The search indexing has been completely redesigned to ensure that the most appropriate results are returned first. As part of this it was necessary to build in some intelligent algorithms, so that you don’t have to be quite so specific, accurate, or get the left-to-right’ness correct to find the company or officer you are looking for. Previously, searching was causing a lot of customer difficulty, and needed addressing.
The upshot is that we unavoidably get more matches, but the likelihood that the customer finds what they were looking for within the first few results is high, even with imprecise search terms. The more search terms you supply, the higher the chance of finding the result (a single word term is usually unhelpfully noisy…) and as you descend the result list, they will of course become increasingly less relevant, but still useful if you deliberately used a rather general search term.
The result ordering is now relatively predictable, whereas previously it looked a bit of a jumble. It is complicated to describe, but in general the results attempt to group themselves by relevance, and then alphabetically within that group. In all, the ordering “looks” right to users, and they no longer question the ordering. Consistency is always good!
There are also nearly three times the number of companies indexed, as we have included the last six years of dissolved companies. This will also increase the number of matches you receive, but unless you are pretty closely matching the dissolved company name, it will be ranked less important than an active company, as a decay function of it’s dissolution date.