analysing YouTube’s Revenue, Part 2: Reader Input


Several readers weighed in with sharp comments on our YouTube revenue analysis.  In this post, we discuss these thoughts and incorporate some of them into updated estimates.  The overall message doesn’t change: YouTube’s revenue will likely be immaterial to Google’s overall performance for at least a year or two.  Five years from now, it could be very material, at least on the top line.

Liz Gannes of NewTeeVee says YouTube product manager Shashi Seth told her that the video ads would be sold at a $20 CPM whether or not users clicked on them (Liz’s post on the ads is here).  We assumed that the ads would be pay-per-click like the rest of Google’s advertising, so a flat impression-based CPM simplifies the analysis.  Liz’s input also reduces a concern that others had, which was that click-through rates would decline far below the rates we used (if the ads are impression based, CTR is irrelevant).

Another area of push-back was the percentage of YouTube videos that would ultimately have ads (dancing cats not being prime advertising inventory).  Specifically, Liz and others felt our range of 10%-50% was too high.  Here, we’re going to hold fast.  First, we think that as viewers and advertisers get more used to video advertising, concerns about advertising on or near inoffensive UGC will fade.  Second, and far more important, some videos account for a much higher percentage of overall streams than others, so even if, say, 70% of videos never have ads, we still think at least 50% of overall streams could (the long tail is long, but the curve isn’t flat).

Liz and others also noted that YouTube will have to pay a licensing fee in the form of a revenue share to most of its content partners.  Presumably the $20 CPM the company cited was gross, meaning that its net is likely to be far less than that, even when it gets the full rate. Based on this and other intelligent lower-CPM arguments, we’ve reduced our CPM assumptions from $10-$50 to $5-$30.


This input results in two changes to our model:

  1. Change to flat impression-based CPM from PPC. This increased the overall revenue.
  2. Reduction of CPMs. In most scenarios, this more than offset the above increase.

The adjustments, therefore, modestly reduce our revenue estimates.

Adjusted estimates here.
Original estimates here.
Original post here.