Stop using Postman in your interviews
Recently, I was conducting a pair programming interview that required writing some code that interacting with an API. I’ve run through this problem several times and in the last year I’ve seen an uptick in the number of people that use Postman to help them check routes and see the data returned. At first, I appreciated it — candidates were taking the time to build their mental model by looking at the data and walking through it. Last week I came to the realization that it was a crutch at best and an anchor for a cruise liner at worst.
What is Postman?
Postman is an awesome tool for hitting network endpoints that helps you build your API and diagnose issues easily. It’s great, it’s awesome and I definitely recommend using it… just not for interviews.
Why not?
The core idea behind using postman is right — a successful interview requires building a proper mental model of the problem you’re trying to solve. Otherwise, you’re trying to build a solution by slapping json blobs on top of each other like stickers on the back of your laptop (you know who you are). Are you working with a graph of nodes? Are you dealing with streams of data? Do you need to handle pagination? Answering these kinds of questions is vital to success in solving your interview questions.
But postman is wholly overkill in getting to this point. Consider the following code:
A straightforward, simple piece of javascript that does the exact same thing as postman. So why is this better than using postman?
You’re solving the actual problem.
Getting results and data in code means that you’re already halfway through solving your problem. Postman makes it easy to send requests and get results, but translating that to code takes additional effort because Postman does most of the work for you. Adding headers, authorization, body parameters is just different and so you run the risk of wasting time trying to finagle things into Postman to get a successful response.
On average I’ve seen candidates spend anywhere from 10–30 minutes of a 90 minute interview poking requests with Postman. Then many of them spend another 20 minutes trying to get the same response in the code they start writing.
Ask the right questions
Always make sure you have the right mental model when you start working on a solution, but build it by asking the right questions and using tooling that will *help* you in your interview. Postman is a fantastic tool for developing APIs, but without proper expertise it can really slow you down in an interview. Leave it at home when you are working on an interview solution.
P.S. One tool that actually impressed me in an interview: Ngrok. A candidate was building a frontend solution and shared a url that I could follow along and poke with.
