How would you adapt a growth model for network effects and k-factor?

WHAT THIS TESTS: This question tests whether you can move beyond linear paid-acquisition models and build a compounding growth equation that treats existing users as acquisition channels. The interviewer wants to see if you understand that network effects change the fundamental shape of growth from additive to multiplicative.

A GOOD ANSWER COVERS: First, define the k-factor mathematically as invitation rate multiplied by conversion rate. Second, explain the critical thresholds: a k-factor above 1.0 produces exponential growth where each cohort generates more than itself, while a k-factor below 1.0 means virality only offsets churn and you still rely on paid spend. Third, introduce viral cycle time as a multiplier on the k-factor because more cycles per period compounds faster; you should note that Dropbox achieved roughly 0.65 from 2.6 invites and a 25 percent conversion rate, yet scaled to four million users in forty months by compressing cycle time from sixty days to seven days. Fourth, state that you do not need k above 1.0 immediately; a k above 0.7 with a short cycle time and a power-law distribution where the top 20 percent of users drive most invites can still yield sustainable compounding. Fifth, anchor the model in product moments of peak user delight rather than forced prompts, such as Slack triggering invites after a user misses a notification in a public channel or Calendly embedding virality directly into the calendar invite itself.

COMMON WRONG ANSWERS: Treating k-factor as a static vanity metric without connecting it to cycle time. Insisting that any k below 1.0 is a failure and ignoring the optimization path. Proposing generic referral programs that add friction instead of identifying the native viral moment inside the product. Forgetting that network effects also improve retention and engagement, which changes lifetime value and therefore the allowable customer acquisition cost.

LIKELY FOLLOW-UPS: How would you instrument the product to measure true viral cycle time versus just invite sends? If your k-factor is 0.5 today, what is your three-step roadmap to 0.8? How do you distinguish between viral growth and artificial incentive-driven growth that churns? How does the k-factor change as the network matures and the easy invites are exhausted?

ONE CONCRETE EXAMPLE: Suppose you are modeling a marketplace with a typical k-factor range of 0.4 to 1.0 and an achievable target of 0.7 or higher. You might assume an invitation rate of two invites per active user per month and a conversion rate of 30 percent, giving a k-factor of 0.6. Rather than killing the program, you model compressing the viral cycle time from fourteen days to four days by auto-copying referral links to clipboard and removing one click that was dropping conversion by 20 to 30 percent. You also model that your top 20 percent of power users will send ten or more invites. Combined with paid acquisition, this viral layer reduces blended customer acquisition cost and creates the compounding curve investors look for.

Source: thegrowthterminal.com