Using underscores in the meta data field of knowledge articles - Yes or No?

[AndrewRR]

Hello

I have spent quite a few hours researching the best practice of the meta data field in knowledge articles. Specifically, is the use of udnerscores still advised, or not?

So hard to find relevant documentation about this in SNow - but moving along. 

Older SNow forums and documents (prior to 2020) do mention the sue of udner scores to assist with Search, but I am now starting to doubt this.

Can anyone confirm whether it is still necessary to use underscores in meta data?
Thanks in advance

[amarks]

Hi @AndrewRR ,

This was driving me crazy too with a lot of conflicting information out there, so I finally did my own analysis. (Keep in mind, there might be environmental variables based on your ServiceNow environment.)

BLUF

In our environment, the underscore is valued as a space, and appears to be no benefit to using underscores.  Our users had been informed that phrases such as "computer_browser" would be ranked higher in a search when people searched for that phrase. This wasn't the case.

Based on this, some authors were putting into the Meta field every permutation (ex. computer, browser, computer_browser, pc_browser, laptop_browser, mac_browser).  Not only did people search not match on the actual phrase, but it boosted the term browser in the same way "keyword stuffing" boosts the ranking (it sees this as browser x 5).

In our environment, the heaviest weighting goes to Short Description.

The second heaviest waiting goes other fields we have indexed.

Finally, Meta weighting comes in third (of those three, but there are other factors).

From what I've found, the Meta field should be used for

• Unique terms not found in the other fields
• Synonyms (that aren't already in your synonym dictionary)
• Abbreviations (that aren't already in your other fields)

I hope that helps.

-Adam

[AndrewRR]

Thanks Adam
Appreciate the updated info. Especially about how it can be environment specific, and how in your environment the Short Description has the heaviest weighting.