Pete Bernardo / User Insights / December 13th, 2013

Iterative Research: Better Data, Sooner

By now, we all know that user feedback is a vital component of a healthy website strategy. At 352, we have a strong belief in research and incorporating that research into our sprint cycles – sprint, review, iterate and sprint again. Just as important as gathering user data is your timeline for implementing changes based on the feedback you receive. After all, what use is collecting data if it goes stale before you incorporate it into your website?

Since I feel strongly that the cycles for gathering feedback need to get into shorter and shorter timeframes to accommodate the rapid rate of change of user behaviors and expectations, I was happy to provide some techniques that everyone can use to start getting better data sooner with iterative research.

Why?

If you haven’t noticed over the last few years there are a lot more devices, apps, and new types of interactions people engage with on a daily basis. And as more people become connected the behaviors of different groups continue to diverge slightly. With every new application, website or marketing initiative there will be a different audience. Understanding your particular  audience is critical to providing them the best experience as well as improving your opportunity for success. Google’s The New Multi-Screen World does a terrific job of showing how user behaviors have changed dramatically in the last few years.

So how do we get better feedback sooner? How do we adjust our workflow to better understand our audience, but also keep in mind that their behaviors and attitudes are always evolving? While also keeping in mind that the audience might change depending on the medium?

First, I think it is important to understand that you are never done. When you launch a product or campaign is the point that the project really starts. If you don’t have this mentality that you must always keep iterating you will fail in the modern marketplace. Geoff Wilson, 352’s CEO, gave a great talk recently about our workflow for projects and how it has helped our clients.

Since I wanted to focus on tools that anyone can implement right off the bat, these tools are only meant to be a starting point for you to gather and learn about your audience. You will only become an expert through experimentation, so don’t be afraid of doing it wrong.

Analytics

bar

The Goal: The “eye in the sky ” watching how your site is being used and how people are finding your product.
The Tools: Google Analytics, Adobe Marketing Cloud, KissMetrics and MixPanel

The first thing you need to get a handle on is how your current product is being used. What are the historical habits of your audience? What has brought them to your product? But before you do that you should take a step back and make sure your analytics track all of your product. If you have an app, are you tracking how it is being used? If you have a web app, are you also tracking how the site is being used after someone signs in as opposed to unregistered users? More data always tells a more complete story, so it is important to make sure that as your product, website or device support evolves you also make sure you maintain your ability to track in all the new places.

While tools like Google Analytics can certainly be implemented after you launch a new website or product, we feel it should always be baked into your development project. Being able to track growth and site usage should be a vital component of any launch, not something tacked on as an afterthought. Good data lets you prove design and development hypotheses, so it’s important to ensure you have solid data from the get-go.

Once your data is healthy, we want to use our analytics not only to track success, but also to inform the other tools I’m about to go through. By taking a deep dive and understanding how visitors are reaching and using your site, you can start to formulate theories you want to explore through other techniques.

Surveys

pieThe Goal: Find out what your users value without asking how they do something
The Tools: SurveyMonkey, SurveyGizmo and PollDaddy

With surveys, we want to measure Attitudes, not Behaviors.

Survey respondents tend to over/understate what they actually do, so you tend to want to ask about what they like, not how they would do something. Tools like SurveyMonkey, SurveyGizmo and PollDaddy are great low-cost ways to start collecting responses. You can add a call to action to  your website or email a list to seek responses.

Additionally, some tools will provide an audience to you for a cost. For example, SurveyMonkey allows you to specify different criteria for desired respondents (age, sex, location, device, etc), and will cost anywhere from $1 – $2.50 each. You can get really nuanced with your targets, but you may have to ask for a quote.

We have seen a lot of success with clients who have run surveys on their site before they engage with us. Simply asking users what they would like to see added or what is not working best for them can tell you where you should focus development – invaluable insight when building a custom site. If you’ve never written a survey, have no fear: Many of the services will have pre-built templates to give you excellent starting points, like these SurveyMonkey examples.

Remote User Testing

The Goal: Get a video back of someone using your product and hear/see them talk through what is good and bad about it.
The Tools: Usertesting, Loop11 and YouEye

Remote testing is a technique that really has taken off over the last few years. In the past, you would need a lab environment for proper testing, but now there are quite a few services that can provide deep insight into how your product is being used. That’s nearly instant feedback that surveys and analytic research just can’t compete with. And remember, you don’t have to wait for launch to have users test your site. User testing during development can provide invaluable feedback to your team, so don’t be afraid to let users see a site in the middle of development.

Like the survey tools mentioned above these services will also help you with pre-built templates to get you started. Whether you are testing for desktop, tablet or mobile (see below) you can get great feedback with out pushing your product public. Once you have the video feedback you can share with your team and make the appropriate adjustments. Just like the actual testing, seeing is understanding so watch the video below.

Split Testing

The Goal: Be able to test different versions of your website without your users knowing it and see which works better.
The Tools: Visual Website Optimizer, Optimizely and Google Content Experiments

You’ve probably heard terms like A/B testing and multivariate testing – each of these are Split Testing methods. This technique allows you to easily make changes – big or small – to your website to test which design or content choices help you reach a goal. For example, you can see if your signup button should be “Sign up”, “Register”, or maybe “Join for Free – Today.”

Like any sort of testing, you’ll need a decent sample of traffic to get good data, but you’ll be able to move forward with the version that works best. I love these tools because they do most of the hard work, so understanding web development is not a requirement to run these tests. I suggest that you focus on color, copy and emphasis. That is to say, play with the color of a button or call out. Try a few different headlines to see which work better. Lastly, reduce the content on a page and emphasize a certain message.

This long but excellent video about split testing from Google Ventures and Optimizely will show you how sophisticated and powerful split testing can be.

Better Sooner

All of these tools and techniques provide a great starting point for anyone, even beginners. But no matter what your expertise, it’s important to understand that you must incorporate user feedback into your process when developing, and then consistently seek it out in order to constantly optimize your site.

Photo credit: Alexandre Delbos