This is a great question! As background, when a visitor is included in an experiment on your site, this visitor is cookied with a unique user ID, the experiment ID, and the relevant variation ID. This cookie lasts 10 years and it ensures consistency -- the same visitor will always see the same variation.
- We de-duplicate all goal conversions (except revenue) based on the unique user ID mentioned above. So, if the same user clicked the same button 5 times over the course of 5 days, this would actually only count as 1 conversion in Optimizely.
- The cookies mentioned above allow Optimizely to track conversions outside of a specific session. If a visitor comes to your site and is included in an experiment but does not convert in that session, Optimizely will still count the conversion for the experiment & variation if it occurs at a later point in time.
Does this help clear things up for you? Let me know if you have any additional questions or if you'd like me to provide more detail on any of the above.
Could you please give me a little more detail on this.. I understand that if "User 1" clicks on a link on my conversion page to "Display contact number" for instance, and then comes back the following day and clicks on the same conversion event on a different page, then the conversion will only be counted once by Optimizely.
However, if I have 2 different conversion events on the page, for instance "Display contact number" and "Click to email", and "User 1" clicks on both of those conversion opportunities in the same session, will 2 conversions be counted in Optimizely, one for each conversion type for the same user?
We are not certain to what degree we may be double-counting conversions for the same users.
Hey @timgregory - We de-duplicate the conversion events for the *same* goal based on the unique user ID. So, if you had two *different* goals, and one user converted on both, you would see 1 conversion event for *each* goal.
So, in your example:
You have 1 goal called "Display contact number" and 1 goal called "click to email".
User 1 comes to your site and clicks "Display contact number". This will trigger 1 conversion event for the "Display contact number" goal on your results page. User 1 then clicks "click to email". This will trigger 1 conversion event for the "click to email" goal on your results page. You now have 1 visitor and 2 conversion events.
Now say User 1 comes back the next day and converts on both goals again. No new conversions will be reported since it's a duplicate event and user as before.
Does this help clear things up?
thanks for the detailed explanation. You answer does make sense, but sometimes we would like to be able to handle conversion goals differently, and I wonder if there is a way to manage the following situations:
1. In some cases we would like to be able to group different goals together and only count the first conversion in a group. Referring to the example above, "Display contact number" and "Click to email" are both lead conversions, just different mechanisms of creating a lead. When testing, we track clicks on each of these, and I also track a conversion called "All Leads" which is a click on any of the lead conversion links. My understanding, based on your explanation above, is that I will track 1 Visitor, and 2 Conversions when looking at the individual conversion figures, but if a user clicks on both types, I will get only 1 Visitor, 1 Conversion on the "All Leads" conversion.
2. Is it possible to allow 1 Visitor to count multiple conversions of the same conversion type, if they occur on different URLs? Typically, our users convert multiple times as they send enquiries regarding different properties. This is not critical to us as the de-duped goal conversion is a good proxy for this behaviour anyway.
thanks for your assistance
1. Unfortunately you can't "group" individual goals together. But what you could do is create one goal that tracks a click on any number of buttons. You are already doing this with your "All leads" goal and your understnading is correct. If a user clicks on both types, you'll see a conversion for each of the individual goals and only 1 conversion for the "All leads" goal.
2. If you are tracking the same element on several pages, the goal de-duplication still applies. A way around this would be to create goals for each of the pages, but I'm not sure that this would effectively let you track the effectiveness of your experiement. You could also integrate with Google Analytics or other platform so you can see the data (segemented by experiment & variation) there. GA uses session based tracking, so this could be an option.
Let me know if I can provide more details!