jonathan.kalinowski

My school, Savannah College of Art & Design, just let out for winter break, and I've been reflecting on what I've learned over the last 10 weeks. One class that I took was an exploration of Big Data. We answered questions through exercises related to Big Data and we discovered a lot. We even developed a way to fold Big Data into the double diamond design process!

So what'd I learn? Firstly, I learned that Big Data is a tricky term. When people use the term, they mean several things and these definitions are shifting as technology advances. Furthermore, the sources who write most about Big Data are more persuasive than academic. Nevertheless, I've synthesized the idea of Big Data into a couple of thoughts:

Big Data is about correlation, triangulation and interpretation.

What I mean is, Big Data is a catchall term for a set of practices aimed at distilling potential solutions (measured in probabilities). Also, Big Data outputs correlations which need to be reviewed by an individual. Conclusions need to be reached, and there is no formalized process of doing so.

Regardless, those practices represent a refurbished way of understanding the world: through the tactful use of correlation (much like behavioral economics).

Designers ask a lot of questions in order to understand the problem they are working to solve. Our class discovered that we had to prepare to use Big Data by asking specific questions that designers are not used to asking. We usually ask "why", but Big Data answers "what" questions.

So we developed a structure in which to funnel treated design questions. But first we had to determine which questions were important to a given client so that we could feed those questions into our format.

For example, in working with a local food co-op, we discovered that customer loyalty, delivery routes and perception of quality are three (of many) important factors to their business. (We discovered these qualities through user-centered research). 

Our format: "What could indicate xyz?" where xyz is the factor that we are interested in studying.

This format was extremely helpful in identifying potential databases for our process of understanding the problem. For example, while thinking about delivery routes, we determined that optimal routes may change depending on a plethora of conditions. We asked the question, "What could indicate optimal delivery routes for supplies?" and brainstormed some correlated ideas: forests have animals, animals run across roads, cars drive on roads, etc. Therefore, it was posited that an optimal delivery route should account for the existence of forests along a delivery route (and avoid them when possible).

We also discovered that Big Data is the input in a process of insight development. Big Data analysis methodologies/techniques transform that raw input into useful information. This rough IPO (input - process - output) framework allows the designer to understand his relationship to Big Data.

The designer can therefore augment his process to magnify the effect of his tools. Through a set of discussions, our class demonstrated that Big Data can be used to prepare or even change existing tools. For example:

In preparation: Blue Ocean Strategy's four action framework requires the designer to think about specific attributes of a business to augment. Big Data can help to identify the most pertinent attributes to change by sifting through a chosen set of databases. Important to note, Big Data focuses and magnifies the creativity of the designer.

In use: Real-time data can be analyzed and folded into existing data sets from self-documentation exercises.

In sum, designers stand to benefit from Big Data by folding it into the tools they use in their design process (ours is the double diamond).

Our class has identified two useful ways to think about Big Data. The first is through asking the right questions (our "What could indicate xyz" format) and the second is through understanding when/how to use Big Data to augment tools (in preparation or in use).

I am writing on this side of an design research study around how users of print technology organize their content. I've been reflecting on what I've learned and wondering if I can formalize some sort of a research plan for design researchers. Hopefully this will help to cut away logistical problems so that more time can be spent understanding the participants of the research study.

Ok, let's start.

Refer to the double-diamond design process as a way to guide your flow from project definition to project completion. The double-diamond design process formalizes a way to look at the design process in an intelligible way. This is a win for designers who want to be taken seriously by the business world, which desires predictable processes.

The first phase (discovery) of the double-diamond design process can be completed in as little or as short of a time as you would like, but here are some suggestions:

Spend one week conducting secondary research. This large time frame allows you to explore the corners and details of the subject that you are studying. Study journals, white papers, grey literature, blogs, articles, tweets, etc. You can remain loose and 'non-denominational' at this point, because we are trying to understand human behaviour around a specific thing (e.g. printing). Furthermore, one week allows you to reflect on the information that you've gathered.

