mDAU, DAU and the Growth Hacks for High-Growth Companies¶
Facebook first popularized MAU (Monthly Active User / People) in 2010 when it surpassed MySpace became the social network on the web. Unlike traditional metrics such as total registered user or daily new user, MAU accounts for both growth and retention. Ideal for measuring growth of the online community, it is adopted pretty widely in a short time frame.
MAU metrics is not without faults. Facebook traditionally under-reports its MAU numbers (their milestone announcements such as 100M MAU and 1B MAU both comes a few days, even weeks after stabilized around that number). It can also be affected by seasonality. Holiday seasons and New Year can see positive changes to MAU that are temporary.
While MAU is a reasonable metrics for many investors to look at. New generation of social media companies such as Snap are advertising DAU (Daily Active User) metrics. Mark Zuckerberg, was reportedly looking more at DAU and WAU (Weekly Active User) after the company's shift to mobile in 2012. On mobile, if a user only opens the app once every month, it is not a meaningful interaction and not so much monetizable at all.
The recent changes for Twitter's quarterly reporting to mDAU (monetizable Daily Active User) is an interesting move. This cuts to the core of how different companies report DAUs and why an accurate reporting of DAU is difficult. Thus, prompts me write this article to give a quick primer on DAUs.
Monetizable in DAU?¶
There are several things interesting in Twitter's announcement of this change. Many other social media companies don't put the emphasis on monetizable because it is implied. It simply doesn't make sense for Instagram having DAU that is not monetizable. Users can only interact with Instagram on the Instagram app, and Instagram sells ads globally. The only exception, for Facebook / Instagram probably comes from less than a million people in the holdout group that have never seen ads from the very beginning. Facebook holds this tiny fraction of people to continuously gauge how much advertising load will start to hurt its user engagement.
It is certainly true in 2012. At that time, Facebook's newsfeed didn't have any ads and its mobile apps only support newsfeed, not the ad units. You probably could, in theory, exclude these people from mDAU metrics as Twitter defines it.
For mDAU in Twitter's case, they specifically called out "people who saw embedded Tweets on other websites" or "viewing Twitter content without login" are not mDAUs. It can be more than that sometimes.
It is no secret that social media companies track every interaction you have with their websites and apps. What constitutes as active sometimes can be a contentious topic internally. For example, if every time you receive a push notification counted towards an action, the DAU number can be inflated dramatically (Twitter's DAU to MAU ratio is around 39%, push notification enable rate is around 50% to 90% depending on apps).
Facebook defines the action as logged in and visited the site, that include viewing, commenting, liking or posting contents. Other social media apps could use tap to open as the indication of active. However, this can be tricky. A program bug or crash could artificially inflate tap to open metrics, causing the counterintuitive effect to inflate DAU metrics.
From When to When is a Day?¶
DAU number is susceptible to cut off times. If a social media's peak traffic time is around 1900 in Eastern Standard Time (United States) and if their cut off time is around the peak traffic time, a user who uses the social media app every the other day can be counted as a whole DAU everyday because their interactions are around the cut off time, therefore, counted towards both the previous day and the day after. Depending on the cut off times, DAU metrics can be inflated intentionally. Many social media apps use its low traffic time as the cut off time, which could change over time as their international presence expands.
Ramadan holiday is famous inside social media company circles because it shifts the traffic pattern considerably in Middle-East countries. Their peak traffic will be pushed a few hours later to after the sunset. For some social media companies whose cut off time is global, this may deflate their Middle-East metrics a bit.
When to Conclude the Day?¶
We cannot manufacture another hour in a 24-hour day. However, the data pipeline, especially the ones with mobile client involved can be delayed. If a user opens the app at 11:55 p.m. with limited network connectivity, should we wait until that data uploaded to the backend or should we cut loose? If we are going to wait, for how long then? Some companies take 6-hour cut off (i.e. wait until 6 a.m. next day to run the batch job to compute DAU). Some companies wait a whole day. Sometimes it depends on at which point, the extra wait time doesn't generate meaningfully more DAUs; sometimes it depends on the data retention policy in your Term of Services.
What is the DAU Reported Quarterly?¶
The above analysis ignores the glaring fact that fake accounts and bots exist. Sometimes, these will be counted towards DAU by mistake or simply the inability to identify and filter out.
Because DAU is a daily metrics, the reports in many quarterly earnings from social media companies are actually less rigorous than many thought. Facebook reports the DAU as the averaged DAU everyday over last month of the quarter. Since Facebook traditionally under-reports its engagement numbers, they also subtract unspecified amount to make the number looks more stable. Twitter and Snap report the DAU as the averaged DAU everyday over the whole quarter with unspecified rounding mechanism. All in attempts to smooth the DAU number to be stable quarter-over-quarter while quantitatively meaningful year-over-year. Facebook's calculation are more current while Twitter / Snap's numbers are laggy.
What That Means to Investors?¶
mDAU / DAU could be a reasonable number to glance over every morning as a routine for management. But it is only one of the many metrics (W6/7, MAU, Core Interaction etc.) competent management team tracks to guide decision-making. For serious investors, the resolution of quarterly reported DAU is simply too crude to base any conclusions alone. Do user studies, gather engagement patterns, understand demographics and geographics, the peak traffic patterns and then make sense of these DAU metrics. The monetization potential lies within the engagement pattern, not the absolute number of DAU.