DAU/MAU Ratio

How much a product was being used got reported as a single headcount for a long time. Monthly active users (MAU) — the number of people who used the product in a month — going from 8,400 to 9,100 was accepted as evidence of growth.

That number leaves out intensity. Someone who opened one document once in a month and someone who opens the product every morning to work both count as one. So two products with the same 9,100 can be entirely different things: one is wedged into somebody's daily work, the other comes to mind once a month. A month with an event or a mention somewhere inflates the figure especially hard, and the following month gives it all back.

The DAU/MAU ratio exists to expose that difference as a single proportion. Usage metrics split into a branch that counts people and a branch that counts what those people did, and this value is a ratio built from two numbers in the first branch. Next to the figures that describe size, it holds the seat that describes depth. Neither question answers the other, which is why both stay on the page.

Divide the number of people who used the product on a day by the number who used it in the month. Four things have to be settled.

1. Pin down what counts as active before anything else. Does signing in count, or does it take a core action like opening or creating a document? Counting sign-ins usually returns a higher number, because it includes people who left a tab open and did nothing. Whichever rule gets chosen, changing it later makes every earlier month uncomparable.
2. Average the daily figure. Using daily active users (DAU) from one particular day means reading the day of the week. Business products routinely fall to a third of weekday levels on a Saturday, so take the mean across the month.
3. Divide two figures from the same month. A daily average of 900 against 8,400 monthly actives is 11 percent — roughly one in nine of the month's users shows up daily.
4. Count people, not visits. Someone who signs in five times in a day is still 1. Mix this up with sessions, which count visits, and the ratio can exceed 100 percent.

Products nobody has a reason to open daily sometimes use a week instead of a day. Weekly active users (WAU) divided by monthly actives reads more easily for a business tool whose users work in it three days a week. Either way, write the numerator and denominator down and keep counting the same way, since the ratio only carries meaning against its own history.

What counts as good is set by what kind of product it is. Messaging and social products, where people exchange things daily, land between 40 and 60 percent; a document tool teams share is healthy between 15 and 25 percent. A product for running payroll or filing taxes, used once a month by design, is fine at 5 percent. Demanding a document tool's 20 percent from a product like that guarantees nobody reaches it.

That makes comparison against other companies nearly useless. Comparison needs identical counting, and the definition of active differs at every company. This value gets read along a single product's own timeline. Move from 9 percent last quarter to 12 percent this quarter and that three-point gap is something you can attach a reason to.

Being a ratio brings one trap with it: it also rises when the denominator falls. Monthly actives sliding from 8,400 to 5,600 while the daily average holds at 900 moves 11 percent to 16 percent. On the chart, intensity improved. What actually happened is that the occasional visitors left. This is why the value never gets read apart from the two numbers it came from.

Raising it has a known location too. It means building a reason to open the product daily, which in practice means finding the place where the product can sit inside someone's working routine. Push notifications to drag people in instead and the value climbs for a month or two, after which more people mute the alerts or leave. That's why retention has to sit next to this number whenever it becomes a target.

"We're at 11 percent. That means people come back about every tenth day."

Translating the ratio into an interval between visits.

"That went up because monthly actives dropped, not because daily users grew."

The denominator shrank, so the rise isn't a good sign.

"Nobody needs our product daily, so we read it on a weekly basis."

The numerator moved from a day to a week to match how the product is used.

Pagedeck reads this value in two pieces. Overall it sits at 13 percent, but among users on paid teams it's 31 percent and among free solo users it's 6 percent. Two different kinds of use are living inside one product, and the combined figure describes neither of them.

After Mia brought that split to a review, the overall figure came off the goal list. The easiest way to raise the combined number would have been to bring in fewer solo users. The paid-team figure gets measured quarterly instead, and the solo side is watched through activation, the point where a new user first gets something out of the product.

The most common mistake is reading it as a number that should always go up. It measures intensity, so higher looks like better. But the right level depends on the rhythm the product is naturally used on. Wanting a product that replaces a monthly task to be opened daily is asking users to invent work they didn't have. Set that as a goal and the team drifts toward whatever makes people open the app — notifications and badges, usually — while the thing the product was good at slides down the list.

Treating it as the same thing as retention is the other one. Both deal with continued use, so they blur together. What they look at differs, though. Retention takes everyone who arrived in the same period as one group and follows how many remain as time passes. This ratio photographs everyone present this month and reads the share of them who come daily. So in a month with many new arrivals, the ratio can fall while retention holds steady.