Getting into the Weeds re: Content Gap Analysis

When reviewing our Searches with No Results report, have you ever noticed queries for which YOU KNOW you have content? Well, I’m here to tell you that you’re not crazy. For better or for worse, because that report pulls session data, queries with results are included along with queries with no results when they occur in the same session. In other words, during the course of a session, if I enter three searches, and the first two return results but the third doesn’t, all three of those searches will appear in the Searches with No Results report.

Why is this important? As an admin, if you’re trying to get a snapshot of queries that really and truly produced zero results to help you identify content gaps or opportunities to add synonyms to the NLP Manager, the only way to do it using the default Searches with No Results report is to test every query. Doing so is time-consuming and can be more than a little frustrating when many of the queries you test actually do have results.

So what’s an admin to do? Don’t give up! I have a workaround just for you.

Instead of referencing Searches with No Results, go to Search Analytics > Conversions > Session Tracking - Details.

1. Be sure you’re on the right search client and have set the date range within which you want to work.
1. Please note: If you want to perform this analysis over a long period of time, such as several months or a year, I highly recommend pulling the data for one month at a time, then later consolidating everything into a pivot table, which I’ll discuss below.
2. Click the Envelope icon in the upper-right corner of the report to email it to yourself.
3. Enter your email address and select the date format you prefer.
4. Under Activity Count, select Per Activity.
5. Tick Include Advanced Facet Details. 

1. Finally, click Send.

Depending on how large the data set is, it may take several minutes to receive the report in your inbox. When you receive it, you’ll need to extract it from the zipped folder to view it. By default, the extracted file will be a .csv in your Downloads folder. I recommend saving it as an Excel file in a more permanent location on your computer.

When you open the spreadsheet, you’ll see many different columns going across. The one you need for this exercise is called Search Result Count.

Filter on that column, deselect Select All, tick 0 (zero), and click OK.

The number of rows in the sheet will be reduced into a much more manageable data set, which you can now easily organize into a pivot table to identify the top queries that have repeatedly produced zero results over the given date range. If you do want to create a pivot, I recommend opening a new tab and pasting in the following columns:

• Session Identifier
• Email
• Activity Detail (which is the column that contains the queries that you’ve now filtered to a zero-results count)
• Facet Type
• Facet Value

Including the Email is important for two reasons:

1. To see how many different users tried searching for the same zero-results queries.
2. If a subset of your content is public and the rest behind a log-in, then one or more zero-results queries could have occurred because the content was not publicly available, but unauthenticated users tried searching for it.

Similarly, including facet details will help you determine whether a number of users are filtering themselves out of existing content. For example, I recently observed a clients’ customers consistently selecting “Known Issue” as a facet but getting zero results because the client hadn’t tagged the relevant content as a known issue. If you repeatedly see your customers identify something as a known issue (or anything else), then maybe it kinda is, so you might as well tag it as such, thereby respecting your customers’ input and facilitating their success.

The penultimate task in this activity is to test each query in your target list. Does it represent a true content gap, or is it more appropriately a synonym of language your business regularly employs? If your data set spans several months or a year, testing is also important to eliminate queries tied to previous content gaps that have since been filled.

When you’re done, you can turn the list over to your SMEs/KDEs/Docs team and walk them through your process. They can rest assured that you’ve done your due diligence and that every query on the list is something multiple customers were trying to find for themselves – but failed – because the content didn’t exist. Now it’s their job to fill the gaps.