I got a question from a reader here recently and after posting a comment back on it, decided that a full post would be best for everyone. Comments can get buried at times and this topic is worth taking note of.
The question involved the use of Google Analytics Funnel tracking and went something like this.
Is it possible to set up a funnel with a determined starting point and ending point, but where the pages in between don’t matter? In other words, a user comes in on Page A, then they could either go to Page B, C, or D, and then from there go to Page E. All I want to know is how many users who came in via Page A ended up at Page E. I don’t care what happened in between. Is that possible?
Taking note of the true purpose of funnel tracking will shed some light on why not caring about what happens in between the initial click and the end goal can backfire on you.
The Purpose of Setting up Funnel Tracking
One primary objective of any e-commerce site should be to drive visitors down a “specific and intended path” guiding them all the way to then end result (a sale.)
Providing that path gives users direction. The direction you provide builds customer assurance which in turn builds confidence. Both lead to more sales.
Allowing customers to “aimlessly wander” through your site will result in less than desirable results and lost sales. If you want them to buy, you need to take their hand and “show them the way”.
Funnel tracking is used to measure how effective a particular process is at leading your visitor down an intended path then converting that visitor into a goal (for e-commerce sites that ideally is a sale.)
It provides a detailed path analysis depicting where in that process you have problems (also called barriers) which prevent your visitors from buying. Funnels provide a “starting point roadmap” for addressing sections of the site that need adjustment and attention.
Leaving the “meat” out of a funnel (all the pages in between the initial click and the end goal) only defeats the purpose of setting up the funnel in the first place.
Having said that, if you really want to setup a funnel without any pages in between mattering, then it is possible (but not advisable) and, you have to be able to read and interpret the data figures to make sense of it all.
This Example Explains Further
Joe has a website. He wants to increase sales. To meet his objective and as a first step, he knows he must have an analytics package in place. He adds Google Analytics to his site and sets up funnel tracking.
His funnel consists of two pages. A starting point on step one (of 4 total steps) of the checkout process and an ending point at the sale confirmation page.
He sets up no pages in between those two start and end points — big mistake and you’ll soon see why. Joe says he wants to know “how many people enter the process” and how many people “make it to the end.” The problem is that because of the way he set this up, the data that is going to be gathered will be useless in the end and provide no real value.
Joe launches the site and it runs for a few weeks while data is gathered by the analytics program.
Upon looking at the analytic reports, Joe sees he has a definite number of visitors that arrived at step one and also sees a definite number of visitors that have ended the process (arrived at the goal.) He feels he got the numbers he was looking for — how many started the process, and how many ended, but he also sees that 100% of the visitors who started at the first step apparently abandoned the process at that point?
This statistic prompts him to ask “why did 100% of the visitors abandon the process?” Furthermore, he next asks “how can 100% abandon the process but still arrive at the end goal?”
Questions Answered? Not Exactly.
The answer to those questions is simple, the “where they went” portion is not as easy because the funnel was not setup correctly in the first place.
The reason Joe sees a 100% abandonment rate at step one is directly based on the way he setup the funnel to track visitors. The abandonment rate will always be 100% with this funnel setup. Why? Because there are no middle steps in the funnel process. So, immediately upon leaving the first step in Joe’s funnel, the tracking system reports an “exit” (increasing the abandonment rate.)
The tracking is only reporting what Joe asked it, and in fact, it is reporting accurately that 100% of the people who enter step one abandon the process at that point (as it has been reported according to Joe’s incorrect funnel setup.)
Let me explain it another way.
Every visitor who enters at step one of this process is tracked only to the point of where the tracking allows them to be accounted for. Let’s say 45% of them who enter at step one make it to the end. Of those 45%, everyone has to move from step one to step two onto step three and finally to step four (the goal) at some point.
However, because the funnel tracking Joe setup only allows them to be accounted for at step one and step four of the process, each of those visitors who actually makes it to the end seemingly abandons after step one (which is incorrect as we now know.)
Remember data is only as good as the tracking setup allows it to be. It’s the classic GIGO (Garbage In Garbage Out) principle. You put garbage into something, you can expect to get garbage out in return.
The Real Question and Answer
Sample conversation excerpt with Joe:
Joe: “100% of visitors abandon the checkout process at my store.”
