Star Date: 250413. 1600 CMBUK
One of the biggest complaints I've had from users about using Confluence is the search results. Without getting to specific, many many people have told me that it doesn't work for them, that they get too many results or none.
As someone who does 99% of their work in Confluence (as admin, content designer, tech writer etc) I'm very used to the environment and searching. I'm also used to using the more advanced search techniques such as AND/OR/NOT and double quotes, so I've never really found searching to be a problem.
But what is your experience? What have you done to improve the built-in search? For example, have you used any of the plugins, have you restructured your content to improve navigation and thus lessen the need to search?
Anything you can tell me will be welcome, I've got to do something about this for our wiki, but my research in the Atlassian user info and the plugins hasn't filled me with hope.
If the Confluence install is public, using Google with the site: paramter gives you a ton of power. For example: https://www.google.com/search?q=site%3Aconfluence.atlassian.com+greenhopper+error+&aq=f&oq=site%3Aconfluence.atlassian.com+greenhopper+error+&aqs=chrome.0.57j58.14091j0&sourceid=chrome&ie=UTF-8
I hardly use it, because I have a couple of pages that I frequently visit and access them directly by url autocompletion.
What I really find annoying is that spaces can only be found if you start the spacename at the beginning. A space named abc def ghi won't be found when searching for "def" :(
I think the cure imho is a richer 3rd party search indexer that Lucene.
AppFusions has done a couple integrations like these and they are sweet - calibrated results, additional metadata, more filtering options, embedded results in Confluence like you are used to etc.
In defense of Atlassian - search is not a nothing technology. Atlassian focuses on the techs they are good at - purposefully, not getting distracted by OTHER significant tech areas.
Conversely, for 25+ yrs, there are corps that dedicate their lives to this problem and only this problem - and they are good at SEARCH.
Integrating is best of both worlds - imho. And bonus is that you can also mix in OTHER content repos into the results (beyond Confluence - with federated results).
Email me if want to make this happen in your org.
It is one of my favorite topics in the house (b/c of other integrations where it also comes up!).
Thanks for your info and thoughts. I had a look at the plugin searches but I wasn't convinced that any of them would really make a difference, at least not one that justified the cost. I'd be delighted to be proved wrong though. :)
Which engine would you suggest? It'd be handy to do some background reading on it.
I'm running a survey at the moment on how people use the wiki and I know that the biggest moan is going to be the search. It's a long term goal for me to get it sorted, so I'm more than willing to have a chat about it.
Hi Mick -
Its not a plugin-only solution.
Well I shouldn't limit it to that - here's a big list - take your pick, and as you can see - a pretty flushed out industry of flavors.
Large corps will often have a box in house, and they can get Confluence integrated for federated search.
But in my humble opinion, its not a battle that Atlassian should pick solely - just like most other colab platform vendors do not (unless they want to BUY a search tech outright, which I doubt will happen since they are a collab and dev sw company, not a search engine company).
Lucene and Solr are the most popular index engines bundled in Enterprise collab systems - so Atlassian's design decision here is not at all outa whack.
But if you want the uber power - you need another search indexer integrated - which brings other benies too.
I'm beginning to think part of the cure will be to make the content as clean as possible (i.e. no duplication and only relevant info to reduce the amount of similar results. This means archiving out of date or deprecated info out of the main space and into an archive space).
This also means restructuring the content so that people find it easier to find info by navigating to it.
If you group content, then you can give people landing pages (a page that contains links to a set of relate pages) where they can search from. Once they get used to this idea, they will navigate to the places they're expecting to find info, rather than guessing their way to it.
In some instances these groups can be sub-pages of the grouping page, in which case you can add the search sub-page macro so that only those pages willl be searched.
I'd also remove the macro that auto-suggests results as you type as this invariably lists pages that do not match in anyway, which is annoying and demotivating for users.
What else do people do to improve searching?
If I get feedback from a user that he/she was searching for something and the most valuable results didn't come close to the top, I add an appropriate label to the page(s) that should be found.
Same, if I realize that users usually tend to search for a certain term, but we're using another (similar) term in our documentation. I add this alternative term as a label to the pages I want to be found.
Sorry for reopening the old thread but this seems to be the most fitting place to ask. I was wondering how you proceeded since 2013 to improve the search functionality in Confluence? What measures did you take if any at all?
My colleagues and I have the same problem today and I would like to do something about it.
I use https://developer.atlassian.com/server/confluence/lucene-boosting-strategy-module/ to weight the search correctly (Atlassian's default is better suited to social media than anything else, which is the one thing no-one uses Confluence for)
Connect with like-minded Atlassian users at free events near you!Find a group
Connect with like-minded Atlassian users at free events near you!
Unfortunately there are no AUG chapters near you at the moment.Start an AUG