Using Website Unique Visitors To Estimate PC Digital Distribution Market Share

We’ve used the Easter break not only to collect eggs and visit our beloved. No, we actually worked too:)

As with our Top 20 Download Charts here is another attempt to better understand and quantify the games download market. So far there have been some vague numbers and discussions with regard to how the different download shops are doing.

The only hard facts I could find date back to November 09. In its customer report Stardock (operator of the Impulse online store) estimated that Steam operates as some sort of monopoly with 70% market share of the digital distribution business. Impulse follows with 10% while all others fight for the remaining 20%.

I found that quite intriguing. Not only that Steam has built such a strong market position but that there would be so few competition. So we decided to do some research on our own.

Method and Limitations

Since we aggregate deals from all major digital distribution platform we somewhat know the market and its main players. Since real sales data from are (yet) hard to come by we needed a different metric. We went for unique visitor as the core metric, asuming that the more people visit a store the more people buy at that store.

This obviously is not perfect since we don’t know how well the indivdual shops convert visitors to actual sales. Another limitation may be that some platforms (like Steam or Impulse) don’t require their users to go to the website to purchase. Users can also use the desktop client to access the shop.

Further, all traffic data itself are estimated and therefore present another imprecision. For the final analysis we used www.compete.com, which seems to be a bit US focused and thereby propably bringing in further imbalance for non-US stores. I compared all results against analysis with Google’s Adplanner and Trafficestimate. While there are some differences (between 2-5%) the overall picture doesn’t change.Finally, I estimated all shops that are not on the list with a conservative guess of 10%.

Our Findings

ShopMarketShares_April2010

These results obviously do not fully reflect the few currently available data (see Stardock). As discussed earlier, all data are based on estimations and a number of assumptions so they are prone to error. However, we see them more as a starting point for further analysis and hopefully a critical discussion.

Also, I’d very much like to hear your own interspretation of this market structure.

[Talk to me at blog.deals4downloads.com or visit www.deals4downloads.com.]