Upcoming LunaMetrics Seminars
Seattle, Nov 3-7 New York City, Nov 17-21 Washington DC, Dec 1-5 Los Angeles - Anaheim, Dec 8-12

Tracking Google Product Search Traffic

If you’re an ecommerce site, chances are that product comparison sites are an important source of traffic for you.

Recently I had a customer ask about Google Product Search, and how to distinguish traffic from there from regular old Google organic searches.

Campaign tagging

One option is using campaign-tagged URLs. If your product data shows up in Google Product Search, it’s probably because you’re providing the data through Google Base. You could add campaign tags to the URLs back to your site that you provide. You could make the Source be “google base” instead of just “google”, for example.

But unless you have an easy, automated way to do that, it might require changing lots of URLs.

Easier: Use a filter

(***Using filters as a form of tracking is no longer a feasible way of measuring search traffic. Google changed the beginning of the URLs so they are the same on both regular search results and within Google Product Search, so a filter no longer works to bring you actionable data. However, campaign tagging is still an effective means of gathering data.)

It turns out, it’s pretty easy to use a filter to distinguish regular Google searches from Google Product Searches. If you do a search on Google, you’ll see that your results page has a URL that looks like this:

http://www.google.com/search?a-bunch-of-other-stuff 

It starts with “/search” after the “google.com” part. But on Google Product Search, it looks like this:

http://www.google.com/products?a-bunch-of-other-stuff 

It starts with “/products” instead. So we can use a filter that using the “Referrer” field (that’s web-geek speak for “where the visitor came from last”) to distinguish between these and change the Source field in your analytics data.

Here’s a screenshot of what it looks like:

google product search
The first part looks for a Referrer like the Google Product Search URL above. The second part limits it to only organic searches (since AdWords ads appear on the Google Product Search pages too, and we don’t want to screw up the attribution for those). The third part changes the Source to “google base” instead of just “google”.

Here’s a sample of what you’ll see in the All Traffic Sources report:

traffic-sources

Now traffic is listed separately as “google / organic” and “google base / organic” so you can distinguish plain old search from product search. And if you drill down, you can still see the keywords for each of those individually.

Remember that filters only apply from the time you create them going forward; this doesn’t reprocess any of your existing data.

Jonathan Weber

About Jonathan Weber

Jonathan Weber is the Data Evangelist at LunaMetrics. He spreads the principles of analytics through our training seminars all over the East coast. The next seminar he'll be leading will be a Google Analytics training in Boston. Before he caught the analytics bug, he worked in information architecture. He holds a Master’s degree from the University of Pittsburgh School of Information Sciences. Jonathan’s breadth of knowledge – from statistics to analysis to library science – is somewhat overwhelming.

http://www.lunametrics.com/blog/2009/07/16/tracking-google-product-search-traffic/

43 Responses to “Tracking Google Product Search Traffic”

We did the same recently. The big problem for us is that you can’t see transactions (or goals) for the newly filtered Source/Medium. And since you’re screenshot is only the site usage tab, my guess is you have the same limitations.

Actually, I’ve had many experiences with that sort of filters (ex: true search term filter) and it’s always the same: transactions do not show up.

Can you comment on that?

PS: I was surprised when a well known GAAC recommended us the filter approach. They finally ended up recommending a custom GA script to track Google products, which we still haven’t implemented unfortunately.

Ran says:

Thanks for the filter tip! Absolutely brilliant to track traffic and user engagement (bounce rate, time on site etc).

I’ve used the first method many times, though it has one major drawback in terms of SEO. Many times customers will bookmark the product URL WITH the tagging and very quickly these pages are indexed and act as a duplicate of the original page (quick Google search for [inurl:utm_source + inurl:utm_medium + inurl:utm_campaign] will illustrate my point).

So, for the ordinary GA user, I guess the options are track conversions and goals, but risk some potential SEO issues or filter the traffic, but get less visibility on conversions and goals.

Jonathan says:

Louis — I don’t have the data in front of me to say one way or the other on whether conversions hold up under this filter, so I’m going to have to do an experiment to find out. Look for a followup post in a few days after I try this out and get some data. (Off the top of my head, though, I can’t imagine why you *wouldn’t* get conversion data just from changing the source via a filter.)

Jonathan says:

Louis — I tried this out, and I didn’t have any trouble getting conversion data (both for goals and for ecommerce transactions) for visits where the source was changed via a filter.

