How do people find local information online?
Primary research undertaken by the writer strongly suggests that most locals try to access all forms of information about their area utilising a global search engine (Google, Yahoo etc) in the first instance rather than a dedicated web page or an online directory /community portal - locally based or otherwise.
In fact most of these users do not appreciate that in virtually all instances, the sought after information is provided indirectly (through a hypertext link from the Search Engine Results Page (SERP)) by a human edited directory - in descending order of importance (Page Rank):
| International | - | Wikipedia |
| National | - | Business Directories |
| Local | - | Portals & Local Directories |
What’s more local directory-type search queries overwhelmingly comprise only two keywords – most often the product/service, first, followed after a space by the geographical location e.g. ‘plumber’ ‘wollongong’
This search approach, whilst intuitive and often furnishing an adequate result, has a number of significant limitations and biases, which are discussed below:
Firstly, the choice of product or service as a keyword:
Search engines index (compile) web pages in a number of ways and virtually all processes involve text in some form or other. By far the most important factor, in terms of SERP ranking, is the use of keywords in the URL and Title.
This factor alone virtually guarantees that directories will always appear above standalone web pages - notwithstanding their relevance and/or quality of content.
It should be noted keywords in the body copy (number, density, format etc) do have a positive effect on page ranking however in most instances this is not sufficient in itself to elevate a website’s positioning to the critical first page of the SERP – usually about ten listings. NB: nearly 50% of all searchers do not proceed past the first page of SERPs and only about 15% continue to the third page.
The order of keywords can also negatively impact the quality of search finding (order in SERP) however the cognitive process involved makes the current user approach unlikely to change in the short to medium term – certainly for the majority user group i.e. non-professionals.
Secondly, the choice of location as the second keyword:
Ideally a search engine would prefer to use a Zip Code as a location reference for indexing purposes however from the user viewpoint these numbers are non-intuitive (in terms of position and size) and generally speaking most people are not aware of nearby Zip Codes i.e. consecutive numbers are not necessarily indicative of contiguous areas.
To date very few directories in this country have used Zip Codes as the sole or major means of describing their footprint. In addition users do not usually wish to limit their search to a small area and much prefer to make their choices from a larger selection within easy driving distance.
Not inconsequentially directory advertisers also wish to reach the maximum number of potential customers. The critical minimum viable size of a local directory (in terms of listings and population served) is also yet to be determined.
A number of approaches have been attempted in this country and the optimum solution has yet to emerge although community based directories (such as SMARTPAGES) do have some natural advantages.
Thirdly, page rank:
On-going longitudinal studies by the researcher indicate, all other things being equal, a high PR web page will appear on the SERP above a low ranking website with more relevant and/or unique content – if the PR difference is more than two e.g. a perfectly keyword matched PR3 website can only jump a less relevant PR5 in positioning. This finding is particularly noticeable since a Wikipedia (PR8) related web page will nearly always dominate a SERP notwithstanding the indexed presence of better qualified (in terms of keywords) alternative options.
Fourthly, listings:
The quality (in terms of comprehensiveness and freshness) and number of listings within a directory is not currently used as a determinant of Page Rank, when indexed by Google et al.
In some instances internationally based directories with very small numbers of relevant listings rank higher on SERPs than more comprehensive local directories.
This factor is a major impediment to improving the quality of local search, which currently does not match the standard of global search.
Fifthly, domain size:
Search engines index all web pages by matching keywords. When a user searches for information the search engine creates a SERP comprising all web pages indexed in descending order of importance (PR). E.g. As at January 2009
Google indexes 4,170,000 web pages with the keyword ‘wollongong’
Google indexes 443,000,000 web pages with the keyword ‘golf’
BUT only indexes 212,000 web pages with the keywords ‘wollongong’ and ‘golf’
AND indexes 211,000 web pages with the keywords ‘golf’ and ‘wollongong’
As a consequence when searching locally the domain sizes involved are relatively small i.e. in the hundreds of thousand of web pages versus tens or hundreds of millions.
Page ranking algorithms originally developed to sift through many millions of indexed web pages (to find the critical best 10 SERP results from numerous web pages of similar page rank) cannot function optimally and become biased towards non-local websites.
This factor is another major impediment to improving the quality of local search, which currently does not match the standard of global search.
A modified extract from a technical paper, titled An Analysis of Local Web Search, prepared in 2008 by Andrew Connery as part of his doctoral research at the School of Computer Science and Software Engineering, University of Wollongong
DISCLAIMER: Andrew Connery is the Founder and Managing Director of
Your Online Community Pty Limited publishers of SMARTPAGES online business directories