Building a New Product: Phase 1 – Measure for Success

Once you’ve done the hardest step, launching the product, you need to measure your users.  The reason to launch early is to see how users interact with your product or service and if you have a viable product.  To do this, you need to start tracking key metrics and user behavior.  It is important to focus on the activity and metrics that really matter in determining if your users are interested in your product as well as how they interact with it.  If your product depends on users registering and contributing to the community, then measure that.  To do this, we can use a variety of tools and services, free and expensive, using one or both of two methodologies.

Conceptual magnifying glass on a spot light - ...

Free Tracking Tools

  • Google Analytics – Great free tool from google that measures user behavior and traffic sources but also allows you to set goals to measure the KPIs that matter to you.  You should set this up immediately.
  • Userfly – A free user session tracking tool.  Records the actions of your users and lets you watch a video playback.  Gain great insight into your users by seeing exactly where they go on your page and where they hit stumbling blocks or are confused about the next step.
  • Feedback Form – Let users give you feedback and suggestions directly.  The insight gained from this qualitative data can be great.

Expensive Tracking Tools

  • Omniture SiteCatalyst – Robust tracking solution that is highly customizable and scalable.  You’ll need to add tracking code all over your site, but the reports and data mining possible are powerful.
  • Tealeaf – Records user sessions and compiles into reports, breaking down users by behavior types so you can see where users are failing to do what you expect.  You can then watch individual sessions to gain insight into what they were thinking to figure out why.  This service is great at combining quantitative and qualitative data, but can be very expensive.

3D render of test tubes in rack

Testing Methods

  • A/B Split Testing – Present one group of users with one experience and another with a slightly different experience and measure the groups against each other to determine which experience is more effective.  In an ideal world, you would A/B test everything before you put it on your site.  This means everything from the color of a box outline, the placement of a button or link, to the wording of a caption.
  • Multivariate Testing – By testing multiple variables at once, you can gain more insight into how multiple elements or experiences affect the metrics in less time and users than with multiple iterations of A/B testing.  You can actually test as many variables as you want, but the more you test, the more users it takes to get statistically significant results.

Leave a Reply