I don’t know your specific situation, but it’s possible you have other filters that are conflicting or some changes to your tracking code that are interfering with this.

That might be true indeed. Will remove filters one at a time & will report back when/if I find the one responsible. Thanks for the update Jonathan.

Brian says:

Great post. I’ve been looking for a way to do this for a while. I’m surprised Google Analytics doesn’t have something built in to accomplish this since it’s inaccurate to say that something that was the result of google product search is organic which is the way it shows up in GA.

My question is, is this retroactive? I just added this filter and I don’t see any results even though I know the client has been using Google Product Search.

Jonathan says:

Brian — Filters are never retroactive, unfortunately. You’ll see the data from the point you applied the filter going forward, but not for any of your past data.

Isaac says:

Great post. We already set up the tags to track google products is there a benefit to switch to the filter? What happens if I were to use the filter and keep the tags does?

flyers says:

i agree with dominic, the answer lies in the script. Although it depends on how much value this info brings to you and if its worth the time to change and upload.

Jonathan says:

Isaac — There’s probably no benefit to switching now that you’ve already done the tags. The only downside of the tags is, someone can bookmark or link to them and that might throw off your numbers a bit.

Lee Willis says:

You have to be careful about what you’re seeing here – I would disagree with the second method as a useful metric.

Most of my clients see a significant portion of their product traffic coming from blended search results (Where Google show a handful of product results in with the main SERPs). Under your second method this would be classed as organic search traffic – maybe causing your clients to stop putting effort into their Google Base listings.

What you should do is both, then you can understand the traffic is about the google base item, and whether it’s come from a standard, or product search.

Hope that helps …

Ian Miller says:

I agree with Lee, most our of clients get their clicks via the blended results rather than dedicated product searches. Implementing both allows you to see the customer journey better (and give due credit for shopping feed sales)

Kevin says:

We have been searching for a way to distinguish regular se traffic from base traffic. Filtering sounds interesting.

Internet Security says:

A quick search on google will also list some other filters Compression filter

Flood Lights says:

This is a great info.Thanks

BJ says:

Louis – Were you able to get the converstion data to start showing? Either we haven’t had any conversions, or I’m having the same problem getting conversion data to show.

Mark West says:

Great post and filter tip! This is clearly valuable information.

Kath Dawson says:

Great post – it was exactly what I needed although I didn’t realise that you could do this with an advanced filter. I’ve applied it and its started splitting the data into a new “google base” group.

A question though, because the site largely tragets a UK audience and is based in the UK in the refferal field I entered google\.co.uk/product. This seemed logical but should I have used .com or does it not make a difference. In the UK our google search results default to google.co.uk even if you first go to google.com to enter your search term.

Thanks,
Kath

Adam says:

I’m also having the no transaction issue using this filter, but I can’t seem to find an update on this? Any news on what would be causing this?

Cory Howell says:

This filter looks great, but how are you tracking the inline Google Products results that don’t reside in the /products filepath?

Sample query:
http://www.google.com/search?q=big+%26+tall+sweater

See the 3-4 product results listed? The clickable URLs do not match the filter, nor do they look unique enough (quick look) to differentiate from the regular organic results.

Thoughts?

Francis Shovlin says:

Great question, Cory. I would be interested in seeing a filter to track that data.

Also, not sure if this was already asked, but wouldn’t the filter need to be “google\.com/products” instead of “google\.com/product”?

Jonathan says:

Folks with conversion/transaction problems: I tested and had no trouble seeing conversions and transactions. Likely you either (1) haven’t had any from visits from product search or (2) have another filter that’s interfering.

Cory – the URLs for inline product search results are really no different from regular organic search. There’s no way to differentiate these from the regular results except by using campaign tags in the URLs you provide in your product feed (as discussed above).

Francis – because of the way regular expressions work, you don’t have to match an entire string for the match to occur. Even “google\.com/p” would match these URLs (although it wouldn’t work very well to differentiate any other google.com URLs starting with a “p”). The general rule of thumb with regular expressions is, be as specific as you need to be.

Adam says:

Thanks Jonathan, but I’m still not seeing any conversion/transaction info, and I’m running only two filters – an IP exclude and the google base filter. I’m assuming the IP filter wouldn’t be an issue, is that correct? I did a segment for google base traffic, and I can see in the content report that a lot of visitors have in fact reached the submit order page (that’s a transaction). I’m baffled. The site is http://www.mydigitaldiscount.com if anyone can see something that I’m missing and wants to offer a fresh set of eyeballs. Thanks.

