Data-Driven Design

Data-Driven Design

Regardless of your goals, data-driven design can help creators move beyond best practices create digital experiences that convert. Many creators project their own behavior and assumptions onto their users and make decisions exclusively based on their own thoughts and experiences. Data-driven design helps creators get a better understanding of their user's needs, and high-converting products serve the user's needs first.

Getting Started with Data-Driven Design

Building a design strategy around data requires adopting a specific set of principles to ensure that you use the available information correctly.

Here's what to do:

  • Establish Realistic Goals. You'll have to consider feasibility, cost, timing, and other factors once you start applying the information you plan to gather. It is best to make data-driven decisions right from the onset. You'll need to consider your organization's unique needs when deciding how data can help you.
  • Create a Hypothesis. Now that you have established your goals, you can begin to develop a hypothesis. A hypothesis must be able to be tested. You can construct hypotheses as "if X, then Y" statements. If the X condition is met, then the Y result will occur. It refers to a cause-and-effect relationship between two factors.
  • Ensure a Sufficient Sample Size. Determining a hypothesis's authenticity as it applies to a large population can be impractical for many reasons, so it's common to determine it for a smaller group called a sample. However, if your sample size is too small, you will increase the margin of error, and the value of any data you gather will be unreliable.
  • Eliminate Confounding Bias. Confounding bias is the result of having confounding variables in your model. A confounding variable (also called a confounder) is an extra variable that you didn't account for. Confounders can ruin an experiment and render your results meaningless. In an experiment, the independent variable typically affects your dependent variable.
    • For example, if you are researching whether adding a video to your product description leads to more purchases, then the video is your independent variable, and the purchase action is your dependent variable. Confounding variables are like extra independent variables that also affect your dependent variables.

Quantitative Data Collection

Data made of measurable quantities, values, or numbers is called quantitative. It is often seen as more reliable and objective due to the use of statistics to produce and analyze it.

The following are examples of quantitative data collection:

  • A/B Testing. A/B testing is an experiment where you separate your audience into two groups to test several campaign variations and determine which performs better. Typically, you would show version A to one half of your audience and version B to another. When performing an A/B test, it is imperative to ensure that only one variable is changed and that the control and experimental groups are similar in size.
  • Surveys. A good survey should avoid leading questions and ensure that the question's purpose is clear.
  • Heat Maps. A heat map is a data visualization technique that shows areas of high interest. Some use eye tracking to understand where users are looking on a screen. Others may track clicks/touches. Patterns derived from heat map testing can be beneficial when re-organizing content assets or redesigning a UI.

Qualitative Data Collection

Data that is descriptive rather than numerical in nature is called qualitative. Unlike quantitative data, qualitative data is generally not measurable and mostly gained through observation.

Examples of qualitative data collection methods are:

  • Interviews. Interviews are a great way to gather qualitative data from users. Although time or budgetary constraints might limit the number of interview subjects, the insights gathered through conversation will be more in-depth than what you could get from a survey alone.
  • Competitor Analyses. Competitor analysis involves examining another company's product to identify any comparative strengths, weaknesses, or areas for improvement. Simply imitating competitors is not always practical. Instead, use competitor analyses as a means of gaining inspiration, with the understanding that what works for others may not always work for you.
  • User Flow. User Flow is a data model used to conceptualize how users are interacting with your product. The information you gather from your user flow can help you identify potential weaknesses.

Data-Driven Creatives Use Case

I have used data-driven design in my work with a California based healthcare provider. The company needed a strategic plan to better reach their target audience and be the first brand patients considered when searching for primary care. Using quantitative research methods, I first identified their target audience and then created a data-driven plan which focused on enhancing patient touchpoints.

I overhauled their website's UX and UI features and used data insights to make it easier for patients to interact. The results were spectacular.

A full case study on the project will be available soon.

Posted in Digital on Dec 10, 2020.

Share this article:
divider graphic
arrow-up icon