Reading the competitor gap

With a project, two screens read one run sideways. Domain overview’s Competitors tab, opened on your own site in your project’s market, lists your site and the rivals you track, measured or not, with what each one is worth and how much of your keyword set each one overlaps. Keyword gap shows the matrix of who ranks where for each keyword.

Keyword gap, with a project and without

With a project, Keyword gap reads the run described below and sorts each keyword into missing, weak, strong or unique to you. You can also name your own domain and up to 4 others and compare them with no project involved: the same comparison is then built from Domain overview’s answers, and for each domain you have not yet looked up it offers a press that fetches it, one search for each. Domains already looked up are read back at no cost.

It reads one run, and says which

In project mode, everything on the Competitors tab and in Keyword gap but share of voice comes out of a single run — the newest recent one that actually holds domain rows, of any status, because a run that failed on its fifth rival still measured the four before it. Both say which run they read and which market it asked in, on every tab, read off the run rather than off the project, since a project’s market can be changed and the run’s cannot. When Keyword gap compares named domains instead, it reads the answers Domain overview fetched and says the oldest day they were fetched on, rather than a run.

Traffic value is an advertising equivalent

The money column is what a domain’s organic traffic would cost to buy as advertising clicks each month, estimated by the keyword provider. It is not revenue, and nothing here can say what a competitor earns. The traffic column beside it is visits.

Overlap, and the difference between unknown and none

Overlap divides by your own measured keywords rather than by the union of both sets, so it answers “how much of what I rank for does this domain also rank for”. A rival whose keyword step failed has no overlap at all rather than nought per cent, and a rival the run never measured draws no column in the matrix — it is named in a notice instead.

Share of voice is the tracker when the tracker can answer

Where you have tracked keywords and a weekly round that can answer, share of voice is the tracker’s own figure — the same definition the Dashboard prints for your own site, so the two cannot disagree. Where it cannot, the run’s own estimate is used instead, and the column heading says which of the two you are looking at, because the instrument changing under a comparison is the one thing a reader has no other way to see.