Here’s a crazy one.

Wordtracker only shows keywords with at least 2 searches in a 90 period on Metacrawler, which can be more or less equated to the total searches each day on all engines.

I’ve found most targetted keyword niches exhibit more or less normal power-law curves (as opposed to Nielson’s drooping tail, when graphed double logarithmically against a true power law trendline, keyword lists often have slightly drooping heads). So, speaking very roughly the second term has half as much traffic as the first, the third has a third, the fourth a fourth and so on.

When you’re building a keyword list with wordtracker, the “end” of the tail for you will be where the keywords have only 2 searches. So using our understanding of power laws we can estimate that this 2 searches point is at keyword number N, where the traffic of the very first term divided by N is 2. If the most popular phrase in your list has 400 searches the power law curve would suggest keyword number 200 has 2 searches a day.

So a quick rule of thumb to guestimate keyword niche breadth is to divide the traffic of the first time in half.

I’ve seen it be a decent rough estimate in nearly every “normal” niche I’ve looked at.

If you liked this post, don't forget to subscribe to my RSS feed or my email newsletter so you never miss the science.