True, if you dive deeper into multi-channel reports, it might lead to inaccurate data and bad business decisions. So it’s really in the way you use and interpret this report. And best of all is to work without sampling of course.

And there are other solutions to measure multichannel behaviour without having the effect of sampling.

I do agree that managing spend should not be done in Google Analytics. At a minimum in conjunction with other tools.

Thanks for the detailed response. The sampling issue presents a problem because even though it might be accurate – you are getting information that is generally flawed (sometimes more flawed than normal). Even if the error is small overall, on a per dimension basis, this small error can be game changing.

Optimizing on trends – that’s not an ideal way to manage spend. (Since guess is not ideal and advertising channels trend over time as well which compounds the problem) What I mean here is that google analytics is not the tool you should be using to manage spend at all.

Thanks for your comment! The message in “yellow” here is directly related to multi-channel reporting whenever a filter is applied.

It’s good to interpret it in context. If you only exclude one or a few IP addresses, the data will still be very accurate in the corresponding view. If you have a lot of filters applied (also on traffic sources), it might skew your data.

In general I can say use Google Analytics to optimize on trends (large buckets of data and differences between segments) and use your back-office / other tools for the most accurate and absolute data.

So yes, pulling analytics data out and combine it with other data sources can be worthwhile to do. It all depends on the goals that you have.

*It’s not that I don’t trust google – I do trust them – that they will maximize their shareholder value ;)

Thank you for your comment Wyatt. I totally get your point here.

If you want to measure on revenue stats, it definitely makes sense to use e-commerce tracking.

And like you say, goals are only measured once per session so you won’t be able to measure across sessions or multiple clicks per session.

Another thing to consider would be to add a set value (same value as set CPC). to your click event (opt_value). You could built a user based event tracking report and calculate revenue stats per user.

However, I think e-commerce tracking provides you with the most accurate information here.

You could even create an automated report in Google docs/sheets (or another tool) via the Google Analytics API based on e-commerce tracking. It will provide you with interesting insights much faster than if you analyze the data in Google Analytics itself.

Hope this helps!

I was hoping you could weigh in on my situation.

I have a website that reviews consumer products (currently have ~1k products), and on each page has ~20 affiliate links to stores where it can be purchased.

We receive a set CPC for each click. We are currently tracking them using Events as Goals, but we cannot easily see:
– Lifetime value (multiple clicks per session, or over user life, what other products a user clicks on)
– Which products are most popular

We were thinking of setting up e-commerce tracking since this is essentially e-commerce data. Does this make sense?