Cintreuk

DAUs MAUs and LIEs

DAUs MAUs and LIEs

Don’t you get surprised when you see numbers like “Heathrow had 69+million passengers during 2011”? You then consider that the United Kingdom has about 62 million inhabitants, and you imagine people there travel a lot! Things get even more extreme if you consider that Heathrow is not the only airport in the UK. Where are all these people going? Then you start the “data mining” process, and you realize that most people going through airports are not migrants. Consequently, most of them then will come and go back somewhere (or go and come back). Most times, the same passenger is counted twice, and you can then quickly realize that those 69+ million passengers are at most 35 million people (I’ll round up). Yet, that is still a very large number, if one considers those to be different people. But are they? The average number of trips abroad is affected, for example, by businessmen. Some of them will travel abroad at least once a week (some even more!).

Between those that travel once a year and those that travel more than once a week, there must be some distribution. My guess would be as good as any other for such distribution. I started with 40 people out of a 100 traveling abroad once a year. I then decrease that number in exponential moves, while trying to keep the total of 100. And I didn’t want any number to have a percentage of the total of less than 0.25%, because after a while, we really are likely to have the same probability of having people traveling 22 or 27 times per year (or so I believe!). My resulting curve has points (number of trips, %): (1,40), (2,20), (3,10), (4,5), (5,4), (6,3), (7,2), (8 to 10, 1), (11 to 20, 0.5), (21 to 52, 0.25). This curve has a weighted average of about 5.785 for the number of trips (let’s round up to 6!). Consequently, if you assume such distribution, the average number of trips abroad would make the number of passengers be divided by 6 (and don’t forget the “round-trip” effect!). From a start of 69+ million passengers, we end-up with a guess of about 6 million people travelling through the same airport.

When seeing social networks claiming to have 900 million monthly active users, the only thing I can think of is: how are they counting? Am I considered a user, despite having a registration mostly to prevent others from registering my name? How many would be really active users? Obviously, those that play games are likely active, since social games require people to continuously monitor their “farms” and other similar virtual items. The number of daily active users for such games is about 20 million. How do I know? I’ve once – within the last year – attended a “social games” conference just to peek at such MAU (monthly active users) and DAU (daily active users) numbers for different companies, and the maximum for DAU was around 20 million (and somehow nobody cares about the weekly active users!). Obviously, the MAU > DAU, but you cannot make a business based on users that log in once a month. Engaged daily users are the ones that can be sold to advertisers as targets for ads. You cannot have a campaign targeting monthly active users if it will last a week, target the Valentine’s day, and such users aren’t guaranteed to login until after the campaign is over. Now you know why social networks are desperately sending you e-mail because you are “missing activity” by not login in daily!

What would be the distribution of “active users” for social networks? My reasonable guest is that among all registered users, half are simply frauds, or experiments, or investigations. I myself registered a few times in social networks just to test if I could see data from my real account given different privacy settings. Don’t even start me on the “lonely Russian girls” trying to befriend me… I’m likely being very conservative here thinking that only half of the registered users are frauds. Among the remaining “real people”, how many are “really engaged”? If I was to consider my data: 1/7. So, if you start with a social network claiming to have 900 million registered users, take half away due to fraud, and keep just 1/7 of those as “engaged”, you end-up with 64+ million active social network users. Those people keep posting, liking, tweeting, retweeting and commenting among themselves, giving everyone the impression that there is a lot of market for ads in social networks. My suspicion: even that is a lie. Many users are logged in from devices that are “always on”, counting daily activity as the pull of data from servers. The MAUs and DAUs lie may continue for a while. It won’t be able to continue forever…

Available link for download