The Shape Of The Curve

And what it says about your product

A picture is worth a thousand words, and in business, so is the shape of a curve. In product roles — and now as an investor — I have found charts to not only be the best way to assess the performance of a product but also a great way to communicate a story, especially a phenomenon or consumer behavior that’s not well understood. The storytelling ability of charts is particularly interesting to me because the shape of a curve reveals a lot more nuance about the product than snapshot metrics ever can.

Everyone is familiar with the “up-and-to-the-right” growth curve but when trying to understand how consumers are really interacting with the product, here are some charts I like to look at.

DAU/MAU

The most commonly used engagement metric is the DAU (Daily Active Users) to MAU (Monthly Active Users) ratio. The higher the ratio, the more of a daily habit the product is for consumers, like wearing a watch or viewing TV. Most consumer software products have a DAU:MAU of 20% or lower. Zero products get to 100% but some rare ones come close to this ceiling (like Facebook 65% and Whatsapp 72%). For me, it’s not the flat number, but how the ratio behaves over time that is most revealing.

On a graph where the timescale is a year or more, here are three shapes to consider: 1) Up and then flat, 2) Stagnant, and 3) Steady decline.

1. Up and then flat: It’s obviously terrific to see the curve go up, but it often plateaus out at some max point and ends up looking like an S-curve.

Upon closer inspection (shorter period), the curve won’t seem as smooth as it first appears. The up-swing part of the curve is usually spiky. The bigger spikes usually represent product or feature launch that garner a lot of user attention, a portion of these will leave once the novelty factor wears off.

What’s interesting about these spikes is that they tend to have a higher landing point than the corresponding launching point, implying that the new feature has converted more monthly users to daily users, thus increasing the overall engagement of the product. That’s a successful product launch!

2. Stagnant: The stagnant DAU:MAU curve usually appears to be slowly increasing or slowly decreasing.

But the curve usually fluctuates within a stagnant band and generally finds it hard to break out of the upper band into an S-curve.

Implications: Users aren’t going to use the product more as-is, and the company has to innovate or pivot out of the stagnancy. Frequently the curve breaks through the lower band into declining usage territory, this is a red flag to watch out for because it implies that the company had product-market fit but may be losing it.

3. Steady decline: The declining curve also displays spikes upon closer inspection but the shape of the spikes look different from the S-shaped up-and-flat curve above.

In this case the landing point of the spikes is lower than the starting points of the curve.

Implications: Product launches aren’t having any effect. And if the spikes have particularly high crests, we can deduce that marketing efforts are failing too. At this point, it’s time for a major product pivot or prepare to be acquired.

Hourly DAU

The DAU curve usually looks pretty similar to the DAU:MAU curve — either up, flat, or down. For products with an upward or flat, looking curve, I like to zoom-in to the DAU curve that’s broken up Hourly.

The hourly chart usually looks like a sine curve and it’s a really good sign for business if it’s smooth.

If the shape of the curve is a sawtooth wave, then there’s probably some inorganic phenomenon going on — the most egregious of which are bursts of external stimuli (e.g. mobile push notifications) at various times during the day to get people to use the service. These bursts momentarily raise the peak DAU, and is generally hard to see in the standard DAU chart.

Tenure chart

Finally, if your core product has been around a few years and you’ve launched some major product initiatives over that time, it’s quite enlightening to look at a tenure chart. Tenure is essentially a reverse growth chart displaying the number of active users (y axis) the product has had, starting with the most recent users (least tenured) to the oldest users (most tenured).

Let’s walk through the above chart from right-to-left:

1. The line ends at 6-year mark — that is when the company started.
2. The product hit a peak at the 5-year mark. This is the initial product launch (a successful launch, that is).
3. The number of users declines after that indicating, a) there were no subsequent major product innovations, and b) the product hit a market cap. Basically there are not many potential new-users out there remaining to fuel product growth. One example would be a country-specific product having a natural ceiling of the given population of that country.
4. At the 2-year-and-less mark we see an increase. This indicates entry into new markets or some major product overhaul (e.g. the company’s product launching on mobile or a new platform).
5. The curve will appear to be artificially high at 100% when it hits the y-axis at the “0' year mark. Since the number of users that joined this month are, by definition, active this month.

There are plenty more metrics we can extend this form of analysis to — not only to consumer products but also to enterprise products. The shape of each curve tell a fuller story that is hard to unearth from a static set of metrics alone — you just need to know what you are looking for.

I co-wrote this post with my friend Sagar Savla, along with lots of help on chart design from Ilwon Yoon.