Ask someone what they want, and they’ll tell you they want a faster horse. Provoke them, and they’ll tell you they have a ‘get there faster’ problem, an ‘equine waste disposal’ problem, and issues with total cost of ownership.
Thought Provoking
If your requirements elicitation session looks like the photo above, you’re doing it wrong. However, just asking people what they want and confirming that you heard what they said is also not enough. Active listening is important, but to capture great requirements, you also have to provoke thought about why someone is expressing a “requirement.”
Adrian Reed wrote a great article this week (hat tip to Kevin Brennan) on asking provoking questions that leverage lateral thinking techniques to get better insight into the true requirements. Adrian presents eight questions, such as “Imagine if we fast-forward to 2 years after the implementation of this project, what will the organisation look like?” Some of his questions remind me a lot of the ideas behind Enthiosys’ Innovation Games (and Luke Hohmann’s Innovation Games book). The remember the future, and product box games immediately come to mind.
Unprovoked Thoughts
Most good subject matter experts I’ve met, when asked about the important problems to be solved, try and be really helpful and incorporate elements of solutions in their descriptions of problems. They will say things like “the system must integrate with [other system] to do X.” They may even ultimately be right, that this particular system integration is a constraint, and that “X” is the only acceptable (by policy) way to achieve “Y.” But usually, neither constraint is a requirement – it is a solution approach.
Subject matter experts who are not as good at having and sharing insights about their domain often confuse problem manifestations with their underlying problems. By analogy, it is requesting treatment for a runny nose, when the problem the you have is the flu. You can dry up your nose and still feel horrible.
Provoking Questions Reveal Real Problems
Adrian’s questions are designed to help you understand that you’re treating the flu, and not a runny nose. Requirements gathering is a lot like diagnosis of a medical malady. You have to discover the real problems. The problems that people are willing to pay to solve. You have to uncover the latent problems that are “hidden” behind problem manifestations.
In a (rare for me) American football analogy – the problem manifestation is that your quarterback is completing 1 of 20 forward passes. Replacing the quarterback and receivers will not solve the problem. The problem is that your offensive line is not able to give the quarterback sufficient time to throw higher-probability-of-success passes.
Asking questions that force people to describe their objectives differently is a good way to bypass solution-design answers. It also creates chinks in the armor of problem manifestations. Completing more passes is not the future you’re looking for, winning more games is the goal. When you’re treating your flu, your goal is not to be sick – but with a dry nose. Your goal is to be well. When you ask someone to remember the future, they will will describe being not sick, not being dry-nosed. The product box will be a description of a winning team.
Check out Adrian’s list of questions, and ask yourself, how do you get to the root causes? Ishikawa diagrams (also known as cause and effect or fishbone diagrams) provide a great visualization tool if you’re a spatial thinker or a whiteboard-talker. In the example below, you can quickly see that spending too much on fuel is part of the real problem – that the cost of operation is too high. You can likewise see that under-inflated tires are a source of poor fuel economy. Check out the Ishikawa article for an explanation, or this article on providing context (with Ishikawa diagrams), and this article on buyer and user personas for more examples of problem decomposition.
If you’ve got any examples of problem-statement-turned-problem, chime in below…
Scott,
Good insights, yet again. I’m impressed with the consistent quality of your posts.
As you suggest, going beyond the surface, first-impression responses is often critically important to making real progress and getting to the most useful, actionable knowledge. This reminds me a lot of a fantastic online presentation I recently saw by Eric Ries. The entire 1 hour presentation is one of the best I’ve seen in a long time. Especially relevant to your post is his 2 minute discussion of “the 5 whys,” discussed in the video here: http://ecorner.stanford.edu/authorMaterialInfo.html?mid=2296
– Justin
Thanks Justin!
That’s an awesome video, too. I’m going to re-use that clip for explaining it too – better than I could have done. I love the “do the first hour of your eight-week training” trick too. Great technique for incremental improvement – and for bypassing the manager-speak for “I don’t want to do that, so I’ll make it unpalatable.”
I often listen to great television or radio interviewers for ideas and best practices that I can apply when talking to customers. One thing I’ve noticed is that great interviewers don’t inject their own thoughts or opinions into a question. A lot of people (myself included) load a question with an answer by phrasing things like “don’t you think that …” instead of asking blunt, straightforward questions.
I think we do it because we want to sound smart when talking to a customer or prospect, or we already have an answer in mind and are just looking for confirmation. However, when I focus on *not* loading questions and actively listening to what the person says, my requirements gathering has gone much better.
Thanks, Tony, great suggestion!
There is a case where I intentionally add my own ideas – when trying to dispute what I’ve just been told, I’ll often make up examples that are clearly (or clearly, to me) wrong.
For example, if a person says that they need “A plus the absence of B” I will respond with “A plus B” or “B alone” or “the absence of A” or variations of “A” and “B” that they did not describe, to help the SME clarify their statement into (or away from) the more general cases.
But I love asking the ‘unloaded’ question tip! Thanks again.
Great post Scott. This is one of the biggest problems I face in my current position, getting people to ask these sorts of questions to get to the underlying problem instead of a potential solution. This is especially key when you have many clients using your product. If you don’t do this, you end up building in “enhancements” to your product that really only work for one client because it was the solution they suggested instead of adequately addressing the problem which many of your clients are experiencing. Since your first attempt didn’t solve the problem for everyone, you end up implementing several other “enhancements” to solve the same problem for other clients and in the end, your product becomes overly complicated and confusing to use.
Hey Matt, thanks, and welcome to Tyner Blain!
If you happened to subscribe to the comments on this article, thought I’d add an update. I just wrote about part of this in a new article, instead of a giant comment.
http://tynerblain.com/blog/2011/02/03/everything-old-is-new-again/