Skip to main content

How to Create a Take-home for Open-ended Tasks

Premium

The primary way in which open-ended tasks are different from defined tasks is that they rely on you to define a goal or interesting area of exploration. We’ll describe how to do so for two scenarios:

  1. If the dataset is related to the product or business
  2. If the dataset is not related

Dataset is relevant to the product/business

If you’re given a dataset that’s relevant to business, try to evaluate it in ways that might be advantageous to the business.

To do this, think about what their business goals might be (e.g. more users, more revenues, lower costs, more engagement) and how you can do something with the data that might deliver that outcome.

Frame a few ideas, directly link them to these higher-level goals, and show how those goals will improve the company's business outcomes.

With these “problems” or ideas in mind, explore the data and follow the same approach you would for defined tasks.

Since you’re trying a number of ideas here, try to quickly assess their feasibility (e.g. checking if there’s enough data to do the kind of prediction you’re trying) before spending a lot of time getting into the thick of your approach.

This specific scenario is quite rare and usually indicates that the company really values creativity. With this in mind, try to think of interesting questions or ideas to explore.

For example, let’s say you get a dataset on weather. In that case, you might explore “how bad global warming has gotten.” Now, think about what you would need to quantitatively measure to answer that question. You can observe how much deviation “from expected weather” there is in the observed data. This could lead you to making forecasting observations, and then measuring how far the observations are from the forecasts.

Consider reaching out to the recruiting team to get guidance on what the take-home evaluation is based on.

Common pitfalls

These pitfalls apply to both defined and open-ended tasks.

  1. Failing to define a clear goal. Without identifying and stating what you want your work to address, it will be hard for the interviewer to follow your work.
  2. Not answering the problem statement or goal. Make sure any analysis you do and conclusions you draw actually address the core problem.
  3. Allocating too little time to understand the data. If you miss something significant here, you’ll have a hard time creating a compelling solution. However, try to avoid spending too much time on rabbit-holes, since most companies aren’t trying to trick you with the data.
  4. Not proactively stating your assumptions. The interviewer or hiring manager will lose sight of the logic behind your solution without a clear presentation of your assumptions.
  5. Spending too much time on the assignment. It’s very easy for any given task to bleed into several days. A good strategy to prevent spending too much time is to draw a hard cut-off after your second session working on the take-home.
Next Lesson