As you comb through your secondary research, keep a pack of post-it notes near you to capture important quotes, facts, correlations, etc. Make sure that you note where you found the information, and remember - one idea per post-it note! You're setting yourself up for success if you take notes on post-its, because these will be transferred to a wall for affinitizing. Conclude secondary research on day 6 or 7 of the study with an affinity session. Put one post-it note up, divine the underlying concept and place other post-it notes that have some sort of similarity next to it. Sometimes it is good to put two unlike post-its next to each other to uncover gaps in information. The important thing is to get everything that you've learned, externalized. You can rest assured that you won't have to rely on memory.

On day 6 or 7, depending on how you're making out, choose some business tools to analyse and sort the information that you've gathered from secondary sources. Develop 2x2 matrices, market analyses, strategic canvases, SWOT analyses, five-forces analyses, trend analyses and cultural analyses. Spend some time googling these and don't be afraid to try new models!

Congratulations, you've completed the first step of an ethnographic study! Next, it's time to sift through all that information you've just affinitized and:

1.identify gaps in knowledge,

2. identify patterns in behaviour and;

3. identify correlations.

Be very strict with this part - it is the only way that you will be able to go into the field, feeling equipped to gather information. Spend one day (e.g. day eight) resting after affinitizing your secondary research, and then spend one day (e.g. day nine) sitting in front of your affinity map, writing down every question you have.

Guess what? It's time for another affinity session, this time centred around your questions! You'll eventually get really good at identifying patterns and spotting correlations, so don't worry if the process takes some time to learn - it will actually make you a better researcher as you learn to observe.

Here is where you identify you target group, as well. This is essential, because you won't know who to study during primary research. Using your business tools, you can identify gaps in markets, and create segmentations of the market using user behaviours to distinguish groups. For example: brand loyals, brand switchers, brand disloyals and non-users. This segmentation is based on the use of a product/service.

Make a choice, and live with it. It won't feel perfect, but it will have to do. You will discover more through a focused approach than you will if you try to study everyone.

So you have some questions, what now? There are several paths that you can take from here, employing various ethnographic methods, but my favourite way is to slowly move from a passive observer of something related to your problem to a more active one. I usually enter the field to shadow people completing tasks around the problem. I take copious notes and photographs to refer to later on in the process. I also spend 15 minutes after each shadowing session to write down my thoughts and anything that stood out to me. Sometimes, however, shadowing is insufficient. Take printing as an example: users don't print very often in their days, so it is important to economize time by making sure that I am shadowing or observing a directly or indirectly related activity. I used a task-analysis in order to nudge users into participating in a directly or indirectly related activity, and then was able to observe and shadow.

It is useful to have a framework (or a series of buckets) to organize your thoughts in. The AEIOU framework is a popular one, which stands for Activities, Environments, Interactions, Object and Users. No framework is better or worse than the other - just pick one and stick with it. This will allow you to eventually compare and contrast the categories that you choose to work with.

After each primary research activity, comb through your notes and prioritize the data that you want to put on post-its. Ask yourself, can I move on in my research if I disregard this data point? for each note you take, during the debrief session.

At this point, set up a place to deposit your findings. I suggest organizing your entire space around the framework you choose for this portion of primary research. For example, reserve spots for Activities, Environments, Interactions, Object and Users. After each primary research activity is completed, go to these spaces and deposit your post-it notes. Give yourself a LOT of room to work with - as much as you can afford.

As you place post-it notes up on the wall, moving systematically through you framework (e.g. AEIOU), move post-its around in a first-level affinitization. Doing this on the go will make you more familiar with your data points. You're probably already starting to make connections. If you have an insight at any point, STOP, write it down, and place it in an Insights pile. Make sure you refer to which post it you've derived it from.

This brings up an important point - make sure that you come up with a system to organize your research activities. If you shadowed Alex, who was the first user that you researched, you can put Sh-U1 or Sh-Alex at the top of the post-it. Just make sure you have a consistent reference system.

