Create a searchable knowledge base on Confluence

[Ian H]

As I am sure is very common, in my Confluence Cloud instance I have a huge number of mixed purpose pages. Some are of lasting value as "knowledge". i.e. documentation of large ERP systems we develop. But there's a whole host of pages which are of merely transient value - ideas, bug replications, Xmas party plans, ideas to add to our garden, cakes we've baked (I could go on).

A very common complaint from users is "we can't find anything". When we do a search, it returns pages which are simply not valuable "knowledge". If I want to understand some functionality, I don't typically want to know about bugs we fixed in that functionality 3 years ago.

I'm struggling to work out the best way to enable users to search only "knowledge" pages. I have a large number of uses very willing to add a label "knowledge" to every relevant page - but I am then reliant on users to know to filter a search for that label. I have many users very unfamiliar with Jira and need the very simplest search option. I've tried via automation to copy all pages with that label on to a dedicated knowledge space and then set up searches for that space only, but the automation I am finding very problematic for many reasons, not least it timesout on large pages with many screenshots.

I'm basically nervous that I am overlooking something really really simple.

Anyone have a similar challenge or any brilliant ideas how to approach this?

Much appreciation in advance

[Aron Gombas _Midori_]

If you don't want to actually fix the chaos, then you could try to convert the Confluence contents to an external site using Scroll Viewport. If it is possible to filter the content to convert during the transformation, then you could leave behind everything else, but the "knowledge" type content. The result of the transformation can made searchable by the app.

If you want to fix the chaos, and get back the control over your ever-growing content, then use the Better Content Archiving app:

1. Set up custom statuses "Irrelevant", "Expired", "Duplicate" and the corresponding CQL queries to find irrelevant content (party plans), outdated ideas, bug replications, etc. Then the app will classify each page and blog post and assign one of these statuses. With that, you have an overview about the statuses (see below!).
2. Then you should set up automation notifications to ask people to update outdated pages, archive the duplicates and so on.
3. It won't happen overnight, but people will start go through stuff, decide what to do and ultimately improving the overall quality of your content.

This app has been developed to solve that typical problem you are describing and then continuously working on maintain the quality so that the problem doesn't occur again.

(Discl. it is a paid and supported app developed by our team.)

[Ian H]

Thanks for this, Some useful ideas.

I probably did paint a chaotic picture of what we have - and we do. Archiving old irrelevant rubbish is absolutely needed. However, when I refer to "knowledge" I am talking of product documentation I want to share with customers ultimately. There's lots of valid pages that are old and I do not want to archive - like documented processes and best practices for example. So solving the chaos is I feel only part of the challenge. 

I will for sure look in to the option you've suggested. 

[Aron Gombas _Midori_]

There's lots of valid pages that are old and I do not want to archive - like documented processes and best practices for example.

Absolutely! We have that type of content, as well. Like all teams have.

For that exact situation, the Better Content Archiving offers the action "Confirm". It means that "I reviewed this content, it is old but still valid, so I confirm it requires no changes".

Or, if you are completely sure that a page is old, will never change and will always be relevant, you can mark that with a custom status like "Evergreen".

The lifecycle options we built to the app are hyper-flexible.

[Ian H]

We've made some very good progress - thanks  for the support here.

We feel Refined is potentially our end goal but before we can justify costs and additional effort it would need, I am still on a crusade to gather all the "knowledge" articles in a usable structure. I've got a series of pages working for me using the "Live Search" macron filtered by the labels I have been implementing, which I believe can get me to a place I can demonstrate the potential.

Now battling with the unordered mess from many years to add those labels and deeply wishing there was a usable bulk update option I could use (either to move irrelevant pages or add labels en masse based on whatever selection criteria) but I see plenty of other posts on that topic without any real solutions.

Thanks again to everyone who contributed.

[Laura Campbell _Seibert Media_]

Hi @Ian H Do you have your Confluence instance organized into multiple spaces? If pages are organized into different spaces, you can add a search macro to the welcome page of each space and limit the search to that space. That way people who go to the "Office life" space and search from the home page will only find gardening suggestions and cake recipes, and people who go to the "Product documentation" space will only find documentation.

You will probably need to do some user training so people know there is a search option available from each space home page, or that they can essentially do the same thing by searching from the top menu bar and limiting by space. But other than that, it's pretty simple to set up (you might need to move pages between spaces in the beginning).

[Ian H]

Yes it is in multiple spaces.

What I regard as "knowledge" is spread across many spaces. It is created by many departments - L&D, R&D, Support etc.

Maybe we are very territorial - but I see users very happy to create pages in their "own space". When I try to establish spaces based more on their purpose, it breeds sense of reluctance. Users will publish things once they are perfect. "Perfect is the enemy of good" comes to mind. So I am trying to find a good balance here. Let them work in their own territory and publish from there what is valuable. More curation needed undoubtedly but on balance I believe it is worth it.

What I have done, rightly or wrongly, is used automation to copy everything that is valuable long term knowledge in to another space so that I can do precisely as you suggest with a search macro on the welcome page in there. But I am getting various problems with that and constantly question whether there is a better approach available.

[Laura Campbell _Seibert Media_]

Instead of copying with Automation, what about using the Include Page macro? That way any updates or changes in the original page will be visible in the separate space.

To help combat the reluctance of users to publish pages, the native page status option is helpful. Perhaps create some custom page statuses to better match the writing process of your colleagues. Another method is to lock down visibility on a page (for example only the author can see it) until it's ready. But I would recommend going with page statuses first to promote open sharing of info.

And I don't think anyone has mentioned it yet, but if you go with your label solution, you can insert the Content by label macro on a page and list all the pages with your "knowledge" label so people don't have to do advanced searches. If you have tons and tons of pages and it makes for a very long page to scroll, try using expand macros. I need to try it, but you should be able to put the Content by label macro within the Expand, and if you give it a nice title, users should understand what list of content they will find if they open the expand.

[Ian H]

For the Include Page macro, I would be reliant on someone manually creating a page in my Knowledge space and adding it for every article deemed to be knowledge - right? 

[Ian H]

I like the page status option. That has some potential to help for sure. I worry they would never mark it "done" or whatever we decide on but can see for some use cases I can make it work. Thanks.

[Ian H]

We are talking, ultimately if my ambitions are realised, of many thousands of pages.

[Ditte Simard]

Hi @Ian H !

I ran into a similar problem on one of my Confluence Spaces.  I ended up using "Refined for Confluence" (https://www.refined.com/for-confluence) to guide users to the documentation and pages that they actually needed.  It is pretty easy to use as long as you follow the tutorials on the Refined website and it has the added benefit of making a Confluence Space look more "professional".

I hope this helps!

[Ian H]

I had another similar tool suggested to me some time back and I bailed out on the idea due to the additional admin overhead it would introduce. But if I am to achieve my ambition of making this customer facing, perhaps that's just something I have to take on the chin. 

Thanks very much Ditte. I'll take a look.