How to Build Data-Informed UX Process For B2C E-commerce
User experience (UX) design has been the bedrock of most design team processes for years. Typically, teams heavily focus on qualitative methods to uncover customer needs and validate solutions. However, the need for objective, quantitative data in this process is growing. While the majority of teams (71%) report leveraging some quantitative data, almost everyone struggles with how quantitative data plays into the UX design process. And that’s a problem for customers. The result of that on the business side is lower customer loyalty, higher churn, and reduced revenues—all from a very fixable problem.
Data-Informed, Not Data-Driven
Quantitative data rarely tells the whole story when it comes to customer needs and wants. It might tell you that someone isn’t clicking the button 90% of the time, but observing a user and asking unstructured questions is still the best way to reveal why they aren’t clicking or how they got to the button in the first place.
Nevertheless, a quantitative, data-informed process reveals journey-wide trends that UX teams can then validate with qualitative analysis. Qualitative data captures a moment in time, which is why it’s so powerful for understanding motives and validating assumptions, but quantitative data helps us take a more holistic, birds-eye view of the journey.
UX researchers regularly need to “unlearn” previous assumptions so they can keep up with new trends and emerging ideas. Quantitative data provides an easy baseline that can be updated, repeated across multiple tests, and re-built for any context. The result is that UX researchers can ask questions based on up-to-date baselines, making the unlearning process easier because there is a new anchor to focus on.
When the UX design process becomes data-informed, the whole process becomes more sustainable, scalable, and traceable. This ties directly to creating unique digital experiences, something the vast majority of customers say is critical to their purchase decisions.
How to Build a Data-Informed UX Process
When it comes to creating a data-informed UX process, there are four key steps:
Step 1: Audit and foundations
The outcome of this step is to identify what data to measure and to set up hypotheses you want to test. Break down the UX process into individual elements, just as you would in a qualitative process, auditing for the following factors:
- Relevant: Is it relevant for the context?
- Attributable: Can uplift in this data be attributed to changes in the experience design, solely or at least in part?
- Testable: Is there a delta in the metric that we can test for?
From here, either research a baseline—if you have the customers or existing data—or look for third-party research to help you set up a baseline.
Step 2: Identify triggers and set up measurements
The outcome of this step is to make sure you have the tools or software necessary to measure key metrics. In particular, look at these three factors:
- Recordable: Have you put relevant analytics software in place to view, analyze, and visualize the data?
- Reliable: Is there a large enough sample size? Is there noise or unpredictable variance that you need to account or control for in the data?
- Valid: Have you ensured your analytics system is “data-friendly” (i.e. is it able to both capture data that is valid and confirm that its source is reliable)? Are there any biases or other obstacles to overcome to ensure data is attributed correctly?
From here, choose the methods you’ll use to collect data, such as: A/B testing, heat maps, traffic data, or clickthrough testing. Quantitative UX research methods vary widely depending on the context and goal, and some companies may benefit from outside advisors to help them set up the right experiments.
Step 3: Analysis and insight generation
Taking the hypotheses generated in step one, use an analysis of the data to strengthen or revise each hypothesis. Leverage any qualitative processes you have in place to more deeply understand the hypotheses and why the outcomes may have differed from your original assumptions. From there, build a story that explains the narrative and prompts next steps, looking at data that is:
- Unambiguous: Valid and reliable data that clearly strengthens or weakens each hypothesis.
- Prioritized: Focus on insights that create the most value if acted upon.
- Narrative: Connect quantitative data with qualitative understanding in order to tell a story about user intent and behavior.
Step 4: Action and future planning
After supporting next steps with both objective data and a user intent-based narrative, start the two phases of taking action:
- Design: Identify where proven patterns can be leveraged versus where bespoke experience patterns are necessary in order to make a change based on hypothesis test results. As well, develop new hypotheses about the change to assess the progress made by each experiment you deploy.
- Implement: Because each change is its own test, leverage a building methodology like headless architecture in order to easily integrate changes and run multivariate tests without disrupting current systems.
With a build that’s set up for testing and incremental improvements, the UX team can continue their research and optimize the digital experience over time, leading to improved customer experiences and business outcomes.
Data is the Bedrock of Enhancing Digital Experience
Traditionally, customer experience is considered a capital initiative when it launches—backed up by projections of ROI of increased sales, retention, and customer satisfaction—but maintenance and optimization are typically considered cost-centers afterwards. However, once the UX process is data-informed and linked to the whole customer journey, real brand differentiation can begin to emerge. When that happens, empirical data from data-informed UX processes provides an opportunity for measurable improvements. The result for businesses is that optimization becomes a function with measurable return for the brand.
Activating customer data is an iterative process that has a bearing on both a brand’s commerce operations and customer journey. And given the level of nuance that often accompanies optimization, businesses need a platform that can respond accordingly.
With a composable commerce architecture, where customer data sits at the forefront, businesses can both foster unique customer experiences across channels and rapidly respond to changing customer expectations and behaviors. While customer data is a key pillar of a composable commerce architecture, it’s one of three building blocks. Learn more about the tenets of composable commerce.
Product Manager, Myplanet
Brooke is a Product Manager at Myplanet, where she shapes the roadmap and priorities for our Composable.com™ Accelerator. She focuses on synthesizing inputs from design teams, customer teams, development, marketing, and sales, ensuring our accelerator is expanding capabilities to meet the needs of future-focused brands
Subscribe to Our Newsletter
Sign-up for our industry newsletter to learn more about composable commerce.