UX Magazine

Defining and Informing the Complex Field of User Experience (UX)
Article No. 620 February 15, 2011

Getting Guerrilla With It

Who among us hasn't found themselves working on a project that had so much potential to be great, but didn't have enough time or budget to do the research needed to inform the design?

Research is a wonderful thing. It gives us the ability to inform our designs with reality. Through research we find out what real users think and how they behave when they're in certain scenarios. We find out how right or how wrong our assumptions are. Research allows us to see the difference between intention and reality once our designs move from concept to production. Research provides direction in place of guesses, and data rather than opinion or speculation.

But this doesn't mean that we get either the time or budget to conduct research. In many cases, research is perceived as a very expensive and time-consuming effort that eats up weeks, if not months, of a project. Too often, it's seen as less expensive to fix a mistake post-launch than it is to engage in the research prior to beginning the design phase. Of course, no one really talks about how mistakes in the design are going to be identified, or how and when they'll be fixed.

For years, many of us have worked in environments where research wasn't included in the design process. The design process was driven by a list of project stages tied to a budget, and hat list didn't include research. And why should it? After all, if we're UX experts, why do we need to do research? Shouldn't we just know the answer already?

The mentality that research is an unnecessary part of the design process—something that has a cost without any real value—puts unwarranted pressure on designers to get their work right the first time. Without any kind of research there's no data to back up a design decision, so product design can become driven by committees and subjective opinions. With research, designers have data to fall back on—something more quantitative and measurable. Research-led design is the holy grail of design; it's art + science.

Guerrilla Research: The Gateway Drug

When it comes to research for information architecture, interaction design, and UX generally, some is better than none—and often it's a fight to get any at all. Guerrilla research methods are faster, lower-cost methods that provide sufficient enough insights to make informed strategic decisions. Guerrilla research can be a lot more palatable to many bosses and clients who struggle with understanding the value of research. In fact, using these methods is a great way to introduce companies to the value that research can bring to projects.

Often when stakeholders push back against research, they're not saying "no" to research; rather, they're saying they're not willing to spend a lot of time and money to see results. They are saying, "What can you do that doesn't cost us a lot of money and time, and still helps us keep from getting egg on our faces at launch? We'll weigh our options." (Or something similar, at least.) If you are fortunate enough to hear this, you have an opportunity in front of you, and if not, well, it doesn't hurt to ask for the opportunity.

Guerrilla research methods aren't a replacement for The Big Research, but they're better than no research. But if all you have is Big Research in your arsenal, you're chances of getting any data to inform your design are slim-to-none when you are battling against time and money limitations. If you can get valuable results more quickly, then you have a chance to build at least some research into your project, and to demonstrate its value to your stakeholders.

All projects move at a fast pace—and if you think we're kidding, sit in a scoping meeting with a client and listen to them negotiate and bargain about their own ability to provide feedback and approvals. Time is of the essence—we all get it. But once stakeholders have had that first taste of the difference that a little research can make, they're hooked for life.

Characteristics of Guerrilla Research

Guerrilla methods are generally differentiated by three characteristics—rigor, time, and cost.

It's important to note that you won't get the same rigor with guerrilla methods; there tends to be just enough rigor to determine that a problem or opportunity for improvement exists. And "just enough rigor" can be difficult to grasp without the context of the audience that is your user base. Guerrilla recruitment methods do not often benefit from multiple rounds of qualified candidate filtering; in many cases, the user base for your research may be as basic as a group of people who you know are available. Ideally, your participants will fall well within the middle of known user types, of course.

Time and effort are probably the biggest factors your client or project sponsor is concerned with. Deadlines are deadlines, and guerrilla research methods give you a way to incorporate research into your design process without derailing them.

Let's look at usability testing, for example. In-person usability testing, whether at labs, homes, places of employment, or other locations, can take a lot of time and effort. In-person testing presents challenges for scheduling, travel time, setup, paperwork for clearance, and other administrative minutiae that all to add to the cost meter before the actual testing even has a chance to begin.

