Long tail, droopy tail, hump tail

Nielsen dabbles in $/page analysis

[ronburk]

Hopefully this isn't a dupe, but I only now noticed the Nielsen article on logarithmic analysis [useit.com].

My eyebrows raised at the thought of trying to do traffic analysis to identify whether or not there's "enough" content to satisfy the low end of the Long Tail.

But I was also surprised at the overall references to $/page payout. I must have missed when he moved from a focus on usability to start including what are essentially AdSense publisher payout issues.

[darkmage]

While I understand the point about picking the right type of analysis, the rest of it is either grabage or useless. Why would you actively plan to create low volume pages and why does he assume the curve would stay the same if you did? If you went to a site that was full of useless garbage would you keep going back or find another one that is concise?

Also, the logic is back to front. You can't create a page and say 'this will get exactly x impressions'. Impressions change over time and you can guess, but I have seen pages move around wildly with SERPs, new links in, a mention of the topic on TV or no reason whatsoever.

Then you are assuming the traffic mix is the same.

I have seen this type of 'analysis' in the real world by people who have a great understanding of graphs but not of people - it ends in disaster.

[ronburk]

Why would you actively plan to create low volume pages

Oh, possibly because you've exhausted the traffic you can get from higher-volume searches and you have the ability (via twigging your own reasonably active forums) to get others to generate the low volume pages for you.

For example, I've just generated a low volume page for Brett for the infrequently searched term "droopy tail analysis". Won't bring him much new traffic, but it didn't cost him much either.

One recurring question in this forum is how to decide when a website/topic has reached the point where it's time to move on. Looking for the droop in the tail is at least a data point to consider when making that decision.