Matt says:

Hi there,
Can you confirm that this filter, as well as filtering off the Google products into it’s own group, it also removes these results from the original group of Google organic, so that no longer includes Google products (with the exception of inline product results).?
Thanks
Matt

Tom V says:

I came across this blog today when researching the GA Filter for Google Base. I’ve had it in place for several weeks now and, like a few others, I don’t see any transactions or dollar amounts attributed to Google Base although it is accounting for over 8% of the traffic to the site. So, I had a thought; create an advanced segment for Google Base. I did, and when I apply the segment it now shows the transactions I suspected would be there. It still makes me wonder why I have to use a segment to see them and they don’t show up under the ecommerce tab like the other engines.

I’m curious to see if anyone ‘solves’ this.

Tom

I’ve was wondering how this would apply for Google.co.uk and so used rubular.com/ to test and find out. This confirmed the following expression was required: google\.(com|co\.uk)\/product

Phil says:

Could I suggest the an improvemtnt to the RegEx…

https?://(www|froogle)\.google\..{2,8}?/(|m/)products

OR https?://(www|froogle)\.google\..{2,8}?/((|m/)products|.+[?&]oi=product_result_group)

This also catches mobile searches for google.com/m/products and oi=product_result_group results.

Thanks

Phil.

Jonathan Weber Jonathan says:

Phil – nice!

will says:

Thanks for the post and for all comments.
I also had the same problem with tracking transactions, I followed someone live coming from froogle and his order # showed up as Google/Organic
Not sure how to create a segment to fix this problem, your additional advise is appreciated

Could I just repeat my earlier query Jonathan,
When these referrals from Product Search are filtered off into their own source, does this mean that they are removed from the original source Google Organic So that now represents only Google Organic results (plus the inline products) or are they still represented in the numbers for Google Organic. Your advice is appreciated. Matt

Sheesh, my learning curve for SEO is straight up. I don’t think I’m ever gonna learn it all. OK, guess I’ll spend the next half hour studying all that I see here. Thanks, it hurts but it is real valuable:)

johncarney says:

have your own facebook theme here free :http://smal.ly/facebookthemes

Nadia says:

(Same question as Matt above)

In your screenshot, you have 1326 visits from google/organic and 86 from google base/organic – are the google base vists included in the google/organic visits? I am thinking the answer is “no”, but your insight would be appreciated!

Any word from Google if they are going to make tracking of their various products (image, products, instant) more seamless? This seems fairly cumbersome…

San says:

I’ve installed the GA filtering, but my numbers don’t match up between the two.

Our Performance Report (in Google Product Search) is displaying one number that represents our clicks (not the weekly report, but the graph at the top right). However I’ve applied the applicable filter in Google Analytics but it’s showing significantly less clicks. Why is there such a big discrepancy? Thanks!

Jonathan Weber Jonathan says:

San — note that this doesn’t capture clicks in “universal” results where products are mixed into other search results on Google. Also, this was written a long time ago in web-time, so I’m not really sure if the URLs have changed at this point and the filter is still correct; I’d have to experiment to be sure.

Joe says:

This is a great tip! Thanks for sharing!

Mia says:

I’m also having trouble tracking conversions and tracking using this filter – and I’ve had plenty of traffic coming from product search. I also only have IP filters that shouldn’t conflict with the google base filter. It seems the only way to resolve this is to set up advanced segments @Tom V or anyone else – How do you set up advanced segments to resolve this problem?

digginchina says:

i think it can’t work well now, you need change the regular expression to tbs=shop in field a. if google change its urls, you need follow it….

leanne says:

does the referral field need to have the slash before the .com? “google\.com”. sorry if this is laughable but I am a beginner!

@leanne – yes, you need the \ before the .com. Otherwise it treats the period as a special character that will match any one character (the fields use regular expressions).

Here’s our eBook on Regular Expressions – it may help:
http://www.lunametrics.com/regex-book/Regular-Expressions-Google-Analytics.pdf

Carlos says:

Mia, you can create an advanced segment with the following settings: Include Source Containing google base

Now your analytics will only show Google Base stats. From there, go to Ecommerce to view its transactions.

This was such a really great post.