TikTok·Product Manager·Onsite - Product Sense / Strategy·Intermediate

Intermediate

Jun 2026

Summary

PM interview at TikTok, got a estimation question about the social graph. Pretty niche angle for a product role but I guess they care a lot about how you reason through platform data.

Questions Asked (1)

Estimate the number of mutual follows the average TikTok user has.

Product Analytics & MetricsProduct Sense & Ideation

Author's notes

I started with DAU, then tried to segment users by how active they are since a casual lurker and a creator have wildly different follow graphs.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Break down the problem by defining 'mutual follows' and estimating the average number of followers and following per user, then use probability to find the expected number of mutual follows. Consider TikTok's social graph characteristics, such as its emphasis on content discovery over social connections, which may lead to lower mutual follow counts compared to other platforms.

Pro tip: Acknowledge that mutual follows are a measure of reciprocal relationships, and in TikTok's algorithm-driven environment, many follows are one-sided (e.g., following celebrities). This shows you understand the platform's unique dynamics.

1. Define the metric

Clarify what 'mutual follows' means: a relationship where both users follow each other. This is distinct from one-way follows or friends.

2. Estimate average following and followers

Estimate the average number of accounts a TikTok user follows and the average number of followers a user has. Consider that these distributions are skewed: many users have few followers, while a few have millions.

3. Model mutual follows

Assume that for a given user, each of their follows has a probability of following them back. This probability can be estimated based on platform norms (e.g., reciprocal follows are more common among friends than celebrities).

4. Calculate expected value

Multiply the average number of follows by the probability of reciprocation to get the expected number of mutual follows. Adjust for skewness by considering median values or segmenting users (e.g., casual vs. creator).

5. Sanity check and refine

Compare your estimate to known benchmarks (e.g., average Facebook friends ~200, but TikTok is more asymmetric). Refine by considering TikTok's focus on content consumption over social networking.

Key Points to Mention

TikTok's social graph is more asymmetric than symmetric; many follows are one-directional (e.g., following creators).
The distribution of followers/following is highly skewed; using averages can be misleading, so consider medians or segmented averages.
Reciprocity rate is likely lower than on friend-focused platforms like Facebook, due to TikTok's entertainment-centric nature.
Active users vs. passive users: passive users may have very few mutual follows, while active social users may have more.
The estimate should be a range, not a single number, to account for uncertainty.
Tie the estimate back to product implications, e.g., how mutual follows affect features like DMs or friend suggestions.

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