Getting inside your head: Analysing search behaviours
It’s useful to examine the way people find and use government websites, as this can reflect underlying needs and preferences. And the more we understand about those underlying needs, the better job we can do of designing content and services to meet them.
Most people start with Google, so we thought we’d start there too. Usage analytics tell us where people go and what they do online, but the words and phrases people use in Google tell us something more - we get a glimpse of what they already know, what context they have, and their starting intent and expectations.
Patterns of behaviour
We engaged Donna Spencer from MaadMob to help us analyse around 13,000 search terms from a bunch of Australian Government websites, and quite quickly some patterns of behaviour started to emerge.
A person uses Google to locate a specific website or agency, including:
- Entering a URL (or approximation of a URL) - “ato.gov”
- Entering a version of an agency name - “ato Australia” “tax office”
- Asking for contact details - “contact medicare” “immigration office Melbourne”
Known item search
A person knows what they are looking for, including:
- Using a specific term that gives a specific answer - “auskey registration” “visa subclass 489”
- Using a specific term and defining a specific format - “2014 tax calculator” “medicare claim form”
- Using a general term that has a specific answer - “when is spring in Australia” “who can certify documents” “what electorate am I in”
- Starting a transaction - “ABN lookup” “my gov au login”
- Checking on a transaction - “track my tax return” “check my visa status”
A person only has a rough idea of what they are after, and may not have the terminology to describe it or know who might have the answer, including:
- Using a specific term that has a broad answer - “age pension” “self managed super fund” “justice of the peace”
- Using a broad term that has a broad answer - “Australian visa” “medicare levy” “business plan”
How can we use this information?
We will use the information to design better services that match these behaviours. So we’re working on:
Better search engine optimisation – global search engines (like Google or Bing) do a pretty good job of helping people find what they need, but we need to optimise our content to make sure we return the right result for common behaviours, and accommodate the language people naturally use, variants, misspellings and so forth.
Better interaction pathways, for example:
- Direct specific queries to specific answers, and broad queries to broader information that allows users to understand context and drill down further.
- Direct contact requests to an appropriate contact page, with the right context. For example “contact Agency X” should go to that agency’s contact page, “contact service X” should go to a contact for the service if it exists.
- Include explanations of how people can check on the progress of transactions and advice about processes and timelines. Make these explanations easy to find without having to login, even if the actual progress check requires a login.
Content targeted around user behaviours, for example:
- Creating content around common queries – if everyone is asking a particular question, we should have a whole-of-government answer.
- Using whole-of-government sites to answer common general queries that cross multiple agencies . This saves people from navigating across multiple sites and having to digest multiple agency-specific answers.
- Using whole-of-government entry points to provide topic and life event overviews for people searching broad term
It also draws us to some other gaps and opportunities — one that comes to mind is how we might be able to use smarter search technologies or virtual assistant style technology to help match natural language queries to the right information, or help people refine down broad queries. We plan to design and test some of our work against these behaviours over the next few months and see what the impact is.
What do you think?
Do you recognise these behaviours in yourself or in the users you provide digital services to? Do you have ideas about how to design for these behaviours?