Organic CTR

Growing search traffic has long been treated as a synonym for improving rank. Then rank improves and traffic doesn't follow in proportion.

One of Pagedeck's help articles did exactly that. A month of work took it from eighth to third, impressions quadrupled, and visits rose by a factor of 1.3. Ravi found the cause by opening the results page himself: the title used the internal name for the feature, so nobody searching recognised it as the thing they were looking for. There is a step between rank and visits, and this number is where it shows up.

Divide clicks by impressions and write it as a percentage. Sixty clicks on 2,000 impressions is 3 percent.

1. An impression is one appearance of your row in a results list. It counts even when the searcher never scrolled far enough to see it, which is why deep positions accumulate impressions and almost no clicks.
2. The overall average is rarely useful. Break it out by query, by page, and by rank band. The average moves with whichever queries happened to get impressions that month, which tells you nothing about whether anything improved.
3. Compare within a rank band. Five percent is low at position one and high at position eight. Line up two pages at different positions and you'll mistake the rank gap for the quality of the title.
4. When rank improves and this number doesn't follow, look at the title and the description. The test is whether the searcher's own words appear in the title and whether the two-line description answers their question.
5. The layout of the results page moves it too. Four ads above the list, or a featured snippet answering the question first, lowers the number at an unchanged rank. A search that ends with no row clicked at all because the answer was already there is a zero-click search.

"We're third with an organic click-through rate of 1.8 percent. Start with the title."

Too few clicks for that position, so the title and description need work.

"Impressions are up four times and visits are up 1.3 times."

Rank brought impressions and they didn't convert into clicks.

"This query has ads filling the top of the page, so the number is low by default."

The layout of the results page suppresses the figure at any rank.

Ravi rewrote the titles of twelve help articles using the words people actually searched. Rank stayed where it was and this number moved: 2.1 percent on average to 3.4 percent, with visits rising to match. That it cost less than a ranking project was the part that made the quarterly report.