Using detailed interviews to build personas

This is the fifth post in my Customer Research series.

There are many awesome posts on personas, including this primer by Steve Johnson at Pragmatic Marketing, and this post about how personas serve the CEO and the executive team.  So I am not going to belabor the point.  I’ll simply provide a case study of exactly how I go about building personas.

Before we begin, we must decide whether we are building buyer personas or user personas.   Since I have always worked on the PD/PM side (instead of the PMM side), I almost always end up making user personas first.   Here is the process I use.

1.  Pick a market segment.  Size it.  Make sure the total addressable market (TAM) is big enough to matter.

2. Generate prototype personas.  Draw on your company’s tribal knowledge / theories about customers and prospects, as well as from your own life experiences for this.  Make sure you state assumptions about needs, wants and expectations.

3. Create a screening questionnaire.  Work with a recruiter to find subjects.  Evolve screener as necessary (tweaking the prototype personas as needed).

4. Plan the interviews.

• Pull together the equipment: a mini DVD camcorder, a tripod, fully charged batteries, camera, laptop for notetaking. (Mini-DVD camcorders result in the least postprocessing overhead.  Finalize the disk and you can watch it instantly.)
• Staff each interview with 1 researcher who will lead the conversation, and at least 1 support person (in the role of audio/video monkey and notetaker).
• Every researcher should plan on going to several interviews (overlapping with each other for some of them).
• Cycle as many staff members through the support role as possible. This exposes everyone to the process and gets people out of the office and into customers’ environments, which is always good.
• Plan mid term and final debrief sessions with the research team.
• Plan other work around this work.  It takes at least 1 – 1.5 person days to process each interview. You also need time to reflect on the results.  Expect each researcher to be completely consumed for a good 75% of the time.
• Share intermediate work products with the team early and often.  Transparency is key to buy-in.

5. Do the interviews. Here are some tips I’ve collected over the past 14 years (your mileage may vary).

• For best results, detailed interviews should be done in the target environment for product use. That said, an interview at a coffee shop or some other neutral environment is better than no interview at all.
• Start with an introduction and a warm up section where the subject gets comfortable with the researchers, then ease into the topic of interest.
• Keep the interview guide short and sweet. Questions should be open ended, inviting the interviewee to talk in their own style and to show their personalities.
• Be prepared to take the conversation where it wants to go, not where you want it to go.
• Make sure the subject talks more than you.
• For certain types of products, you may have to match gender. (e.g. research around a feminine hygiene product will require a female interviewer).
• Don’t even think about showing a product or product concepts. Discuss your product at the end only if there’s time left over.

6. Build the personas.  With luck, you will have recruited carefully to get the right people, and after 10-20 interviews, they will have self-organized.  You should now be able to build composite personas in each bucket of subjects who share key needs, wants, expectations, and attributes.  Check these personas against the people you met and tweak until they work well.

Once the primary, secondary and non-personas are built, you can begin to use them as a proxy to help you envision what benefits they require, what products or services can deliver these benefits, and how those products and services may look like.

Common ethnographic techniques

This is the fourth post in my Customer Research series.

The word “ethnography” has such a grand sound to it.  I’ve seen seasoned executives get intimidated by it. They say things like “Whoa – that’s big agency stuff! We don’t have budget for that!”  To that, my standard response is: “It’s not rocket science!  We can do this!”

And it really isn’t rocket science. It’s just hard, detailed work.  Lots of it. You don’t need a degree in anthropology or psychology. Anyone with an open mind and great listening skills can learn to do it.  If you can’t listen more than you talk… well, perhaps you shouldn’t be a product person after all.

And the budget can vary wildly.  You can spend $250,000 if you outsource it to a big agency and you have a global market.  You can train your staff to do it themselves under some guidance for well under $25,000, using a recruiter to find subjects for you.  I’ve run projects under $2500 where I did the recruitment myself using Craig’s List.  So your mileage may vary.

Now what exactly is ethnographical research in the context of product development?  In the simplest terms, it’s a set of qualitative techniques that places the researcher in the environment and/or the mindset of the subjects.  One listens and observes subjects in the environment that the product is meant to be used in, without showing the product or product concepts. One then derives the needs, wants, expectations and workarounds for the subjects, and uses this information to drive product definition.

There are three techniques that I am a big fan of (mainly due to their high bang for the buck):

1. Detailed Interviews. Researchers meet with a subject for one to one and a half hours.  A researcher asks open ended questions to get the subject to tell a story about the domain of interest.  The interview guide tends to be very high level and the researcher is trained to mix things up and respond to new threads that come up in conversation.  I like to have two to three researchers at each interview so one person can drive the conversation while the other person mans the audio/video equipment and takes notes.
2. Observation or shadowing. Researchers set up their audio/video equipment in the environment where a product or set of products is being used, and simply hang out with the subject while the subject uses the product.  The researchers asks questions when necessary, but by and large they behave like flies on the wall.  They are there to watch and learn, not to talk.  The subject is disturbed as little as possible.  This could be a multi-hour proposition – I once shadowed someone for 8h with a coworker.
3. Immersion. Researchers use the product or a related product for an extended period of time to get a first hand understanding of long term product use. Researchers can record their extended findings via a journal or a photo essay. I also like to write a debrief document at the end of the immersion to sum up my key takeaways.

There are other techniques too, based more on self-reporting by end users. Examples include journal studies or photo essays involving customers.  These are all valid approaches.  When used with detailed interviews and/or observation, they can help round out the picture of the customer’s problems, needs, wants, expectations and so on.

The biggest challenge for this technique is the amount of time it takes to do a good job. Since research is done one customer at a time, and best practices suggest we work with 10-20 customers to allow the data to converge and allow time for tweaking the technique and/or the recruitment criteria, it takes corporate level commitment to pull off a project like this.  It’s a lot of long nights too since sometimes the research must be done after hours when the customers have time to work with the researchers.  But if we keep the focus, and we recruit correctly, generally by the 5th or 6th interview or observation session, a pattern will begin to emerge.

You know you’ve nailed it when you start hearing the same thing from everybody over and over again.  It’s an incredible feeling to get there and know you’ve built the knowledge that will guide the product team moving forward.