The business model is solid. The CPM's are high. The sites are great quality. What does the revenue look like? It is hard to say for sure how much topline revenue these vertical ad networks really drive, but a deep dive into some of the more established networks shows some good signs. So to find what the opportunity might be we needed to use some public data and triangulate around total value. The goal of this exercise was two fold, first to try and better understand how big these networks actually were; but secondly, how well our growth rate predictions fit the vertical ad network model.
To get a good sense of how vertical ad network revenues grow the best proxy is looking at the actual sales from these vertical ad networks. We built a model that includes the growth trajectory for some of the top vertical ad networks in the market today by using public data from the top audience measurement firms.
To build this analysis I took the traffic and total active site count data for Glam, FM Publishing, and JumpStart networks, these networks have much more publicly available data on the composition of their network, than others. The next step was to run reports in Media Metrix, Compete, Quantcast, and a few others on the total impressions served for each network, and contributing site. I triangulated the 3-4 data sources to come up with average impressions count and total unduplicated UV’s for each network. Of course if possible I relied more on comScore or NetRatings vs the others.
Now, the problem with these numbers is really driven off the fact that impressions alone don’t really tell you much more than how big the potential network is. The page views or impressions metrics within all the audience measurement services give an estimate of scale, but don’t really tell you anything about sell-through rates or CPM’s. There are plenty of rumors going around that the sell-through rates on Glam and FM Publishing are low, that is probably true if you measure sell-through as a percentage of all inventory. But if you calculate the sell-through against the high quality (above the fold, homepage, premium content) channels/pages, then the sell-through on these networks is much higher.
So, to make this more complicated, we dropped the impression growth rates into a time aligned curve built off of over 100 networks from the beginning of time (okay, since 2000). This historical growth rate then drove the sell-through rate across these networks; we were making the assumption at this point that these networks will sell their inventory at the rate of the historical trends across the other networks. We already have their site and reach growth rates, so all we needed was the total impressions on top of the estimate of sell-through.
Using this model, the actual impressions on the network is a proxy for the available inventory to sell. In the case of FM Publishing and JumpStart, this actually is an accurate assumption because of the way their networks actually operate. Glam is a slightly different animal and there is a potential within the Glam estimates that their total revenues in 2008 and 2009 might be slightly over stated. The only reason we don’t think this is a big issue is because Glam is still growing their network, which we are not taking into account in the network, and Glam also has a very good remnant deal with Google, which would then support the higher sell-through across their networks, albeit at a slightly lower eCPM.
Finally, to align and calculate the actual revenue, we used the average CPM’s that our market intelligence tells us is the appropriate rate for the categories these networks are selling into. The CPM estimate was built based on taking the category CPM’s and then looking at the composition of these networks by the categories they are representing. Additionally, we have the actual reported/public CPM’s from each of these networks and can make some assumptions on discounting based on rate-card and volume estimates.
The final outcome of this analysis shows a pretty solid growth rate across these three vertical ad networks. The estimated revenues for 2007 were: Glam $21MM; JumpStart $59MM; FM Publishing $24MM. What is the over under on these numbers? Who knows, but the data seems to be in the range of acceptable error.
The growth rates across these networks ranged from 100% growth to 700% growth. This growth is not just a function of the model, but really a function of the size of the network and where the network actually is in the model. Because Glam continues to recruit sites on a monthly basis, their sell-through rate is actually lower because the composition of their network is expanding as they sell more. JumpStart is actually counter to this as they have a limited set of sites they are working with, at least until they start to recruit more sites, and they are somewhat bound to a doubling of revenue year over year.
These numbers seem very reasonable and if you compare to publicly available data the numbers are within the range of error.
We believe these networks, in their 3-4th year of operations are going to enjoy significant growth in total revenue, but also higher CPMs and bigger budgets within their category.
Remember these are estimates and the public data that was used to validate this data is probably subject at best. Please use this analysis with a grain of salt.
