← Pinterest Interview Insights

Pinterest·Data Scientist·Technical Phone Screen·Intermediate

Intermediate
Jun 2026

Summary

Pinterest data scientist interview with a product analytics case around out-of-home advertising reach estimation. Pretty open-ended, more of a structured thinking exercise than anything technical.

Questions Asked (1)

Q1

An out-of-home advertising campaign wants to measure reach. How would you estimate how many people actually saw a specific highway billboard, and calculate the total impressions it generated?

Product Analytics & MetricsAdaptability & Ambiguity
Author's notes

I went straight to traffic volume and kind of forgot to think about segmentation for a while.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Break the problem into two parts: estimating the number of people who pass the billboard (traffic volume) and then estimating the proportion who actually see it (visibility). Use a combination of public data (e.g., traffic counts), assumptions, and a model to calculate total impressions. Acknowledge uncertainty and suggest validation methods like eye-tracking studies or mobile location data.

Pro tip: Mention that impressions should account for frequency (multiple exposures per person) and that you'd calibrate your model with ground truth from a small-scale study. Also, tie it back to Pinterest's visual nature and how OOH could drive brand awareness or app installs.

1. Define the metric and scope

Clarify what 'reach' and 'impressions' mean in this context: unique people vs. total exposures, and the time frame (e.g., daily, weekly).

2. Estimate traffic volume

Use public data (e.g., DOT traffic counts, Google Maps API) or partner with data providers to get the number of vehicles passing the billboard location.

3. Estimate visibility rate

Determine the percentage of passengers who actually notice the billboard, considering factors like billboard size, angle, traffic speed, and time of day. Use industry benchmarks or conduct a small-scale eye-tracking study.

4. Calculate impressions

Multiply traffic volume by average occupancy per vehicle and visibility rate to get total impressions. Adjust for frequency if estimating unique reach.

5. Validate and refine

Validate assumptions with A/B tests, mobile location data, or surveys. Iterate on the model to improve accuracy.

Key Points to Mention

  • Traffic volume estimation using public data sources (e.g., DOT, StreetLight)
  • Average vehicle occupancy (e.g., 1.2-1.5 people per car)
  • Visibility rate factors: billboard size, angle, traffic speed, time of day, weather
  • Use of eye-tracking studies or mobile location data for calibration
  • Distinction between impressions (total exposures) and reach (unique individuals)
  • Frequency adjustment for multiple exposures per person

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.