{"id":375,"date":"2009-05-29T08:04:09","date_gmt":"2009-05-29T15:04:09","guid":{"rendered":"http:\/\/tibetantailor.com\/?p=375"},"modified":"2009-05-30T11:12:01","modified_gmt":"2009-05-30T18:12:01","slug":"design-research-5-mistakes-to-avoid","status":"publish","type":"post","link":"http:\/\/tibetantailor.com\/?p=375","title":{"rendered":"Design Research: 5 Mistakes to Avoid"},"content":{"rendered":"

I am not going to talk about Type I and II errors and stuff like that you can get from any ol\u2019 research methods book. \u00a0These are hard won insights, lessons learned that over time have been hammered into my head. These come from hundreds of interviews and observations and by watching and talking with people. \u00a0These are the mistakes that can keep you from successfully \u201cgetting inside their heads.\u201d<\/span><\/p>\n

Mistake 1: Not thinking hard enough about sampling<\/strong>. This is the most important thing. \u00a0Especially if you have a number of \u201cresearchers\u201d contributing. \u00a0You gotta think about who you are going to talk to or watch.<\/span><\/span><\/p>\n

It is possible to get good results even if you have inexperienced interviewers. \u00a0It is possible to get good results even if you have the wrong questions. \u00a0But it is almost impossible to get good results if you aren\u2019t asking the right people.<\/span><\/p>\n

For good design research you gotta be talking to or watching the right people. \u00a0You mess this up and no matter how well you do with the rest of your plan, tasks, and design, well you might as well be making up your data. \u00a0Who knows, your made up data is probably better!<\/span><\/p>\n

Mistake 2: Not doing any Pre-thinking. \u00a0I<\/strong>f you don\u2019t have a clue about the decisions your decision-makers have to make, then you aren\u2019t giving them as much value as you could be.<\/span><\/span><\/p>\n

Spend sometime talking with your clients, stake-holders, the people who are going to be taking your research and using it. \u00a0What are the decisions they are going to be making that you might be able to help collect data on. \u00a0If you have in mind the decisions down the road, it is much easier to set up your tables, your hypothesizes, and your null hypothesis before so as to use the time as a naturalistic experiment to help them in making their decision. \u00a0It is ten times harder to do this after the fact, and often impossible because of time constraints than it is if you set it up ahead of time. <\/span><\/p>\n

If you can answer one or two or even god forbid three questions that stick in their craw you are rocking and rolling. \u00a0Pre-thinking your decisions leads to pre-thinking your analysis, leads to pre-thinking your data collection. \u00a0Pre-thinking is Good.<\/span><\/p>\n

Mistake 3: Trying to be unbiased<\/strong>. \u00a0Don\u2019t try to be a blank slate. \u00a0It is a waste of time. \u00a0I say \u201cDon\u2019t think about pink elephants\u201d, and what jumps right into your minds-eye. \u00a0Yep, you got it those darn pink elephants. \u00a0Instead, work with what you know already. <\/span><\/span><\/p>\n

Even if you aren\u2019t a professional generalist you know something. \u00a0Or at least you think you know something. \u00a0Spell it out. \u00a0Write it down. \u00a0This is your initial bias. <\/span><\/p>\n

A good anthropologist is not devoid of bias. \u00a0A common misperception is that a good anthropologist is someone who is devoid of bias, someone who has purged their system their brain, their thinking of bias. WRONG! a good anthropologist is AWARE of their biases and takes them into account. \u00a0It isn\u2019t some relativist psycho-babble.<\/span><\/p>\n

Keep a diary AND field notes. \u00a0Use your bias. \u00a0Become aware of things that excite you, piss you off. \u00a0I always say you know you are finding good stuff if strong emotions are involved. \u00a0Try to understand that YOU are the instrument. \u00a0You don\u2019t have to go native or become like the people you study. \u00a0You are you, you aren\u2019t trying to become them. \u00a0You are trying to understand them. \u00a0Use your built in radar of emotion to your advantage.<\/span><\/p>\n

Mistake 4: Not connecting the Pictures Together<\/strong> Many times we focus so much at the mirco level of the customer that we don\u2019t understand the clients macro-view.<\/span><\/span><\/p>\n

Document the macro-system. What is it you are seeking to manage or predict. Seek to understand the system. \u00a0What are you making or saving? How does the system work? What makes it expand or contract. \u00a0What makes it go up or go down.<\/span><\/p>\n

Build a functional model of the business side of things. Build a process model of the micro-customer side of things. Show the connections between the two. \u00a0The amount of insight you can generate for the business owners is immense and just doing this is sometimes worth the entire cost of the project. This helps makes your work easier to understand because you start from your client\u2019s current understanding and build upon that.<\/span><\/p>\n

You don\u2019t want to dunk them into the cold water of a different world-view with nothing for them to hold onto. \u00a0You need to ease them \u201cinside the head\u201d of the customer. The cold hard fact is that they usually don\u2019t care about the customer experience except for how it impinges in a functional sense on the business. So make it clear, make it easy, and connect the dots. It might not be the warm and fuzzies, but it will get clients very excited about the lives and world-views of their customers. <\/span><\/p>\n

Mistake 5: Having More than Five Variables<\/strong>. \u00a0Ok, so here is an easy one. \u00a0Focus in on 5 conceptual variables. \u00a0Not less than 3. Not more than 5.<\/span><\/span><\/p>\n

Usually we think that the more variables we are looking at, the more complicated the system and of course we want to covey the complexities of a customer\u2019s world. \u00a0Unless you are doing extended fieldwork measured in months and years, stick to my advice.<\/span><\/p>\n

Try to determine 5 different variables at the beginning of your work. \u00a0Code your data with those 5 variables. \u00a0These should drive down to operational definitions. \u00a0Don\u2019t worry about scaling theme, likert-izing them. If you can show how one thing influences another thing and if it is a positive or negative influence you are a good way to building a powerful model of what your client wants to change. <\/span><\/p>\n

At least one of these variables has to be the thing that the client or customer wants to change, influence, manage, or predict. \u00a0The other variables can be from the perspective of the informants. \u00a0But the operational definition better be understandable to the client. \u00a0Make sure to use these variables in your Big Pictures you build and it will all start to weave together in a coherent story.<\/span><\/p>\n

To conclude, these are some of the things I have learned over my almost 20 years of investigations. \u00a0And it has taken me a long time to master them. \u00a0They are much more than methods, they are skills, and only through experience do you have any chance at getting better at them. \u00a0But it is a crap-shot at improvement if you don\u2019t reflect on them. \u00a0I hope that if you haven\u2019t thought about these before that on your next project you will and start \u00a0to hone your skills and get better at design research. \u00a0You shouldn\u2019t be satisfied at just knowing a method, but you should hone your skills, better yourself, better your research, and better your results, and in turn better the design that comes out, and better the lives of the people who the design effects. \u00a0They are not methods or tricks, they\u2019re how to reflect on your design research and how to skill lessons. \u00a0Once you understand and apply these to any design research problem you are investigating, uncovering customer insights becomes much easier.<\/span><\/span><\/p>\n