Intranet searching: popular search topics. A case study by Maxine Armitage

Print friendly version (80 kb)
Introduction
This case study analyses searching behaviour on a government agency intranet site. The agency is medium sized and employs over 1,000, mostly clerical, staff.
Search logs were used to look at the topics that people mainly searched for over a two-year period.
I wanted to know which topics were consistently popular so find-ability could be improved through changes to the information architecture, additions to content, review of metadata and search configuration, etc.
This study concerns search topics, not individual search terms.
Popular search topics
The 50 most popular search topics, over two years, in order of popularity were:
| Rank |
Topic |
Rank |
Topic |
| 1 |
flexi time |
26 |
corporate governance |
| 2 |
leave |
27 |
mobile phones |
| 3 |
newsletters superseded by the intranet |
28 |
occupational health and safety |
| 4 |
court of law associated with the agency |
29 |
marketing |
| 5 |
position descriptions |
30 |
call centres |
| 6 |
internal staff directory |
31 |
administrative system 'A" |
| 7 |
dlegation of authority |
32 |
allowances |
| 8 |
core business system 'A' |
33 |
chart of accounts |
| 9 |
penalties |
34 |
records |
| 10 |
abstracts of current awareness articles |
35 |
guidelines |
| 11 |
code of conduct |
36 |
the intranet |
| 12 |
standards |
37 |
recruitment |
| 13 |
privacy |
38 |
business model |
| 14 |
templates |
39 |
inspections |
| 15 |
media releases |
40 |
projects |
| 16 |
business names |
41 |
section 52 of a particular Act |
| 17 |
home building |
42 |
intranet content management system |
| 18 |
pay |
43 |
Christmas |
| 19 |
overtime |
44 |
memos |
| 20 |
core business system 'B' |
45 |
cost centres |
| 21 |
annual reports |
46 |
fax |
| 22 |
personnel handbook |
47 |
legislation |
| 23 |
salaries |
48 |
pawnbrokers |
| 24 |
aboriginal issues |
49 |
booklets |
| 25 |
rentals |
50 |
circulars |
Observations
- none of the five most popular topics had a presence on the home page
- only six of the top 50 topics were accessible from the home page
- types of information (eg standards, circulars, memos, booklets, position & descriptions, etc), as opposed to topics, were very popular
- two core business systems were within the top 20 topics
- the third most popular topic was newsletters, even though these hardcopy publications had been superseded by the intranet
- although it is not evident from the above table I know from the search logs that topics such as 'Christmas' and 'mobile phones' were predictable in their popularity, peaking at certain times of the year or month
Implications
- the most heavily searched of the top 50 topics should be given some presence on the home page
- the position of popular topics within the information architecture should be reviewed. If they are deeply buried perhaps they should be brought further up the hierarchy.
- searchers of this intranet related strongly to 'types' of information, therefore consideration should be given to providing greater access to resources via 'type' metadata. The types 'Forms' and 'Policies' have already been given a high profile on the ever present top navigation bar - that this was a good decision is confirmed as neither of these types feature in the top 50 search topics.
- the two core business systems featuring in the top 20 topics are not accessible through the intranet and little information is given about them. Further study needs to be undertaken into the feasibility of making these systems accessible via the intranet, making reports generated by these systems available via the intranet, or at least providing contact details for system managers. This point is related to the importance of making the intranet relevant to business needs. See the Find It paper Intranet searching: changes in searching behaviour (http://www.finditim.com.au/papers/changes in searching behaviour.html) for further discussion on this point.
- searches for superseded newsletters decreased over time indicating that the message was getting through that the intranet had replaced newsletters. In hindsight it would have made sense to state this on the home page of the intranet during the first few months. Another strategy could have been to force a top search result which clearly stated that the newsletters were defunct.
- information demands can be pre-empted. For example, the search logs show that staff begin enquiries about Christmas leave entitlements in November so information should be posted in November. If this is not possible, then information should be posted indicating that information will not be available until later. This will save staff the effort and frustration of searching for information which isn't there.
Conclusions
Search logs are a valuable source of information about behaviour on the intranet.They can be used to:
- provide sound reasons for tinkering with the information architecture
- allow you to anticipate information needs
- indicate when information should be provided about information which isn't there
- indicate which metadata tags are useful and which terms are commonly used.