How search can help you understand your audience - part 1
How search can help you understand your audience
Subsequently I have been giving a series of presentations on how we can use this information within the BBC to feed into our design processes, our content creation, and our understanding of our audience. I believe that these are principles that can be well applied to any UK based non-commercial website.
It seems to me that many of the metrics with which we measure user interaction with the web are deeply flawed, and provide ample evidence that the internet was invented by physicists and technologists, not marketing and advertising executives. Page impression figures do not convey any information about whether the page view was satisfying or frustrating. Not all users accept cookies, and not all cookies can be associated with a single user when they are issued to computers across corporate networks or public internet access points. Tracking individual session data still doesn't tell us whether the user went to make a cup of tea or answer the door whilst they spent the measured 3 minutes on the page. Beyond that, the effect of caching by ISPs can lead to users not even seeing the latest version of a page, let alone the problem it creates for capturing behavioural data.
However, the action of typing a word or phrase into a search box is a very direct interaction between the user and the site. We can safely assume that when the user typed in that word or phrase they expected "something" back from the site. Few people except the very bored, Google-whackers, or those with a professional interest in search mechanisms enter search strings looking for a null result.
Using search to learn the user's vocabulary
One thing that becomes abundantly clear from even a cursory examination of the search logs at BBCi is that the BBC has an extremely diverse audience, who may be looking for the same information, and for information that we have, in a myriad of different phrases, dialects, and even languages. Without realising it we set up barriers in our site to users finding the information they require, merely by the language we use. This is something that any medium to large scale website will suffer from, for the following reasons.
A well written website will have some sort of style-guide, whether it is standards compliance (w3c), plain use of language (Plain English Campaign), brand consciousness (easyJet), to the downright pedantic (currybetdotnet didn't use to use upper-case on the homepage at all as a design 'quirk').
However users don't come with a built in version of the style-guide. A decision taken in a committee room about whether a site offers 'ecards', 'e-cards' or 'e cards', can have a profound influence on the user's search experience if they are looking for one of the variant spellings.
At the BBC the website can variously adopt phrases like 'Feedback', 'Contact Us' and 'Email the Show' to allow the audience to interact with it. However we know from search logs that the BBC's audience often asks natural language questions like "how do i write to BBC3", and can be viewing a page with a link to 'Feedback' on it and then type 'email address' into the search box.
Correctness vs Common Usage
One area where there can be a particular disconnect with the audience is if you are obliged to use 'correct' terminology, when in the real world 'common usage' is much more prevalent. A classic example of this is with BBC News and the country 'The Netherlands'. A search for 'netherlands' on BBC News yields around 100 results at the time of writing.
In the UK, however, it is much more common to call the country 'Holland'. A search for 'holland' brings back fewer results, and in the first page of results produced by the search technology, the top story is about Jools Holland, and one of the stories is about Tony Holland. Using 'Best Links' we are at least able to insert links to The Netherlands country profile, and the top headlines from europe.
This is where using synonyms and placing 'Best Bets' within search results can play a crucial role in bringing users to the content they are looking for. Where the style-guide allows, we also consider placing synonyms within our navigational structure to allow users to browse towards their goals, even where their vocabulary differs from ours.
An example of this is that the BBC offers a Films site at www.bbc.co.uk/films. If we stuck strictly to our branding and our style guide then we would only list it in the A-Z Index under 'F'. Using synonyms, and the information fed into the design process by search, our A-Z Index team list the BBCi Films site under 'C' for cinema, 'F' for films, and 'M' for movies, providing alternative navigational routes that suit our users needs rather than any internal branding priorities.