looks like an interview, but an interview unlike any you have ever seen. It isn’t a vague, meandering chat about what “the system” should look like. It’s an interview that follows a structured, well-defined path that asks definite questions in a definite manner and order, with predictable results. There are no strange, abstract terms for the users to grasp – like “function points” or “higher-level processes” – there are only their familiar business terms being clarified and fed back to them. The results aren’t a pile of random notes, iconographic representations of concepts, or arrows, but rather a clear, commonly constructed set of row-and-column tables that represent the situation which the users face. These row-and-column tables, and the definitions behind them, are the deliverable of this technique. Oh, yes – there’s one other deliverable. Inevitably, when going through this process, the users are forced to define their business terms with such clarity that by the time the interviewing is finished, the users have a clearer grasp of their business data than they’ve ever had before – something which makes it much easier to define the business processes around the data. This is why even the most skeptical and hostile users have come away from ERA sessions as “believers” – believers that their requirements could be heard and mapped properly.

This direct attack on the problem is one reason why ERA is so astonishingly fast. The general way the industry handles data analysis is as an afterthought, something to fit in after some other elaborate, esoteric process, and only to be handled by technicians familiar with its arcane ways. ERA recognizes that data structure is a direct mapping of the business realities that the users have to deal with day-to-day, and approaches it as though the users are the experts. The users are interviewed, the results documented in a manner that they and the technical staff can understand and verify, and the model is completed in their presence. The results of ERA analysis do not have to be “translated by experts” to be useable. The results of ERA analysis are the translation from the experts (i.e. the users) so the technical staff can understand and use them. This “head-on” approach which goes directly to the users for the data is what makes all the difference.

Extended Relational Analysis developed by Relational Systems Corporation, has been in active use for over 25 years. Through the 1980s and early 90s it was the designated design technique of several major database vendors such as Oracle and Informix. But as these vendors began to offer schema management tools of their own, they dropped ERA in favour of the software tools, a mistake as we need interview and documentation techniques, not just software programs. we need techniques that are designed to be done with users and to deliver a product which can be understood by the users yet easily translated into data structures used by the schema and database management tools on the technical side.

ERA is a proven approach for getting data requirements from the users. It does not depend on any technology or software - it will work with any platform. Furthermore, it is geared for both users and technicians. It can be taught to all people, both technicians and end users, in fact, it is specially designed to help them communicate better. If your data analysis keeps stalling because of poor communication between the users and technicians, then you need the power and flexibility provided by ERA.

A précis of ERA material from ERA practitioner, evangelist and trainer Roger Thomas of JCK.


DecisionOne Consulting hosts an ERA Workshop  24-26 July 2006 at Leeds Malmaison. Full details are available here