Last Minute Ramblings

Tonight I was reading more articles trumpeting a few company’s “amazing growth” and then saw a bunch of blowback responses “but what are your real numbers?”. My first thought was “yeah, what are your real numbers like dau?” but my second reaction was “it’s fine to say growth is great, but how do you know it is meaningful growth and why should we believe that?”

When you are an early or growing startup, everyone wants to know your numbers. Investors, press, potential employees, partners all ask and almost browbeat people into giving numbers to make a quick judgment. And lots of people try to use the numbers to make a soundbite-ready comparison to something else (“up to 2% of facebook in half the time!” or “they only have 5% engagement rate of twitter, it will never work”)

I hear frequent questions about DAU, MAU, DAU/MAU, 2-day retention, 30-day retention, k-factor, invite rate, open graph impressions, conversion to pay, ARPU, ARPDAU, ARPWAU, ARPWOW, DARPA (ok I made the last two up). Most of the charts I have seen presented look the same - up and to the right at a nice fast rate. But it is sometimes quite creative how the x-axis and y-axis are tailored to get there. Often when I ask why someone decided to show their numbers in a specific way, eg something like a 2-day active chart, the answer is often because some advisor or investor asked for that or said it looked the best for growth.

What I feel is missing in a lot of this is getting to the meaning behind the numbers. The most important question to me for any of the products i’ve worked on or am looking at from investment side is: “What do today’s metrics tell you about how you think more and more people will be using the product in the future?”

If we just compare everyone’s DAU we have no idea which companies have the same users coming back every day with little growth, which are massively churning with new people coming every day and just as many leaving, which are growing with a solid base of addicted users that grows more and more each day through organic use of the product (this is the awesome case).

It takes a lot more than a sound bite to understand what real habitual behavior is for a product and to identify which users have gotten that habit, which are on the way, which are on the way out, and how the various vectors of growth are getting users to that point. At Twitter, we identified a group of “core users” which we expected over 90% of them to keep using the product in similarly active ways month after month, and modeled our growth towards helping more users get there. It didn’t matter what overall numbers we reported because as long as we grew this group, we believed the whole pie would get a lot bigger (and ultimately it has).

When comparing everything by dau, mau, dau/mau, etc, we should recognize that only tells us a snapshot of size today. Hopefully going forward we can we all (investors, press, etc) ask questions to understand what are the right core patterns and growth of different products and start rating sustained and direct growth differently.

I would love to hear your ideas. Also if you are interested in working on growth, please ping me - josh at greylock.

Also should give a shoutout to John Lilly with whom I have this same conversation with after almost every meeting we share…