Referrals and viral growth

Referrals are a win in any business as they bring the cost of acquisition down to zero (not taking into account the cost to acquire the referrer). In businesses where scale is required, referrals needs to be tracked and optimised as it is not cost effective to pay to acquire the millions of users required to support the business model.

To speed up referrals (and product sharing) this has to be part of using the product itself. An example of this is when you send an email and at the bottom it says “Sent with X” or “Sent on device Y”. Just by sending emails the product is being shared with others.

Metrics to track for this are:

1. Payload: How many people are included in one share
2. Frequency: How many times shares are being sent
3. Conversion: How many people come to your product/service as a result of the share

To track the growth of a product overall, you need to track:

1. Viral coefficient (k) - the number of new users each current user brings on board
2. Cycle time (p) - the time it takes between a new user signing up, and their referrals signing up

If the viral coefficient is < 1 then bringing in 100 new users this month (for example) will bring in less than 100 next month (assuming no extra marketing) and this will lead to a gradual decline to zero - meaning a constant influx of marketing activity is needed.

If k is equal to 1, then 100 new users will bring in another 100 new users and growth will be linear (y = k x).

If k is greater than 1 then this is when growth starts to get exponential (y = k ^ x). 100 users in month 1, will lead to 120 in month 2, and 144 in month 3 etc.

A high k value is needed, but so is a small p value. If your cycle time is too high it will take a long time to get those new users in.

NB: Both p and k are lagging metrics, you cannot alter these directly, only alter your product/service and check these numbers.