Applying Six Sigma to UX Design

6 min readDec 2, 2015


UX design is an artistic endeavor. Interview users, draft personas, map journeys, sketch wireframes, design graphics, handoff to devs, rinse and repeat. Sprinkle in opinions from execs, product managers, engineers, friends, family, the random guy at Starbucks that always seems to be looking over your shoulder, and you’re basically molding a rowdy high school band into the orchestra that is your final high-fidelity masterpiece. That’s art, plain and simple.

UX designers, much like architects and other functional artists, face a dichotomous creative challenge: success is measured quantitatively. What’s more, there is a huge disconnect between the UX design process (pre-ship) and the data used to evaluate and optimize UX success (post-ship). Apptourage closes the gap analytically, and six sigma offers the perfect framework for incorporating analytics into the design process in order to integrate the art and science of UX.

What is Six Sigma?

At the highest level six sigma is a mentality, a quality mantra. It’s the idea that any process or product can be improved by collecting, analyzing, and learning from quantifiable data.

For the historians in the crowd:

“Six sigma” was originally a manufacturing term used to measure quality; it literally meant yielding less than 3.4 defects per million units produced. From a statistical standpoint that equates to the area under a bell curve outside of six standard deviations (or sigmas) away from the mean.

If you’re like me, the terms “bell curve” and “standard deviation” make you want to dart in the opposite direction. We’ll check statistics at the door, and instead focus on the six sigma methodology that is in fact relevant, and arguably critical, to effective analytic-driven UX design.

Using the Methodology

DMADV (Define, Measure, Analyze, Design, Verify) is one of the core six sigma methodologies that translates perfectly to lean software design and development. Below we’ll dive into each stage of the methodology, and how it can be adapted to optimize UX design.


Define the customer need. Great companies have leaders that set company vision based on customer needs, both current and future. If you’re lucky the leader of your company is an engineer, artist, rainmaker, and customer advocate rolled into one (see above). He or she understands the importance of iterative customer development and empowers the organization to explore needs that customers don’t even know they have, until they’re solved (see every Apple product).

In many cases, however, UX designers are the first line of defense when it comes to defining the customer need. It’s not news that conducting customer interviews, creating personas, and mapping journeys is the industry standard for identifying customer problems and potential solutions. What’s important to remember here is that customer development never stops. Customers change, as do their needs. Your job as a UX designer is to keep your customers at the beginning, middle, and end of every design decision you make.


Measure key customer interactions. Once you understand your customers’ needs in the previous stage, you’ll want to use Dave McClure’s pirate framework to measure acquisition, activation, retention, referral, and revenue in your experience (see video above and SlideShare here). Each pillar of the pirate framework can, and should, be broken out into success-defining actions that you can then analyze in your analytic tool of choice (e.g., acquisition success = app download, activation success = create account, etc).

There are many tools you can use to track success-defining actions, some of the big dogs are Google Analytics, Adobe Analytics, Mixpanel, KISSMetrics, Heap, Flurry, GoSquared, and Quantcast. Choosing the right analytic package for your company depends on a number of considerations. It’s important that you understand the alternatives and considerations, but ultimately your job as a UX designer is to interpret the data collected in order to inform your design process.


Analyze conversion bottlenecks in production. There are two key steps to understanding your customer data: (1) setting up conversion funnels, and (2) running cohort analysis. Ultimately you’ll want to see how different groups of customers (cohorts) convert through a series of success-defining actions in your experience (funnels) for each pillar of the pirate framework. The goal is obviously to increase the percentage of customers that make it through any funnel you create.

There is no golden ticket here. Fifty percent conversion isn’t necessarily bad, and 80% conversion isn’t necessarily good. It all depends on how you’re acquiring customers, how you’ve set up your funnels, and ultimately how valuable your product actually is. Your goal as a UX designer is to target the lowest converting node in your most critical funnel, and iterate your designs until you start to see de minimis increases in your conversion percentage.


Design a solution to improve conversion. Finally the comfort zone, creative design exploration. We all know how this goes: research how other products have designed similar experiences, draw design inspiration from Dribbble, Smashing Mag and others, create wires, get team feedback, design high-fidelity mocks, get customer feedback (potentially), send to developers, and either wait to analyze production analytics post feature ship, or start working on the next experience.

The hardest part about this process is knowing when you’re done. Whose opinion on the team is most valid? How much feedback do you need from customers before you can depend on it? When are you ready for your developers to start coding your designs? Today it’s a judgment call. You need to design, build, deploy, and re-evaluate funnel conversion percentages in production before you know whether to iterate your designs once again. We built Apptourage to solve this exact problem.


Verify UX effectiveness with pre-code analytics. Apptourage allows you to see conversion analytics on new designs before having to code features. You can collect analytic feedback from your own team, test design conversion with real user cohorts on demand, and leverage A/B testing to optimize your UX before writing a single line of code.

Depending on developers and production analytics to guide iterative design decisions is, unfortunately, a huge time and resource drain. We’re looking to change all that.




Precode UX Analytics for Designers. Sign up for free at