Take a look at the Excel Fuzzy Lookup add-in. It is free to download from Microsoft. The tool implements the Levenshtein edit distance algorithm to return a similarity score between rows. Among other customizable features, the tool also allows you to set a threshold for how well the values must match in order to return the approximate result.
Rather than returning the count like you mentioned, the tool will return a table with the most similar match in column A relative to each entry in column B. From there you could simply use a filter and perform a count on the resulting column.
http://www.microsoft.com/en-us/download/details.aspx?id=15011