How do you know if a social app will be a "hit?"

Predicting what goes mainstream - and what stalls out - might seem random.

But after years of leading growth @Snap & investing in consumer social at @a16z, I've found early signs of huge products ๐Ÿ‘‡
When analyzing early stage social apps, there's three categories of metrics I care about:

- Growth
- Engagement
- Retention

I'll outline how to measure each one, plus our @a16z benchmarks based on hundreds of apps we've seen.
First - growth.

Growth is often spiky at the early stages. It's driven by unpredictable moments of notoriety, like @djkhaled's famous jet ski video on Snapchat!

But when you look backwards, your user growth should trend "up and to the right".
Ideally, growth is 90%+ organic. If your app isn't inherently viral, you may not have P/M fit.

For seed companies, here's how we benchmark average MoM growth:

- OK = 20%
- Good = 35%
- Great = 50%

In social, growth is necessary but not sufficient. What matters more?
Engagement! For social apps, you want users engaged daily.

The easiest way to measure this is daily / monthly active users, or DAU / MAU. Here's how we benchmark this:

- OK = 25%
- Good = 40%
- Great = 50%+

Consumer social "giants" have a DAU/MAU around or above 50%.
Another metric is the L-ness curve, or # of days users are active in a given timeframe.

A weekly L-ness curve is % of users who are active 1 day/wk, 2 days, 3 days, etc.

We often look at L5+, or % users active 5-7 days/wk. Benchmarks:

- OK = 30%
- Good = 40%
- Great = 50%+
Our last core metric is retention - the lifeblood of an app.

We measure N-day retention, the strictest definition.

For example, day 30 retention = (# of users in a cohort who are active ON day 30) / (# of total users in the cohort)
Here's how we benchmark N-day retention:

- OK = 50% D1, 35% D7, 20% D30
- Good = 60% D1, 40% D7, 25% D30
- Great = 70% D1, 50% D7, 30% D30

We're also looking to see retention cohorts that are consistent (or improve) over time - and that "flatten out" before D30.
Retention is the hardest metric to move, and it's almost impossible to "game."

At Snap, I saw that improving D30 retention even by 1% at scale was a Herculean task.

Retention can only be "solved" by making the product better, not by marketing.
Check out my full post for more on what we look at for early stage social apps.

If you're working on something with great early metrics, my DMs are open!

And if you're not there quite yet, don't worry - social often take a few tries to hit a winner.
None of the above should be taken as investment advice; please see for more information.
See All