Adwords Product Listing Ads & Google Analytics mis-reporting

[ewanr]

I hope you're sitting comfortably, this could be a long one and loaded with questions!

Cut to the chase:  Why is traffic from google product ads showing as 'organic' traffic in GA?

Here's the scenario:

Google Shopping

I have thousands of products in a feed to google shopping (froogle, google base, google merchant, whatever you like to call it, I'll settle for google shopping for the purpose of this question).

The URLs of this feed is tagged with GA tracking data (notably utm_source=GoogleBase&utm_medium=Product-Search), I have also tagged this with internal tracking which comes through in the back-end to assign orders to that specific source.  In this case 'GOOGLEBASE'.

Adwords Product Listing Ads

As you know, a new (ish) feature of adwords pulls in your products from google shopping so that you get a richer ad (image, title, price) and displays this in the 'advert section' of the SERP.

Once setting up a few of these, I noticed I was getting a fair amount of traffic for these new ads, taking one example, which resided in a relatively specific ad group (advertising Aviation Snips).

Naturally, I wanted to find out which keywords were driving that traffic in order to improve the ads, or kill them if they weren't working.

What was interesting is that I can't find anything about that traffic anywhere in adwords or google analytics.  254 clicks to 'aviation snips' must show up somewhere in analytics, if not the keywords, then what about the product? Analytics is showing nothing like that quantity of visits to those product landing pages where you'd expect.  It's like ghost traffic.

Google Analytics

Since experimenting with product listing ads the organic traffic in GA has suddenly shot up, looking at the new keywords they are all queries which when I test them show up product listing ads in the SERP so it's obviously the paid listing ads driving this traffic.

Why is google reporting these as organic, rather than paid?

I also noticed a keyword appear as * in the PAID segment of analytics.  I thought this was my missing aviation snips traffic, but digging into the landing pages for the * keyword, they are many different ones.

There's a connection between the * and product listing ads, but what is it? Is the traffic being doubly reported?

Back End

Meanwhile we've seen an increase for orders tagged in the back-end of GOOGLEBASE which makes sense - google are pulling in my google shopping feed into the paid part of the SERPs and these are generating sales.

Here are some of my initial thoughts / theories:

1. When google pulls in google shopping results into the organic part of the SERP, these get reported as ORGANIC in google analytics, even if you've tagged them otherwise.  It seems they strip the tags out.  This makes it very difficult to know if your google shopping feed is working well, or if you are doing well on standard organic traffic.

2. Google isn't separating out traffic as PAID with their new product listing ads, completely skewing the reports.  It makes it look like you've gained great natural organic listings when if fact you are paying.

3. With relation to the missing Aviation Snips data - maybe google is showing a huge variety of products for that adgroup (even though it's specific) and therefore I can't see the traffic to the specific products that you'd expect.  This I'm most confused about and wondered if I've missed a trick in setting the product listing ads up?

I've attached a couple of screenshots which I hope will help clarify some of this.  I can see product listing ads being great if you could get proper data to analyse and improve them.

So here are my questions again if anyone can help?

How do I see which keywords are driving the product listing ads?

How do I see the landing pages for the product listings ads?

What is the * keyword coming through in GA?

How can you get GA to report product listing ads as paid rather than organic?

Thank you so much.  If I can gather enough data on this all and work it out I'll try to write up in a blog post to help others.

0rOMM.png GUAE0.png fWPL7.png

[ewanr]

Mike & crew, I think I've got to the bottom of it. (Mike - adding the utm_campaign was useful - thank you, but wasn't the cause of the problem)

Short answer: 301 redirects were breaking the tracking.

On the site in question we have products, parent products & child products (basically distinct SKU products but held under the parent - different sizing, or colours for example).

The site was still sending the original URL of the child product to google shopping (Rather than adopting the parent URL), the child URLs have a 301 redirect to the parent URLs.

So here's what was happening:

1. Someone clicks on the google product listing ad which contains within it the URL of the product.  Google appends it's tracking info on that url (the &gclid=), including the tracking we've set so, utm_campaign, source etc.

2. That URL was being 301 redirected to the parent URL (so the customer eds up where they should!), but this in turn stripped the tracking from the URL.

3.  Since there was no tracking after the 301 redirect, google couldn't see where it originated from.

Obviously it makes more sense to fix the root cause, rather than try to redirect the tracking info too (although I'm sure that could be done).  We're changing the setup so that child URLs adopt the parent URL.

Hope this is clear.  Thanks for all the help everyone.

