If the title didn’t resonate with you, and you have an e-commerce website, go do this right now: In a landing page report, look for your shopping cart. You might just see something like this:
That’s messed up. The screen shot above is from a search landing page report over a 3 day window. There can’t be that many people landing on our cart, right?
Here’s some caveats to this scenario to rule out the obvious:
- The cart is on the same domain as the main site
- No sub-domains or any other cross-domain action happening
- Everything is tracking properly (we’ve tested)
On that last point, I tested things like getting shipping quotes, applying coupon codes, adding/removing items from the cart, etc. None of these actions cause a new session.
So why is this happening? To find out, we need to dig into more specifics.
Understanding what we’re seeing
It’s important to know one thing, these are not new users. Nobody is visiting your site for the first time and landing on your cart. At least, not likely. In every instance of this problem I’ve found, the visits to the cart are actually direct, not search or referral traffic.
Here’s how you know this is the case…
View a Top Conversion Paths report with a secondary dimension of “Landing Page”. Use a filter to only look at paths where one of the sessions landed on the cart.
We see something like this:
This report tells us that in each instance, a user started by taking a traditional acquisition channel landing them on a traditional landing page. Later, they returned via direct traffic and landed on the cart.
Because standard GA reports attribute everything to the last non-direct source, our landing page reports are saying these folks came to the cart via Organic, PPC, etc.
So does that mean people are leaving the site and bookmarking or typing in the cart URL directly? Probably not! Here’s a way to get us closer to an answer to that question.
Getting closer to what’s happening
For one, we can make use of a conversion segment to find out a little more around the behavior here. Create the following segment to look at these users specifically:
Obviously change the URL to match your cart/checkout URL structure.
Once you have this segment created and applied, look at a time lag report. Mine looks like this:
Almost all of these conversions are happening in a single day. The tricky thing is, a user could start on 11/1, come back 11/5, come back a 2nd time on 11/5 (“landing” on the cart) and it would be a 5 day lag. But at the very least, people aren’t generally leaving the site for days and returning directly to the cart. That helps a bit.
So what’s the deal?
My guess? People are adding items to their cart, then doing comparison shopping or coupon hunting. This process lasts over 30 minutes (maybe they wander off to lunch or for a coffee break), and when they return to the browser with your site’s cart open, a new session starts. In that new session, the landing page is the cart and it’s direct traffic. Because GA rolls with the last non-direct source, it’s attributed to search/email/whatever.
But how can we validate this? Here’s one way…
I created a new UA property, as to not mess up ongoing data. In that property, I changed the session handling settings to the max timeout (4 hours). Here’s how you do that via the GA admin:
In the scenario I outlined, it’s reasonable to believe that the majority of those users would do their comparison shopping or coffee breaking in under four hours. So if my hunch is right, this property should report way fewer people landing on the cart.
Let’s look at the same landing page report as before in this new property and find the cart…
That’s a pretty big dropoff. And it’s always possible people snuck into this report from before I made the settings change.
Looks like we got it! If this trend holds up, we should see the correct landing pages getting attributed with goal completions and revenue. At least, the correct landing page based on the last click model, but that’s a whole other discussion.
Caveats and whatnot
The biggest thing to note is that if you don’t get a lot of landing page hits on your cart/checkout, there’s no reason to worry. Also, changing session settings can have negative impacts elsewhere, so I’d definitely recommend testing in a new property and comparing data before just diving in with a change. But if you see landing page data skewed like this often, and your site has a lot of people comparison shopping, you might want to test this out yourself.
And what say the rest of you? Surely there are other possibile explanations for this behavior. What have you found? What can we test???