Querying all counts for the different search results in the same request led to timeouts, so we now only calculate the count for the *current* search results, and request the others in separate asynchronous calls.