[leephillips]

In your graph captioned "Observed Desktop Adblocking Rate vs Predicted Rate", where does the curve for predicted rate come from? It's not explained in the text.

You do not show how much of the decline in ad revenue is caused by adoption of adblock, or even whether the amount of decline attributable to adblock is significant or measurable. You merely point out that revenue has declined and that adblock adoption has increased.

EDIT to add: You show no data on adblock adoption that covers the same time span as your graph of revenue. Your proxy for adbock use covering this timeframe, Google searches, shows a distinctly dissimilar pattern from the revenue graph for the same period. You claim a "correlation" but don't calculate one. You fail to demonstrate anything about causation.

[aw3c2]

It looks like a typical case of idiotic extrapolation to me. http://xkcd.com/605/

[seanblanchfield]

Fair points. We need to provide the source data and a full explanation. We're going to publish a detailed report in August with everything in it.

To answer one of your questions, the predicted desktop rate is based on an assumption that adblock adoption is driven by word of mouth (we have a basis for this assumption), and a compound growth rate is fitted to raw adblock rate data.

It's hard to combine the google trends data into the same model, because it's just a picture of search volume, not total adblock installations. I think it would be a fair to assume that the growth rate of new installations would match the growth rate of search volume, which means that the growth rate of 65% we're seeing this year is higher than it would have been in previous years.

Finally, I want to acknowledge that nothing can grow exponentially forever! At some point everyone who will use adblock will already have it installed. I'm pretty sure some countries are pretty close to adblock saturation already (Germany, Poland, Sweden and Finland).

[ehurrell]

Sounds reasonable, I look forward to seeing the full details in August to get the full picture, but those assumptions seem fair.