Product discovery — what can we learn from House MD?

Elena Borisova
4 min readApr 24, 2022

--

It’s nearly impossible to find practical examples of what good product discovery is.

So one way to learn and improve is to look at other industries and leaders. Even made-up ones — House MD offers a surprisingly useful guide.

But first, what is product discovery?

Any sustainable product design strategy comes down to this: achieving business goals by making the users’ lives better.

Discovery is the process of figuring out how — what are the pain points or unaddressed user needs:

  • What should we do better
  • What should we be doing we are not doing at all

What does discovery have to do with House’s practice?

Uncertainty.

To find a treatment, we need to identify a problem. And we don’t know what to look for and where to look for it. While there’s no clear path with guaranteed success, answering these questions will take you closer to making a reasonable discovery.

Discovery foundation — 4 questions to start with

With a high-level business goal like decreasing customer churn, the first step is to develop a set of hypotheses on what causes users to churn. Answering the following questions will lay down the foundation.

1. Where do users (or patients) explicitly complain about something?

To answer it:

  • Ask people directly what they think
    - Surveys, feedback forms
    Tip: don’t ask them what they want; ask them what’s working or not working
  • Look where users share their feedback naturally
    - CS messages, sales calls, App/chrome/… store reviews, forums
    - TrustPilot or similar review sites (these will be biased towards people complaining. Great for discovery, but do ignore overall score)

2. Where does the product fit in users’ lives?

Get a broader context around users’ product use and where we can cover more of their needs. This is when House’s team usually breaks into the patient’s apartment. Some legal ways of achieving the same for product discovery are:

  • Exploratory user interviews
  • CS shadowing

3. Where do we see pattern or anti-pattern? What doesn’t make sense? (House’s equivalent of “run a bunch of analysis”)

  • Look at quantitative data (such as GA) to look for unexpected funnel drops, surprising distribution, or behavior shifts

4. Where do people use the product in unexpected ways?

  • Conduct exploratory user interviews
  • Look at public resources like Twitter/Reddit.
    Here’s a DeepL user hacking DeepL for grammar correction
  • Recordings of user behavior e.g. Hotjar
  • Open survey questions

We found an issue, what’s next?

Answering the questions above will likely identify some issues and user problems. It’s tempting to jump straight into generating solutions, but is what we found a symptom or a root cause?

Figuring out root causes first gives a broader solution field and the ability to validate treatments more precisely.

Identify possible root causes first
For each cause there could be more than one treatment

Differential Diagnosis

What might cause the issue, and how should we treat it?

Differential diagnosis, for medicine or product discovery

Each possible route cause should be (in)validated. Whether you do it by learning more with research or by testing treatments against that cause depends on:

  • Cost of treatment (development time, eco system cost)
  • Cost of treatment being wrong (as it often happens with House, treating a patient for probable disease A will kill them if they have as likely probable disease B). In product development these could be losing users, wasting time, or reputational costs.

Start somewhere

There’s no “success guarantee” formula for product discovery. Looking at users’ complaints, regular and irregular behaviors helps to reduce uncertainty in product development.

--

--

Elena Borisova

Combining data and psychology for product and design decision-making | Head of Design at DeepL | elenaborisova.com