Why would lead counts in RCE differ from within a smart list or report in Marketo?
Something else to keep in mind is that when you are pulling a smart list, you have a unique count of leads. If you are pulling a program membership report in RCE, for example, the lead count isn't unique (i.e. the same person could be counted in multiple programs). What report in RCE are you looking at?
I created a Lead Analysis Report in RCE and have a higher number than what is in a Leads by Lead Source Report inside of Marketo. However when I pull a smart list the numbers match to what is in RCE.
So I just noticed when I hide the opportunity columns in the Leads by Lead Source Report inside of Marketo the numbers now match. Why would that be?
This is usually because of the assumption that Smart List, Analytics and RCE should use the same reporting formula and calculation approach. If that was truly the case, there would actually be no need for 3 different features. In my opinion, the best and simplest terms that I could possibly use to describe this that helped me and perhaps may help you is that numbers differ because the basis of calculations for each feature is different.
Does Marketo provide documentation describing each type of report and how all calculations are made?
I don't believe I've ever seen any documentation on the differences. You can however, find all kinds of documentation everywhere that will help you understand each type of feature and it's reporting ability a little better. For example: Email Performance Report - Marketo Docs - Product Docs
This isn't to say that it's not out there, I just haven't seen it yet. Marketo is a pretty large product and I've yet to go through every article, community post, blog, product doc, etc. Good luck in your search for the truth.
For now, just know not to compare apples to oranges to bananas because you're probably going to have a bad time.
Choose one system for your major reports.
If you look in University, there are more details on RCE. If I remember correctly, Marketo will be impacted by lead deletion across MLM and RCE.
Your smart list will be real time, while RCE is yesterday.
RCE uses an external database outside of Marketo (hence why there is a 24hr delay due to an overnight ETL job). This database is a daily 'back-up' of data and contains orphan records that were deleted in Marketo or SFDC. This along with how Pentaho queries run versus Marketo explain the count differences.
The way that each report is designed in RCE is by a table (e.g. Opportunity Analysis creates a table of Opportunities that may or may not be explicitly tied to a record in Marketo; thus changing numbers) so a Lead Analysis report is a historical table of leads and contacts.
My recommendation is to always put date threshold and 'Is Deleted in SFDC' flag filters to ensure you're pulling back relevant data. It is very difficult to match RCE to Marketo but know that RCE isn't incorrect, it's just taking a more historical look at your data.