amy@promatt
CG Top Poster Club
- Messages
- 20,561
- Joined
- Jul 16, 2009
- Messages
- 20,561
- Reaction score
- 571
- Points
- 166
Google Engineer Explains Why Your AdSense Earnings May Drop
http://www.youtube.com/watch?v=O1BaOMqcyQY
But he also adds a lot of detail about how the AdSense pricing algorithms 'think.'
From a technical point of view, the question is ill defined. Google does not "consider traffic from social media sites" to be anything. Instead, quality is an intrinsic property of the traffic:
1) Invalid clicks. Google examines individual clicks and impressions, and tries to figure out if they are valid or not. A source of traffic is not low or high quality because Google considers it is, but the reverse: Google's algorithms try to reflect the quality of the traffic by detecting invalid clicks accurately.
2) Conversions and smart pricing. Are users genuinely interested in the ads they click? Are they generating conversions to advertisers? Again, Google does not decide in advance that a kind of traffic is high-value. Instead, Google measures the conversion rate, and applies Smart Pricing accordingly.
There might be patterns emerging from this, for instance related to social media sites. As an engineer, I have no idea. Personally, I believe that there is no such general rule, all depends on the specific case of each publisher.
I think that trying to correlate general problems and disconnected symptoms with rumors or the experience of a few other publishers who are probably very different, then extrapolating what could have happened, is not a successful approach in general. Instead, publishers should try to understand their own traffic in isolation. Measuring, gathering facts, looking carefully at reports (like I did with all accounts I looked at so far) works better, because it does not rely on statistics. Most of the time, the questions of the publisher can be answered by a report showing a peak or an anomaly. (Then the problem is often that publishers think or hope that their traffic is high quality, and refuse to admit that it may not be the case all the time, and that the problem is not Google's fault but their's.)
http://www.youtube.com/watch?v=O1BaOMqcyQY
But he also adds a lot of detail about how the AdSense pricing algorithms 'think.'
From a technical point of view, the question is ill defined. Google does not "consider traffic from social media sites" to be anything. Instead, quality is an intrinsic property of the traffic:
1) Invalid clicks. Google examines individual clicks and impressions, and tries to figure out if they are valid or not. A source of traffic is not low or high quality because Google considers it is, but the reverse: Google's algorithms try to reflect the quality of the traffic by detecting invalid clicks accurately.
2) Conversions and smart pricing. Are users genuinely interested in the ads they click? Are they generating conversions to advertisers? Again, Google does not decide in advance that a kind of traffic is high-value. Instead, Google measures the conversion rate, and applies Smart Pricing accordingly.
There might be patterns emerging from this, for instance related to social media sites. As an engineer, I have no idea. Personally, I believe that there is no such general rule, all depends on the specific case of each publisher.
I think that trying to correlate general problems and disconnected symptoms with rumors or the experience of a few other publishers who are probably very different, then extrapolating what could have happened, is not a successful approach in general. Instead, publishers should try to understand their own traffic in isolation. Measuring, gathering facts, looking carefully at reports (like I did with all accounts I looked at so far) works better, because it does not rely on statistics. Most of the time, the questions of the publisher can be answered by a report showing a peak or an anomaly. (Then the problem is often that publishers think or hope that their traffic is high quality, and refuse to admit that it may not be the case all the time, and that the problem is not Google's fault but their's.)
