I repeat: it is the basic concept of search technology! Think of a simple search algorithm that reads one hundred text documents and counts the number of occurences of each word in each document. The results are stored in an index.
Input comes in the form of a search query. The search algorithm compares it with the index and creates a search results page that may look like:

search query = widget
# 1. document # 12 = relevance score 7 (7 occurences of widget in document)
# 2. document # 96 = relevance score 5 (6 occurences of widget in document)
# 3. document # 4 = relevance score 4 (5 occurences of widget in document)
# 4-100 documents # 1-3, 5-11, 13-95, 97-100 = relevance score 0 (0 occurences of widget in
document)

Note that the documents are arranged by the descending order of their relevance score.
Note that the relevance score gap between the documents ranked # 2 and # 1 is two points, the relevance score gap between the documents ranked # 3 and # 2 is one point, and the relevance score gap between the documents ranked # 4 and # 3 is four points.

A real search algorithm such as Google's does not show documents of relevance score zero; it does not show the number of the document in the index (because the user does not need this information); the consolidated relevance score of each page is calculated by the very complex and secret search algorithm from in-page qualities and from qualities of inbound links.

Most important: real search engines do not show the actual relevance scores because by observing their fluctuations in response to optimization one can understand how the secret search algorithm works!