Expand my Community achievements bar.

SOLVED

Export data from analytics (with Datawarehouse) to feed a database

Avatar

Level 3

Hi,

we export daily data from Adobe Analytics using Datawarehouse tool to feed an external database.
We do that to be able to match those online information (page, marketing channel, ...) with offline data and create audiences or analyze deeper our client behavior.
We notice that sometimes we have 10% of difference between what is in the workspace and the number of records in the exported files.
Did you already experience this?

Do you have alternative to export the data.

 

Thanks

 

Robin

1 Accepted Solution

Avatar

Correct answer by
Community Advisor and Adobe Champion

Without a little more information it will be hard to answer this...

 

But first off, Data Warehouse is sort of a flattened table of data... When you have fields  (such as lists or products) that contain multiple pieces of info, that "one" row will actually present as many.

 

Example, an order with 3 products will actually present as three rows in the export (one for each product)... this also means if you are adding up your metrics, you will actually over count them...   

 

This one "Order" will show the order metric in each row, if it's tracked on a page view, it will also track the same page view in each row...   so "1" becomes "3" because it cannot be de-duplicated.

 

Even when not looking at list type items, metrics like Visits or Unique Visitors can be inflated, again due to the inability to de-duplicate across your rows...

 

 

The better solution is to use the Raw Data Feeds... this will actually provide you with the Raw Data as Adobe gets it... but it's also not for the faint of heart... because now you will have to create all the SQL to process the information like Adobe: identifying your UVs and Visits, excluding the hits that shouldn't be counted, processing all the information... and while you may not get a 100% match (since it is challenging to replicate all the logic), you should be able to get it very very close... 

 

 

 

View solution in original post

4 Replies

Avatar

Correct answer by
Community Advisor and Adobe Champion

Without a little more information it will be hard to answer this...

 

But first off, Data Warehouse is sort of a flattened table of data... When you have fields  (such as lists or products) that contain multiple pieces of info, that "one" row will actually present as many.

 

Example, an order with 3 products will actually present as three rows in the export (one for each product)... this also means if you are adding up your metrics, you will actually over count them...   

 

This one "Order" will show the order metric in each row, if it's tracked on a page view, it will also track the same page view in each row...   so "1" becomes "3" because it cannot be de-duplicated.

 

Even when not looking at list type items, metrics like Visits or Unique Visitors can be inflated, again due to the inability to de-duplicate across your rows...

 

 

The better solution is to use the Raw Data Feeds... this will actually provide you with the Raw Data as Adobe gets it... but it's also not for the faint of heart... because now you will have to create all the SQL to process the information like Adobe: identifying your UVs and Visits, excluding the hits that shouldn't be counted, processing all the information... and while you may not get a 100% match (since it is challenging to replicate all the logic), you should be able to get it very very close... 

 

 

 

Avatar

Community Advisor and Adobe Champion

I agree with @Jennifer_Dungan's last comment about data feeds. Every organization I've been at I've used (or encouraged them to use) data feeds. Normally one of the biggest drawbacks to setting them up is having a data lake to store all of the data in. But if you're doing this with data warehouse, them I'm guessing you already have somewhere for the data to live. Take the time and go through the data feed settings. It's going to get you results that are a lot closer to what you're seeing in workspace. You will have to replicate any segments/VRS settings applied to your report suite, but once you do that, you should be able to get pretty close. For us, we're at about 99.4% accuracy, which is actually really good. So you will see a bit of a discrepancy still, but it should be about 1% or less, no where near the 10% you're seeing now. 

Avatar

Level 3

Thanks all for your feedback. Let's jump into data feed so!

Avatar

Community Advisor

Check to see if in workspace you have any special attribution settings. May have to set to default to compare... 

https://experienceleague.adobe.com/en/docs/analytics/analyze/analysis-workspace/attribution/overview

 

GLTU