Requirements gathering methods determine value

There are lots ways to elicit UCD requirements so I don’t intend on listing them all here, what I will note are some of the effective ways that I utilise. They can be described as structured, unstructured or a mixture of the two, but importantly the methods produce differing depth of requirements dependant not only on the method but on the skill of the facilitator and the characteristics of physical location used. In effect the method used is limited by the capability of the facilitator.

Research Structure

It is critical to determine if there is enough usable information around the project to be able to establish a valid research start point. Get the questions right and the answers follow, get the questions wrong and you’ll waste a fortune. If your unsure carry out a pilot study to find out what the questions are.

Methods

Observational / Reflective

Asking users to carry out their normal activities while being observed does take quite a bit of setting up so that the activity data is not skewed as a result of the observation. It is critical not to participate or lead the user. It is especially common when users are told the fine detail about the project; they will try to give you what they think you want. Often this is because they feel threatened by the process, the old notion of time and motion leading to them getting sacked is embedded in British culture. However this type of method is very useful in a pilot study to determine what type of activities should form the basis of questioning or workshop based co-creation.

Narrative

Narrative is about stories, like the narrator in a play, they see the whole thing being able to talk in the present, remind of the past, note the future or describe a similie that others can relate to even if they cannot yet relate to the main story.

Asking users to describe how they think something should work (Cognitive Walkthrough), what the differences might be for different user types, how it’s done now, is there anything in other technology or experience that they can relate it to. In this way a multi-path picture can be built up from each participant for each interactive pathway within the new system or technology.

Diagnostic

Diagnostic is about testing something this will often manifest through workshops with groups of people. By setting a group a task or co-creation ‘Design the front page of your new website’ spending ‘K10,000,000 that’s Karl’s currency’ on to meet your goals and noting the group dynamic, decision making and outputs, very rich information is revealed by participants.

Show and Tell

Show and tell is imperative to feed back to participants what has happened to the information they have provided. More it can be used to validate key concepts and project directions in a fairly light environment, where participants can be highly critical knowing that everyone else is in the same frame of reference for that meeting.

Some of the Problems

Although the whole process exists to elicit information suitable to influence the project, some participants will have pet requirements that if not given the correct level of sycophantic response they will seek to invalidate the entire project as it does not represent their vision of what the project is any more. This type of attitude is very common when working with low pay low (self) risk bureaucrats, as they can personally destroy projects then blame it on ‘expensive’ consultants.

Case Study 4

On a national census project that was due to be run online and offline a member of the bureaucratic project team insisted that their opinions were correct on the basis of having won an award for a website ten years earlier. The opinions were very dated to the point of being bad practice, incorrect and were a constant disabler to the project. This client refused to consider an open support system, marketing website (to prepare country) or full language landing pages to support the full ethnic background of the country in question. The day to day engagement was painful, but the UX for the several forms, question logic and user interactions were eventually signed off even though the client never signed off the personas based upon the requirements they gave.