Confluence Search Doesn’t Have to Suck

Confluence is great. Really.

This is a guest post by Michael Cizmar, founder and chief architect of MC+A, which also appears on the MC+A Blog.

Confluence’s online editing, extensive add-ons marketplace, and ease of use has made it an ideal platform to write documentation and collaborate. With its deep integration with other Atlassian products like JIRA, it has become the defacto publishing platform for software developers. Confluence is great, but no one buys it for its search capabilities.

Out of the box, Confluence can index and search Confluence pages and the following attachment formats:

• Word (.doc), Powerpoint (.ppt) and Excel (.xls)
• Text
• PDF
• HTML

When performing a full text search, you can filter by: Space, Time Modified, and Article Type. Additionally, you can search by labels. All of which are generally accepted as required features in such a tool.

Search in Confluence Can Work Like Google?

Search in enterprise applications like Confluence and SharePoint not being as good as commercial search engines is a common complaint we hear from many enterprise customers, especially as these systems grow. Apparently, we are not alone. If you do a Google Search for “Confluence Search Sucks” or “Sharepoint Search Sucks” you will find plenty of users with the same sentiments. Impossible you think? Well, not any more. Our guest post from Michael Cizmar, founder and chief architect of MC+A talks about Google Search Appliance, and how it can improve search results in Confluence.

There are some basic problems with this:

1. Modern search technologies go beyond the basic Lucene index to provide relevancy with signals beyond that of a reverse index.
2. As a system, Confluence is a walled community trapping its content within the repositories. Companies very often have multiple Confluence instances each requiring integration.
3. Enterprises often have more than simple Microsoft-based office attachments which they would like indexed and to be searchable.
4. Users expect the ability to refine their search results.
5. Administrators want to tune what they consider higher value content.

Enter the Google Search Appliance

The Google Search Appliance (GSA) is a hardware platform which allows you to conduct searches across all of your repositories saving you time from having to perform individual search executions per content systems using Google’s universally recognized algorithm. Out of the box, it has basic features such as:

• Indexing 220 file formats
• The ability to serve multiple content sources in a search result
• The ability to create filters in the result set based on tags, metadata and elements extracted from the content through advanced pattern matching
• The ability to boost and provide best bets in the results
• A self tuning system which adjust results based on search usage

Having one place for our teams to create content for and collaborate around client and internal products helps us get work done and is a the core of what makes Confluence great. We take knowledge creation seriously; creating both internal product knowledge and external client deliverables. As our team members often find priorities changing as we “reconcile” client priorities against our internal product release schedules we lean heavily on Confluence as our collective memory. This allows our team to shift gears to address issue by priority without completely losing our path.

How Do You Marry the Technologies?

The GSA can crawl a Confluence site via its web crawler. However, web crawling is generally an inefficient process as the crawler has to self-discover changes and once crawled, how do you provide a security context?

Enter the MC+A Confluence Connector for Google Search Appliance, a stand-alone web service which can be deployed on-premise or in MC+A’s cloud. Using the Confluence API, the Connector transforms data into a GSA format fetching designated spaces, the pages, their security permissions and attachments. We designed the Connector to crawl both instance types: on-demand and on-premise. You can set up the Connector to index Access Control Lists (ACLs) with each page for early binding for quick searching. Simply set up user credentials and your Confluence URL and you are indexing Confluence!

Using this Connector not only improves the searching of Confluence content but allows that content to be surfaced in every search result to the GSA. This means that a user does not have to know specifically where the space is located in order to search for it. They simply go to the GSA and execute a search across all of our enterprise repositories.

How Does One Get This?

There are three pieces required for the system: Confluence, A Google Search Appliance, and the Connector. MC+A resells the Google Search Appliance and sells the connector direct to customers. Please visit and complete the contact us form to learn more. Additionally, MC+A recommends you speak with ServiceRocket to ensure that the system is optimized before starting significant customizations.

Is This Going to Be On the Final?

Confluence is a great content creation problem that has difficulty sourcing that to all parts of the enterprise. Using a Google Search Appliance provides your organization with a Google like relevancy with your internal data. We’ve seen increased page views of Confluence content when a GSA helps direct users to the pages that they’ve already created.

MC+A and ServiceRocket can help you leverage that to its fullest. Got questions? Find me online or comment below.