I keep getting the following error message when I set the date range to a year(actually 10 months).
"One of the segments or the search in this visualization contains a text search that returned too many results. Please narrow the text search criteria and try again."
As I understand, this means that the text search criteria in the segments are too broad and therefore, there are too much data that Adobe Analytics has to handle. I've optimized the conditions set in each segment and tried to use "starts with" rather than "contains". However, the error still exists and does not give the result I want.
(As much as I want to, I cannot share how the segments look like due to brand's confidential info.)
Are there any solutions to fix this problem?
Thanks in advance.
If I understand the use case correctly, the standard best practice would be to classify the dimension you are using in your segment and then in the segment create the rules based on the classification.
Let's say my dimension is capturing page url with full paths, and I am checking for the presence of query string parameters in that dimension with my segment. Example query string parameters are 'cid', 'vid', 'gtm_param'.
I could then create a classification for every url containing cid to be classified as 'has campaign id'
Now in my segment I only have to create a rule where my classification is equal to 'has campaign id' and the segment only has to make one equals check, but I still include all my urls that have been classified as 'has campaign id'
I hope that example makes sense, but the concept is also documented here: https://experienceleague.adobe.com/docs/analytics/analyze/analysis-workspace/workspace-faq/optimizin...
"Classifications can also help to consolidate many values into concise groups from which you can then create segments. Segmentation on classification groups provides performance benefits over segments that contain many OR statements or “contains” criteria."
@hyejik65885166 - Your understanding is correct, it means that there are too many hash values returned for the condition specified in the segment. With time, the data in the suite increases and there comes a time when the conditions in the segment starts returning too many data beyond internal processing limits. Usually conditions like "contains any of", "starts with" that translates to "LIKE" internally results in such a behavior that is working per design.
As @Garretth mentioned, the best possible way to get around such issues is to use Classifications - classify similar dimension values as something and then use segment based on that classification - it'll be much faster in terms of processing.