Stop Creating Data Products and Features That Lack Context

Jennifer Fraser | Scott Plewes | November 4, 2020 | 6 Min Read

If you're finding your data science investments and efforts aren't bearing much fruit, there's a good chance it's because your product team is using/analyzing data that lacks context. By humanizing your data and examining its context via Human-Centered Data Science, you can tap into the full potential your data has to offer.

As MedTech vendors jump headfirst into data science, exploring its possibilities and building internal data teams (perhaps even with executive support), some are finding the efforts of this practice aren’t bearing much fruit.

Recent data shows that data science project failure rates hover around 85% and a Gartner analyst predicts that “through 2022, only 20% of analytic insights will deliver business outcomes.”

If your data science efforts haven’t lived up to your expectations or if you haven’t been able to get business buy-in, it could be because you’ve fallen subject to one of the data science pitfalls we commonly see.

But there’s also a good chance that the reason why your data science efforts haven’t created any substantial business impact is because you’ve been analyzing data that lacks context.

By humanizing your data and examining its context, you can tap into the full potential your data has to offer and uncover important insights that would otherwise be missed. This data humanization and contextualization is what we call Human-Centered Data Science.

What is Human-Centered Data Science?

Simply put, we refer to Human-Centered Data Science as the collaborative integration of your research, design, and data practices both at a team level and at a wider organizational level.

Many MedTech product companies have a data science practice, and many may also have a human-centered research and design practice. However, in our experience, these teams are generally siloed from each other. They report into different areas of the organization and rarely collaborate as a part of the product delivery process.

This is a missed opportunity that typically leads to data being analyzed without context, and in turn, can lead to incomplete interpretation, misinterpretations being extracted from the analysis, or the building of flawed models.

Exploring how to better integrate your data science and human-centered research and design disciplines, through both processes and tools, can lead to a clearer understanding of not just what is happening based on the quantitative data you’re receiving, but also why it’s happening.

This “why” space is often where you can find insights and context that can allow you to make better data-informed decisions to increase the value your medical product delivers.

Quantitative data may tell us a lot about what is happening, but not why, especially when it comes to human needs and behaviors.

For example, say you’ve recently launched a device update that allows your device to collect patient data in real-time for surgeons. This data is displayed on a dashboard in the device monitor.

After launching the update, your analytics show that surgeons are barely using this new feature you worked so hard to launch. After studying surgeons first-hand and collecting qualitative data, you uncover that they only find one metric in this dashboard useful; the rest is distracting. Or perhaps you’ve created the perfect dashboard, but your research reveals it’s located in a place that’s difficult for surgeons to access or find, so they never end up using it.

By conducting this qualitative user research earlier in the development process, you would have been able to make smarter product decisions that would enhance the value of your device for your end-users.

Examining Human-Centered Data Science Through Three Lenses

We can think about Human-Centered Data Science as an entire practice and how it’s implemented across three dimensions:

1. Integration Between Disciplines

This refers to the integration of your data science and human-centered research and design disciplines in terms of how they work together and collaborate with their tools, processes, and deliverables.

Ask yourself:

Is there any integration between your data science and your human-centered research and design practitioners?
Are they only exposed to each other’s work through presentations that are shared broadly within the organization? Or, do they work with each other on projects and share their data?
Are they so integrated that their deliverables have started to have inputs from various disciplines?

2. Incorporation Into Product Delivery

This dimension of Human-Centered Data Science refers to the incorporation of the two disciplines (data science and human-centered research/design) into the organization’s product delivery methodology. It considers how the work is planned, when it happens within the product cycle, what types of activities are conducted, and how their outputs inform future versions of the product.

Ask yourself:

Are your data science and human-centered research and design practitioners doing reactionary work based on market pressure and informal feedback from internal subject matter experts (SME’s) and stakeholders? Or, is it based on research activities and data analysis that’s being done both prior and post-release of the product to monitor its use?

3. Leadership and Culture

This dimension looks at how both data science and human-centered research and design are reflected in the leadership and culture of the organization. It considers the types of roles that may exist within the organization under these disciplines, as well as how their work is supported and valued across the organization.

Ask yourself:

Is your data science and human-centered research and design work being done by “unicorns”? In other words, is your data science, data engineering, and analysis all being done by the same person?
Is your interaction designer also doing the user experience research to “test” the validity of their own designs? Or, do you have qualified experts each with distinct specialties within the data science and research/design disciplines?
Is there cross-domain knowledge and fluency across disciplines amongst these experts?

How Well Do You Implement Human-Centered Data Science?

Okay, so we know that Human-Centered Data Science can help optimize the outputs of your data science projects by adding context to your data, but how well are you currently leveraging the practice? And what can you do to become better at leveraging it?

You can implement Human-Centered Data Science as a practice in varying levels of sophistication across each of the dimensions we just discussed. Meaning, it’s possible that your team could excel in embracing and implementing Human-Centered Data Science with respect to one dimension, but maybe not with respect to the others.

We’ve created a framework that breaks Human-Centered Data Science maturity into five levels across the three different lenses or dimensions of examining HCDS.

Our suggestion? Worry less about categorizing your team under a certain overall level of maturity, and use this as a tool to identify the elements you need to improve on to realize the potential of your data and research/design teams and outputs.

Stop Creating Data Products and Features That Lack Context: Human-centered data science maturity model
The above chart provides some high-level descriptions of each maturity level. But if you want to dive deeper into what these levels look like across each specific dimension of Human-Centered Data Science, then take a look at the linked articles below.

Reading these will help you understand where your product team stands today in terms of using Human-Centered Data Science to the fullest and what they can do to build a stronger, sophisticated practice:

  1. Integration and collaboration between your data and design disciplines
  2. Integrating data science and human-centered design into your product delivery process
  3. Creating a leadership structure and culture that supports data science and human-centered design

If you’re still not clear on why you should be giving Human-Centered Data Science some serious thought and attention, here are six reasons why applying it can help you create better MedTech products.

Download: Introducing a Human-Centered Data Science Maturity Model

Assess your Human-Centered Data Science maturity as a first step towards increasing the collaborative nature of your data and human-centered research/design practices. Fill in the form below and we'll email you a copy of the ebook.

Suggested Stories

Increasing Patient Engagement Using Behavior Design

Increasing patient engagement is easier said than done. In this video course, you'll learn how thoroughly understanding the behavior of health consumers can allow stakeholders to increase patient engagement.

Read More

Voice UI Design Best Practices

Voice assistants are poised to transform customer engagement as well as business models. Discover why voice is the next digital frontier – and what you should know about voice-first solutions.

Read More

Structuring Multidisciplinary Software Teams

5 strategies we've learned from working with the biggest names in software for structuring multidisciplinary software teams to get amazing software out the door fast.

Read More