So if you've ever spoken with me about real-time analytics (first - sorry if I rolled my eyes), you may know my opinion. Although real-time analytics have been around since 2011/2012, I have not been a huge fan.
From something I posted a few years ago:
"Real time reporting is a waste of time.
It creates a useless distraction and an incomplete picture. We all know now that last click reporting is flawed. This is because of the concept of latent conversions, meaning not everyone who buys or converts on your site does it on their first visit. Latent conversions are particularly important as people tend to be less impulsive online. Because it is so easy to do just a little more research, people tend to be more pragmatic when shopping online.
Reporting that is based on attribution can show how big of an impact latent conversions make to your site."
I've since calmed down a bit. But still think there is a point worth making. Real-time lacks three important factors:
and, importantly, Accuracy
Real-time Lacks Context
Is this good news? Bad news? How does it compare to yesterday or last week?
Real-time Lacks History
Over the past few years, latent conversions have increased from maybe 20 - 30% of purchased to 40 - 50% (or even higher, in many cases). Latent conversions are the purchases, or other type conversion for non e-commerce sites, that take more than one visit. For example, someone visits the site from a paid Google ad on Tuesday, leaves the site without making a purchase, then returns from a Yahoo ad on Thursday, and makes a purchase. You'd definitely want to know about that Tuesday visit, and the impact of the Google ad.
Real-time data lacks this history since it is only reflective of what is happening on your site now. You can see how big of an impact latent conversions are in Google Analytics. Click to the Top Conversion Paths report (under Standard Reporting, Conversions, Multi-Channel Funnels, Assisted Conversions). This shows the impact - the percent of conversions that take more than one visit, as well as the dollar value.
Real-time Lacks Accuracy
I'll spare you the boring, technical details. Web analytics were 'Big Data' even before Big Data became a thing. There is a lot of processing, sessionizing and de-duping that must be done to the raw data to make it correct.
There is a sub-set of those things that can happen quickly. So the data that you get in real-time is quick and dirty data. You can see this if you save the data, for say New vs. Return Visitors, then compare that data tomorrow (for the previous day).
So there you have it, my beef with real-time analytics. Up next, why real-time analytics may not be so bad:)