Product Experience Analyst (IC6) Interview Experience

Can you speak to the kind of work you do?

I kept this pretty high level and talked through the analytics work I had been doing and whether it lined up with what they needed. It did not feel technical at all. It felt more like they wanted to confirm I could communicate clearly and make sense for the process.

If we're sending emails to customers, how would you think about engagement with that?

I started with simple engagement metrics like open rate and clickthrough rate, then translated that into what I would need in the data to measure those cleanly. The interviewer clearly cared about speed, but he was collaborative and not trying to time-gate me. We spent a little longer getting to the right answer and still got through everything.

How do we determine if X is better than Y?

I did not jump straight into an experiment. I first set the objectives, talked about what the company was trying to do, how users interacted with the product, and what success would mean. Then I laid out the hypothesis and experiment step by step. When they asked about stratified sampling, I explained splitting a sample into balanced groups so you can test multiple variants with matched test and control cells. That follow-up was there to see if I really understood it.

Here is something I’m seeing right now. How would you think about it?

I treated it like a practical problem-solving conversation instead of trying to hunt for some perfect canned answer. The question was not meant to be insanely hard. It was more about whether I could structure a messy situation, think about it in a sensible way, and show how I would approach it on the job.

We have a dataset of advertisers, their spend patterns, and the kinds of ads they spent money on. Build an analysis table to show how much each advertiser spent across the year, maybe segmented by a dimension in the data.

I built it as a straightforward aggregation problem. I would total spend by advertiser across the year, then break it out by one of the available dimensions in the dataset, like the kind of ads they were buying. It was very practical. You did not need internal Meta jargon, but you did need to understand that advertisers and ad spend are core to the business.

For a product like Instagram, what exactly are we analyzing, what metrics matter, and how would you think about setting a target?

I first narrowed the scope because you cannot answer a question like that if you do not know whether it is about the whole product or one slice of it. Then I laid out primary metrics like daily active users for growth, engagement metrics like shares, reposts, or DMs, and guardrails like app crashes. The round was really about whether I could make sense of how the product functions day to day.