Last November, David Karp posted on Tumblr's staff blog. "I have an exciting and slightly overwhelming milestone to share with you," the site's founder and CEO said. "Tumblr has broken into the Top 10 sites in the U.S., and now has a worldwide audience of 170 million people." The post contained this chart, from Quantcast, the analytics company he cited for his "Top 10" claim, which Tumblr allows to "Directly" measure its traffic:
What we didn't know at the time was that Tumblr was likely already courting possible buyers; a few months later, it would be purchased by Yahoo for over a billion dollars. Why? For growth. "Growth in the consumer tech industry always starts with traffic," Yahoo CEO Marissa Mayer told the WSJ in May. Tumblr's users are young and savvy, yes, but more importantly they're numerous.
So this Quantcast chart of Global monthly uniques, from today, is interesting:
The traffic peak is at November of last year. Then there's this chart, showing weekly total people since May.
That peak is on May 20, the day the Tumblr acquisition was announced.
Despite Tumblr touting Quantcast numbers in the past, and the presence of a "Direct Measurement" badge, which implies that traffic is recorded by Quantcast with the support of Tumblr, these numbers shouldn't be taken at face value. Quantcast says it can't comment on individual customers, but Tumblr, in response to questions about the stats, sent this response:
Quantcast does not provide a full picture of our data, as it doesn't include our mobile app traffic, time spent on the site or all of the pages on the blog network because of ajax page loading.
This is potentially major: Mobile app traffic is a major source of growth for many large web companies. The second chart in this post breaks out mobile web traffic — not delivered through an app — shows that it's significant.
But: Asked if including mobile app statistics would paint a picture of overall growth, Tumblr did not respond.
In her announcement post, Mayer mentioned that, "[O]n mobile, more than half of Tumblr's users are using the mobile app." But the question here is not one of mobile growth, it's about overall growth. Facebook grew its mobile user base by 51% since last year, but only increased its total user base by 21%. Even more telling: Last quarter, Facebook increased its monthly active mobile users by 68 million. The site as a whole, including mobile users, only grew by about 50 million. All this is to say that while adding mobile app user data to the equation might change the apparent trends, it wouldn't guarantee overall growth. And Tumblr is functionally different from Facebook in that a lot of these monthly "uniques" don't really realize they're on Tumblr; they might be visiting a Tumblr blog, but they're not using the dashboard. Tumblr's mobile users are almost exclusively dashboard users, which means they have accounts, which means that they probably access the site from a computer every once in a while too.
So, where else can we look? Quantcast competitor Comscore shows Tumblr's multi-platform growth as basically flat over the last three months. Stepping away a degree, data from the BuzzFeed network, a collection of sites with hundreds of millions of users, shows a fairly serious decline in Tumblr referrals:
Mobile app growth does not require desktop and mobile internet shrinkage — a briskly growing property like Tumblr could, and perhaps should, see growth in both app usage and desktop usage.
But this data, and Quantcast's, suggests that Tumblr's presence on the internet, or at least a major part of it, isn't growing, but shrinking. There's a big difference between buying users and buying growth, and Yahoo may have just done the former.