Kind Regards

Ewan

[mikecp]

Ewan,

Very interesting indeed. Because this is a public question, I think it could end up being tremendously helpful if you'd post any further findings. I don't want to take up too much of your time, so no need for a novel.

I suppose it's possible that having malformed tracking parameters on the (organic) Google Product URLs (missing utm_campaign) could be breaking things on the paid side as well.

Mike

[ewanr]

Hi Mike

Thanks for taking the time to check in on this one.  Since I wrote the question, I've done a lot of digging and am actually in touch with google about it.  I've managed to answer some of the questions, but not all of them! (I'll go on to explain).

It turns out I'm not including campaign, something I'm going to get sorted now you've mentioned it <queue kick="" developers="" kindly="">.  Thank you!</queue>

Regarding the remaining issues:

Adwords Product Listing Ads

• I can now see the keywords that have triggered these under the keyword report (you actually have to add a keyword ie [thisisakeyword] in order to see it.  If you don't have one, google doesn't show the keyword report menu!

"I also noticed a keyword appear as * in the PAID segment of analytics.  I thought this was my missing aviation snips traffic, but digging into the landing pages for the * keyword, they are many different ones."

• So these are definitely the keywords that trigger the product listing ads.  If you look at the landing pages that are behind the * keywords you can see all the different products that have been targetted.

HOWEVER, we're definitely seeing a strange correlation whereby we're showing organic traffic which is almost certainly paid (from the product listing ads).  The terms are for highly competitive ones where we have no chance (yet) of ranking for, but when I do a test search our ad comes up in the top right.

What I have noticed, and this may be related is that we're seeing some traffic in our live stats with  <url>.html&gclid=<string>, this actually breaks our URLs and results in a 404, usually its <url>.html?gclid=</url></string></url>

Not sure if any of that is clear, it's been fascinating digging in, if you are interested I could tell you more!

Thanks for the interest.

Kind Regards

Ewan

[mikecp]

Let me ask one simple question before looking further:

When you're tagging your Google Product Feed URLs, are you including source, medium, AND campaign. In the original post, you only mention medium and source. Campaign is also required, and it may just be that not including campaign is causing the whole thing to get thrown out.

Secondly, AdWords product listing ads, as far as I know (and have seen), should ALWAYS report as paid, rather than organic. Assuming you've got auto-tagging set up (in the AdWords settings), I'm not sure how this could be taking place.

Let's start there and work from there.

[ewanr]

EGOL - thanks very much for the response.  I agree with you that there should always be a feasibility argument before getting too wrapped up in specific data and worried about getting everything to add up, however in this instance there was definitely a need.

We weren't talking a small percentage of paid traffic that I was effectively flying blind on, we were looking at a significant percentage of spend.  In this case it was prudent to dig a little deeper!

Luckily, I've managed to find where I was going wrong with the product listings ads.  I mistakenly thought that adding a product listing ad to a specific ad group without filtering it would give me targeted ads.  In this instance, the product listing ad for the 'aviation snips' ad group I mentioned wasn't just showing ads for 'aviation snips', it was showing ads for everything in my product feed (all 13,000 of them!) this means I hit the nail on the head with my 3rd assumption above.

I've just found a great resource which explains the set up in better detail http://www.hmtweb.com/imd/?p=880 (notably filters!).  I have to say google (who only just launched this in the UK) make it too simple to simply go ahead and create these ads for every ad group without explaining the filtering.

http://www.hmtweb.com/imd/?p=880

I'm making the necessary changes and will see if this impacts or makes clear the keyword data in analytics over the next few days, at the moment the other questions / problems still stand!

[EGOL]

oh... you are a numbers person.  So am I - at least I am when it is practical - and sometimes even when it is not.

I have spent lots of time trying to solve some of these same types of questions.  Unfortunately, we either do not have the incoming signals, the tools to capture them, the skills to create our own, or the time to spend acquiring them.

Bottom line.... I have discovered that it is often better to accept schloppy data and use it to make fuzzy conclusions than it is to spend a lot of time trying to solve the problem.  If you make assumptions that go against your ROI calculations then you will at least know for sure that you are doing the right thing when red ink turns black.

Since you are a numbers person I explain it in your language....  the cost of the time that you are spending trying to solve this will go into the denominator of your profit calculation.  If you redirect that time into activities that hit the numerator such as working on your conversion rate or work that will acquire more SEO traffic your profit number will be a lot higher.

Sometimes we must be able to work with uncertainty.

Now... if anybody knows easy ways to get these types of data, I am ready to hear it.