A guerilla approach to usability testing would be remote testing. Remote testing removes many of the challenges related to scheduling, travel, paperwork and setup. You can literally shave off days, or in some cases weeks, with remote testing. That savings in time and effort does come with a trade-off, though. With remote testing you will lose some level of additional insights from the participant's environment, the things you cannot see or experience if you're not right beside them (e.g. the Post-It notes on their screens, their little black books of passwords, the fact that one of their computers sits on a table in the hallway, etc.). The trade-off is usually an acceptable one—better to have some level of reliable data to inform your design than none at all.

So What?

The assumed or unknown costs of research makes many clients and bosses apprehensive, and in many cases that's reason enough for research to be skipped entirely. Guerrilla methods make research more affordable and can show the value and impact that can be had on projects.

Guerrilla research methods aren't cheap and dirty research—they're just less expensive and faster than non-guerrilla methods (a.k.a. Big Research). They may involve less rigor and they take less time and cost less, but they still yield high-quality results. If you're being told that research isn't an option in the work that you're doing, investigate more to find out if you can negotiate for smaller amounts of additional time and money and explore the options that guerrilla methods have to offer.

After all, some good research is better than none.

 

Russ Unger and Todd Zaki Warfel will be be teaching a workshop on Guerilla Research techniques at UX London this April.

ABOUT THE AUTHOR(S)

User Profile

Russ Unger is a User Experience Director for Happy Cog, a web design firm in New York City, Philadelphia and San Francisco. He is co-author of the book A Project Guide to UX Design for Peachpit Press (Voices That Matter) and is co-authoring a book on guerilla research methods with Todd Zaki Warfel due out in 2011.

User Profile

Todd Zaki Warfel is President and founder of Messagefirst, a Philadelphia-based design research consulting firm, where he blends research and design to evolve products in innovate and beautiful ways. Todd is a dynamic speaker and storyteller by nature. He’s rarely short on details. Todd currently lives in Philadelphia, blogs at toddwarfel.com, and is author of the book on prototyping A Practitioner’s Guide to Prototyping.

Add new comment

Comments

39
38

@Fritz guerrilla methods, like anything else, should only been used when they're appropriate. If you're doing research and the audience or intent will require something more rigorous, then you probably should choose a different method. Most of the time, at least in my experience, guerrilla methods work pretty well and have sufficient rigor to identify a problem. Either way, select the appropriate method based on audience and intent.

@Michele what you've highlighted is actually more of a reason to use something guerrilla than lab testing. A few years ago, we were doing research for a large financial institution that provides rates on loans, mortgages, insurance, etc. We used a combination of remote testing and home visits. During both, we noticed a number of participants used pop-up blockers (this was before browsers had the feature built in).

Had we done testing in a lab setting, we would have missed the pop-up blockers. This was a critical finding, as the company was relying on pop-ups and pop-unders to serve ads and promote products. It allowed us to help them reshape their strategy for the site.

If you're focused on trying to have a controlled environment for testing rather than something that more accurately represents the real world, then I'd question the validity of the data you're gathering and the recommendations you can make from that. Just something to consider.

35
42

Russ,

This is a great article and I agree with your premise that guerilla research can inform design decisions. However, I see it differently when it comes to remote usability testing. In my experience, although the travel may be eliminated, for a one-on-one moderated test, recruiting and scheduling the participants takes the same amount of time. Similarly, I've actually had to do more extensive pilot testing to alleviate technical issues than I when I conducted a similar test in the lab. This is due to the fact that participants may be taking the test on a variety of computers, browsers and network speeds vs a few prototypes that are set-up, tested on a lab computer, and tweaked during the study if needed.

36
44

The only drawback I have found with guerrilla usability testing is when the results are inconclusive (due to lack of rigor, poor statistical significance or availability of quality participants). When results are presented to stake holders and they want to know what the "basis" of the recommendation is, I have to tell them that my recommendations are based on my interpretation of results but cannot provide them with quantitative proof. While I agree that some research is better than none; I would also say that sometimes guerrilla testing might be better used internally for the UX department's confirming or disproving of opinions.

40
37

Great Article

34
43

Great article! It’s sad that in-page links are not working :( When could somebody solve this? (I’m starring at the “under construction” text for 2 months now)