StackOverflow revenue estimated

The question of whether a StackExchange-based site can generate a lot of revenue and my answer to it lead to a heated discussion on the OnStartups Answers. So I decided to estimate what kind of revenue StackOverflow actually generates. Apparently, it’s enough to try to go to VCs, but not enough for a self-sustained growth. Luckily, it is relatively easy to estimate.

Disclaimer: the following calculations are unproven, non-scientific, and cannot be used as a financial advice.

Despite a good and long list by Alex Lam of potential revenue possibilities for StackOverflow, they make money only via paid job posts and site ads (rectangle banner in the right column and sponsored tags).

1. Paid job listing is very easy to calculate. They have about 8-9 new job posts a day. One costs cool $350. So let’s estimate it at 9 * 350 * 30 = $94,500 per month.

2. Ad revenue is hard to estimate. Let’s start with the 26M pageviews per month they claim to have. A lot of these pageviews don’t carry any ads at all: pages like About, FAQ, list of users, etc.

2.1 I would take 80% are ad views and 20% non-ad views. This gives 20,800,000 pageviews with ads.

2.2 The ad inventory is not sold out. Every now and then they display an internal ad – ad for job.stackoverflow.com. To make matters worse, the ads are geo-targeted, showing different ads for different countries. This means that ad inventory and its sold out ratio is different for every country. My tests showed it is about 75% for USA and 45% for India and Russia. Let’s take these figures and assume 75% rate for highly ad-targeted countries and 45% for the other countries.

2.3 My estimation is that half of pageviews go from highly ad-targeted countries and another half from the other countries. Originally I thought that low ad-target countries would make most of the pageviews. However, I browsed through the list of registered users and was impressed with the big number of USA, UK, Australia, Canada users. So let’s take 50% – 50% split, which gives average of 60% sold out inventory ratio. Thus, only 12,480,000 pageviews carry paid ads.

2.4 Now the trickiest part, CPM – the cost per 1000 impressions. fabien7474 came up with roughly $5 CPM for StackOverflow and I tend to agree, not going into details.

Drum rolls: the estimated banner ad revenue is $62,400 per month.

2.5 There are also sponsored tags that display ads as well. I don’t think those are sold on CPM basis, and hardly a lot of people click those tags anyway. Also, consider that few tags are sponsored so far. A week ago I saw a bunch of tags sponsored by Adobe and them are gone. Let’s add ballpark $10,000 for the tags.

The total revenue is $94,500 (for job posts) + $72,400 (for other ads) = $166,900 or $167K per month.

A few observations:

  • The job listing brings more money than the ads!
  • I can try and estimate the costs (hosting, traffic, staff), but no matter what chairs you buy for your guys, the costs will be lower than $167K monthly. So, this is definitely cash-positive.
  • This revenue figure is nice and it is a good achievement, but looking for VC money to scale this business? You gotta be kidding.

In the next post I will go through competition and examples of similar sites to get the whole picture…

About these ads

3 Responses to StackOverflow revenue estimated

  1. oweibor says:

    Good analysis…

  2. Now they have stackexchange network.. They are making money in millions.

  3. howdythemes says:

    They have stackexchange network now. So I guess they are making money in millions now.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

%d bloggers like this: