Discrepancy between Facebook orders and Adobe orders

[anuragxadobe]

Hi Everyboby, 

Our client have an ecommerse site running paid campaigns on Facebook whose numbers do not match with what we are getting on Adobe Analytics. We got 185 click through purchases on Fb Ad Manager for last month while in Adobe we got only 71 orders from Facebook for same period of time. 

I'm aware that if some visits from facebook and then again visits the site within 7days from paid search the order will get attributed to Paid search. But still it is a very large discrepancy and the client is very concerned about it. 

Is there any way to track all the orders where facebook has assisted?

(I have also tried using the order at participation attribution but it only shows orders which eventually purchased from facebook only.)

Thanks in Advance

[Jennifer_Dungan]

I believe that Facebook pixels have a default 30 day attribution... (it's pretty greedy when it comes to trying to attribute success).

Our Marketing Manager that handles those pixels changed ours to be 7 days to get it to a more reasonable attribution...  

Also, most people report the same.. FB numbers seem high... I'm not sure I trust FB to be honest....

[TomTheAnalyst]

I am battling the same questions from my "customers". Apart from attribution window differences mentioned by @Jennifer_Dungan what if you could use First Touch Channel instead of Last Touch Channel. Or at least manipulate with custom attribution models within AA?

[Jennifer_Dungan · Accepted solution]

The problem with using First Touch is that Facebook may not be your first touch... 

Facebook works by adding a user being on FB and clicking on a ad... this drops a cookie onto the user's computer (for all we know, users may cancel the page load before the site loads and records our own analytics).

If in the next 30 days (at least by the default setting), that user "converts", because FB sees that cookie on their machine, they will treat that as their conversion... not terribly realistic.

Then you also need to take into consideration your own internal traffic (which should be excluded), Bot / Spammer traffic that you should have excluded, the fact that FB is uing a different way of tracking which may not fit the same "opt out" structure, or same client level "blocking" which we all know we lose data to, etc.

Trying to mix what FB is saying with what Adobe is saying is "comparing apples and oranges".

Given how "greedy" FB is, and considering that users may cancel out of the click before they actually arrive on the site, most people will just defer to Adobe's tracking...but that is a decision that you and your business need to decide on.

Being armed with some understanding of what FB is doing will help you to understand and make that decision for yourself.

There have been other posts this year alone, all asking similar things:

https://experienceleaguecommunities.adobe.com/t5/adobe-analytics-questions/discrepancies-between-ado...

https://experienceleaguecommunities.adobe.com/t5/adobe-analytics-questions/data-discrepancies-betwee...

This is pretty much a universal problem...

[mohammadz]

Hi @anuragxadobe,

First, it is not advisable to compare the metrics across two platforms because it is never an apple to apple comparison.
Still, even if someone wishes to and try to understand the gap, the basic idea is to understand how different platforms track and calculate the same metric. 

The metric might sound/ look similar across the platforms , but it most probably will not be the same when we explore its logic across respective platforms.

Other scenarios can be  - cookie expiration is different across platforms, AA will consider other channels if they are present after FB touchpoint.

For your scenario, one thing that came to mind was to build a visitor level segment with Marketing Channel as FB PAID and "order exists"  within the next 30 days and check for Orders count for the duration in question.
Again, this is not going to be exact and in all probability will overestimate the numbers as it will count all the orders for the visitor, but might be good to get to a point that you are looking for. Here is a sample definition that you can use