V13 Internal Search Keyword

[vvzer]

Hello,

do someone can help us with this question?

We're isolating with event v13 the internal searches, it means search query in website internal search bar.

When we applicate the event v13 and we isolate the only "suggestion" the value returned should include only occurrences related to query suggestion which are available on the website - while adobe returns also query suggestions which are not available on the related website.

We give you an exemple to make it easier: we want to see only all click by users on internal search bar website suggestions (so we want to exclude direct query digits by users), so what we do is: we use v13 "internal search keyword" and we filter only "search results_ suggestion" . The issue is that the results returned contemplated suggestions that does not exist on our website.

Could you help us solving the issue or, in alternative, understand why this issue occurs?

Thank you

[Jennifer_Dungan]

A few questions:

• Does your search only allow for "suggestions" and not freely typed values? Are you sure... if you type in a term like "not a suggestion" and hit enter, without selecting a suggestion, does the form still submit and take you to results? A lot of sites allow for both.. auto-suggested and free typed....
• Where are the keywords being tracked? On the submit search (after selecting a suggestion)? On the resulting Search Results page?
• If on the Search Results page, is the search term part of URL parameters that could be manually updated?

If you feel comfortable sharing your website, I could see your search in action (a private message is fine if you don't want to post the URL publicly) 

[vvzer]

Hi Jennifer! Thanks for your feedback - We try to reply to all your questions 

• Does your search only allow for "suggestions" and not freely typed values? Are you sure... if you type in a term like "not a suggestion" and hit enter, without selecting a suggestion, does the form still submit and take you to results? A lot of sites allow for both.. auto-suggested and free typed.... > our searches in the website internal search bar both contemplate suggestions and freely typed values BUT the events tag are differently depending on the user behaviour. We try to be more clear: free typed are tagged with a different event which is v13 "internal search keyword": keyword:name of query digited - while click on results suggestions by users are tagged with v13 "internal search keyword" : suggestion:name of suggestion. 
• Where are the keywords being tracked? On the submit search (after selecting a suggestion)? On the resulting Search Results page? > keyword and suggestions of the internal search bar are tracked in v7 Page Type = searchresult
• If on the Search Results page, is the search term part of URL parameters that could be manually updated? > we don't have this info as the tagging process is not up to us

[Jennifer_Dungan]

Ok, so the developers or implementation team are passing you the search keyword values, and the "classification" of "suggestion" or "keyword"?

How in-depth have you tested your site? Is it possible that the wrong info is being sent to analytics, causing a mismatch between what is available as a suggestion and the information collected?

How many examples of mis-matching do you have? Is it a lot, or a small subset? Could this be the result of an intermittent bug that is hard to replicate?

See if you can manually test and replicate a mis-match? Try being an "advanced user" and try to change the URL params on the search results (i.e. do a Suggestion Search, then change the values for the "search keywords", in theory, there is a param there that indicates its a suggestion, and if the Keyword is changed then it could track as a mismatch.

If you still cannot replicate the issue, then maybe its an intermittent bug... Does your QA team have a set up for bulk automated testing? Could they create a script to test in bulk and see if they can catch a failure?

[vvzer]

Hello Jennifer, we Try to reply to all of your question:

Ok, so the developers or implementation team are passing you the search keyword values, and the "classification" of "suggestion" or "keyword"? > No, this data is extract from our team of analytics since with observepoint google extension we're able to see the different events that categorized freely typed values or suggestion

How in-depth have you tested your site? Is it possible that the wrong info is being sent to analytics, causing a mismatch between what is available as a suggestion and the information collected? > we don't have this info 

How many examples of mis-matching do you have? Is it a lot, or a small subset? Could this be the result of an intermittent bug that is hard to replicate? > we should take all markets to see the amount of subset - so we should do a different separate analysis to evaluate the impact

See if you can manually test and replicate a mis-match? Try being an "advanced user" and try to change the URL params on the search results (i.e. do a Suggestion Search, then change the values for the "search keywords", in theory, there is a param there that indicates its a suggestion, and if the Keyword is changed then it could track as a mismatch.

If you still cannot replicate the issue, then maybe its an intermittent bug... Does your QA team have a set up for bulk automated testing? Could they create a script to test in bulk and see if they can catch a failure? > we'll report this issue. for us, as analytics team, the issue is that we base our adobe analytics analysis on event tagged on the website and, consequently, visible in observepoint. that's why we are not able to understand why some suggestions that currently doesn't exist on website are however "tracked" with values by adobe analytics. so, finally, we want to understand if this issue can be related to website bug, since the event we are using to track this user behaviour on the website seems to be correct 

[Jennifer_Dungan]

Ok, since you talk about every team as "not you" I really don't understand where you fit into this or where you are seeing wrong data... You really need to talk to your internal teams, show them the issue and get insights from them what is potentially wrong.

There are so many places where things could be mis-matching... it could even be an issue in your Observe Point setup extracting data to check it incorrectly. Since I cannot see the behaviour I can only offer suggestions on how to track down issues.and things to try out...