Under reporting visitors and sales
My optimizely reports show significantly less visitors and transactions than Analytics and my own sales information.
Yesterday Optimizely 238 visitors, to 401 unique users on Analytics. I understand they track them differently but the difference is large.
More concerning is the revenue information Yesterday Optimizely showed 2 transactions, to the actual rate being 9.
I read the trouble shooting guide, and tested the Optimizely data throughout the site and it seems to be working.
The Query string paramaters on the order confirmation page are as follows:
What could I check or change to make it more accurate?
It's tough to know the reason when we can't see all the analytics data, but there are several things than can occur. In my experience, if testing the event has fired correctly with the transaction confirmation - which it looks like you've confirmed using the network tab described in this article, here - then it's likely to be data aligning issue, perhaps with the analytics report. It can also be a timing issue in the setup of the code on your livesite, which means some of the data may pass (if it's setup incorrectly).
It often helps to avoid discrepency if you integrate with your analytics platform - see our article on the different types of analytics integrations we support, here. This will also help make sure you've got the correct report setup, i.e. whether you're comparing the correct set of pages (based on on your targeting conditions ), etc.
There has been known to be some delay with not only results coming into Optimizely's results page, but also Optimizely data populating the Analytics report, so it's important to wait a reasonable amount of time. This can take minutes but in some cases can also take hours (depending on a number of things such as Optimizely / your analytics server traffic, or internet connection, etc.), so please be patient with the incoming data!
It's also a good idea to check your results page too, and align this with your analytics data. Pick a shorter date range and see if the results match. Another idea is to redraw your graph for a different metric (from the dropdown menu in the middle of the goal), i.e. visitors / revenue / revenue over time. This can give you a great idea of what conversions were happening when, and for instance can help if there were large chunks of time where data wasn't being captured (which can suggest a change on your website affecting the code, etc.).
But once again it's hard to say, particularly as your test confirms the transaction working correctly, which you've seen from the network call.
Let us know if you're able to provide any more information!
I was running the experiment on the homepage, but using visitors/conversions on the entire website. Not all visitors/buyers visited the homepage.