Coach: “So, where in the process did they leave Joe?”
Joe: “I have no idea.”
Coach: “Where did they go when they left?”
Joe: “I have no idea.”
Coach: “If you have no idea where they left or where they went, how can you improve their experience?”
Joe: “I don’t know?”
The real question Joe should have been seeking answers to when he setup his funnel the first time around is “where do visitors who do not complete the process actually leave it? And where do they go when they leave?”
That information is something that can only be provided by a properly setup funnel. It is also this information that could guide him toward focusing efforts on the exact sections of his site that need improvement.
Had a plan been thought out, and the funnel been setup correctly the first time around Joe would have answers to questions that would have guided him toward making improvements to specific areas of his site which in turn would have addressed visitor concerns and in the end lead to more sales.
As it stands, he is left with a mound of data that tells him nothing about how and where he needs to improve his site to speak to his visitors and increase sales.
Point made, lesson learned.
Nice article. I’ve been gathering some funnel data for my 9-5 and we have enough to start looking at it. 50% abandonment on the first step, 66% on each of the next 2 steps, with only 22% making it to the end.
Luckily most of the exit goes to other site pages. I guess they weren’t ready to buy at that point (that’s a membership site).
Goals have been setup for the number of quotations requests for moves or removals filled out via the website. Is it necessary to have a funnel if is just one step? Will I get a 100% abandonment rate for this single step or will I be able to see how many leave or ‘reach the end’?
The purpose of a funnel is to track a path from beginning to end of a multi-step process (what ever you deem that process to be).
The funnel could be the pages leading up to the point where the request is complete, or it could be something else.
If you truly do have a one step process then a funnel might not be necessary. You certainly could see abandonment of sorts (for anyone that leaves at the first step and does not complete it), but 100% would not be exactly right. In fact, the abandonment would come in the form of either a “bounce” or an “exit” in your stats since you do not have a funnel setup to track the actual abandonment point.
Here is a situation where it might be nice to have “Page A” included with the funnel not caring about pages “B, C, D” and then caring again about pages “E, F, and Goal”:
Suppose Page A is a landing page from Google AdWords with URL parameters that tell me which adwords campaign sent the visitor to my site.
Now, I may not care if the visitor pokes around on the site a bit (pages B, C, D) before going to my cart and checking out (pages E, F, and Goal). But, I am curious about which AdWords campaign is creating the most goal conversions.
I know AdWords lets you track conversions (this is a hypothetical here), but hopefully it illustrates a case where you may not want a user to drop out of your funnel even though they visit some intervening pages. If you want to track something other than an AdWords campaign, then I think this situation might apply.
That said, I am not aware of the ability to actually create a funnel that allows this type of tracking.
Andy
this is really detailed explanation it would have been more awesome had you shown a picture by picture tutorial of the funnel creation thanks anyhow 🙂
Andy – For your scenario, using GA url tags might be more appropriate. That way you can tell who came in through a particular link, and you can still set a goal (purchase page for example).
Thats all ok, but a hotel webstie, for example, doesn’t have a specific path to bookings. Most hotel websites have the booking button in sight in every single page. So how would you suggest to track that? create funnels for each page?
I’m currently working with a booking engine wich only has two steps; the booking engine is loaded (“reservation_load”), you enter all the data and pay, and it takes you to a “reservation_confirm” page. Thats it.
I created a funnel wich goal is “/Reservation_Confirm”. As steps (not required) i put:
step 1 “/”
step 2 “/.+”
step 3 “Reservation_Load”.
And it doesn’t seem to work.
Any trick i’m missing?
If you switch to the asynchronous tracking code you can use the onclick event handler to trigger tracking on a per button basis. This link might help:
http://code.google.com/apis/analytics/docs/gaJS/gaJSApiEventTracking.html
Nice article. I had no idea that such tool existed, but I will be getting a new client who wants to see what’s really going on with his website so I am reading along gathering info about Google Analytics. Thanks for sharing I’ve learned a lot here!
Hubert
This really took my attention. I’m not an advanced in this web stuff, but this was really interesting. When I think about this Google Analytics Funnel properly, I know I’m just a beginner. Hope you’ll have more articles like this. Definitely bookmarking the page. Thanks