It's good to be sensitive to knowing what you don't know. Consistently build new ideas after each debrief. Ask yourself the question, What do I still need to know, going forward? This way, you are constantly moving forward in your process, never getting stuck in a rut. Get those questions answered in the next research activity you choose. At one point, you will have to say, enough!

A good rule of thumb is to debrief as soon as possible. For example:

Observe/shadow Alex at 1 PM on Tuesday for 2 hours;

pre-debrief for 15 minutes between 3 and 3:15 PM (drink a cup of coffee with your partner if you have one!).

Return to your workspace and debrief with your team (if you have one) for two and a half hours. The first half hour is dedicated to prioritizing your information, the next hour is about discussion and the next hour everyone puts their post-its up on the wall.

But how do you debrief? I'm glad you asked! After your research activity is completed and you've thanked your participant, spend 15 minutes discussing interesting points and questions you have with either your partner or yourself.

Note: it is important to have a partner with you to help the accuracy and depth of gathering data.

Once the pre-debrief is over, go to your workspace and prioritize data that you took in. Use a highlighter, for example, to highlight points you want to share. Once you're ready, gather the majority of your team together and start sharing. Your team is responsible for capturing what you say on post-it notes. You lead the discussion by answering their questions, and making sure that those questions are captured on a post-it note in a Questions pile. Often, conversations will generate unique questions to ask in the next research activity, so don't slack during this part.

So you've completed observation/shadowing exercises, now what? It's time for you to pause and reflect. Take one day to reflect and one day to generate specific questions that you have within the scope of your project. Up until this point, we've been learning which questions are important to the group that you are studying. It's time to develop those questions, and put them down into an interview protocol. Categorize these questions that you have into broad categories, for example: Social Networking Questions. During your more formal interviews, you will be able to have a nice conversation with your participant around these topics, gaining rapport, instead of drilling down a line of questions and extracting answers from the participant.

Note: choose three or four primary research activities to complete for your study; for example: observation/shadowing, interviews, card sorting (google it), artefact analysis, etc. This will give you a good basis to start designing in the future. Remember to be sensitive to how much you know and don't know. If you still feel like you don't know a lot, continue with primary research. You should be an expert when you exit this phase.

Create an interview protocol sheet with topics to cover, and recruit users who are relevant to your study. Use your network of friends and family, the internet and even strangers.

Note: sometimes you have questions that primary research can't answer. What do you do? Continue to do secondary research! Designers work non-linearly, so it's important to supplement your discovered ignorance with information that others have gleaned. Stand on the shoulders of giants!

After each interview, remember to complete a pre-debrief, and a debrief session with the majority of your team. At the end of primary research, you should have hundreds, if not thousands, of post it notes in your workspace.

I won't get into analysis and synthesis of this data - that's an entirely different post, but I hope that you can use my learned process as a starting point for your own design research study.

So, we've covered a lot, right? Let's recap:

- 1 week of secondary research ending with a day or two of affinitization and and a day of modelling your data points (e.g. with the help of business tools).

- No less than two weeks of primary research activities, framed by your insights derived from secondary research (e.g. target groups delineated via market segmentation).

- Stay sensitive to what you don't know; keep on conducting secondary research.

- Pre-debrief and debrief sessions as soon as possible after a research activity.

- Affinitize as you go.

I've conducted a primary research study in one week, and that was way too little time to be able to reflect on any information gathered. Give yourself enough time to work. Take breaks and limit debriefs to as little as one hour, or as much as two and a half hours.

This is by all means non-exhaustive. It is purposefully prescriptive because we all need a place to start. Try new things, ditch things that don't work, but remember, it's ALWAYS good to visualize the data and information that you take in, so that you can communicate it to a larger audience. Don't let this part run away from you!

Sherlock Holmes' process for solving mysteries is nearly identical to the design process: first, observations; second, analysis of observations; third, synthesize bits of data into actionable information (a story, mayhaps?); and fourth, implement a solution.

We could learn a lot from Arthur Conan Doyle's Sherlock Holmes. Especially in the observation phase -- so much of our observations are tainted by personal interpretation.