Debunking StumbleUpon

StumbleUpon is an extremely powerful social site. It is the top referrer of social sharing, even exceeding sites like Facebook. It has an estimated 12 million unique visitors per month and, as seen in the infographic below, is very influental.

StumbleUpon makes up quite a big chunk of my traffic. From May to February it sent about 880,000 visitors my way. It is also completely useless at driving valuable traffic.

The problem with StumbleUpon

I want to show you something interesting. The graph below shows the total amount of visitors coming to this site, in comparison with advertising revenue, from May 2011 to February 2012. As you can see, it has been a roller coaster ride with rather dramatic changes from one month to the next.

For a long time I didn't understand why my traffic looked like this. I thought perhaps I was doing something wrong? Maybe some of my articles were turning people away?

But then I decided to compare it with the number of new subscribers to Baekdal Plus - and the result was a bit of a shock. There is no correlation between the two. Volume of traffic seems to have no impact on why or even when people subscribe.

Note: As a publisher, this is a very important lesson. Advertising and subscriptions don't follow the same pattern.

So I decide to investigate, and as I digged into the stats, StumbleUpon immediately stood out from the rest.

Here is a graph illustrating total number of unique visitors with the number of unique referrers from StumbleUpon. Those roller coaster like curves are caused exclusively by StumbleUpon. The green line illustrate the total unique visitors if we exclude StumbleUpon from the graph. It is quite a difference!

The next thing I looked at was the conversion rate of StumbleUpon - and, as far as I can tell, it's practically zero.

In January, the top referrers to my plus content were: Direct (which also includes referrers using HTTPS), Twitter, Facebook, Google+, Google Search, LinkedIn, and my email newsletter (in that order). StumbleUpon came in on a 16th place with only 8 people in total (out of 138,000 visitors).

That is a Plus referrer rate of only 0.0058%. In comparison, Twitter's Plus referrer rate is 8%, and Facebook's is 9%.

StumbleUpon is just noise. It is really fun to have 138,000 people coming to your site, but what's the point if only 0.0058% of them are valuable?

Another element is advertising revenue. As you can see from the graph above, those big traffic months from StumbleUpon also result in an increase in advertising revenue, but one thing you cannot see is the click-through rates.

When I compared click-through rates in June (low-traffic) with September (high-traffic) I found that every month StumbleUpon referred a lot of traffic click-through rates dropped to less than half of the "slow months".

So not only is StumbleUpon hopeless at bringing people to Baekdal Plus, its traffic is also hopeless at driving value for advertisers.

The reason why StumbleUpon is so bad at creating value is because of how it works. If you haven't used it yourself, here is the short version:

StumbleUpon is essentially the modern version of TV channel surfing - like when you sit down in front of your TV and you just flip through all the channels in search of something interesting. That is what people do on StumbleUpon.

When you go to StumbleUpon, you are presented with a random website. If you don't like it, you click "stumble" to move on to the next ...and the next ...and the next ...and the next ...and the next ...until you eventually find something you like.

The result is that every click is recorded as a "real" visitor, but most of that traffic is just people clicking on the stumble button several of hundred times. They are channel surfing, and your site is just one of thousands of channels they flip through.

The value of this interaction is practically zero. Sure, if you are monetized by CPM based advertising you might make a little bit of extra money, but you do so at the cost of a much lower click-through rate for your advertiser. That means you will not be able to attract premium advertising partners.

StumbleUpon is not a good site. It's persuasive in its sharing power, but the traffic it creates is not the type of traffic you want.

The stats for your site might look different, but I'm seeing the same pattern over at 42Concepts. While StumbleUpon creates a lot of exposure, advertising click-through rates are appallingly low.

I'm slowly coming to the conclusion that maybe I need to remove StumbleUpon for my analytics. It's distorting the image, and causing me to make the wrong assumptions about my traffic as a whole. I won't block it. But maybe I need to just ...ignore it.

Filter it out by default.

Update: Several people asked about the effect StumbleUpon has in boosting search engine traffic and ranking. As far as I can tell, there isn't any. Read more in "StumbleUpon vs. Google Juice".

More from Baekdal