On the path to a conversion, a user may conduct multiple searches and interact with multiple ads from the same advertiser. Google Ads’ attribution models allow the advertiser to control how the “conversion credit” is split amongst the various searched keywords that participated in the conversion process.
Two weeks ago, I made a purchase that I’ve been dying to make for a long time: I bought an Apple iPod Nano! And, for the record, it wasn’t I that broke down and bought it, it was my first generation iPod Nano from 2006 that broke down, leading to my buying my new mini-toy.
Boy, do I love my new iPod Nano! It’s 8GB, which for me is more than enough. It’s smooth, slick, and user-friendly, which is a staple of most Apple products. It even has a built-in FM radio tuner, which allowed me to listen to the end of the Yankees / Rangers game last night while biking! Yeah, I’m <3’ing it. 🙂
Naturally, being in the web analytics industry, it got me thinking about my journey from start to finish in purchasing this item, and how all of it would be tracked in Google Analytics, WebTrends, or any web analytics platform. Clearly, if I were in charge of Apple’s web analytics efforts, I could easily pull up a report within a matter of seconds which would show me the exact purchase scenario and tell me exactly what I would need to do next. Right? I mean, how hard could it be for CoreMetrics or NetInsight to show me that I:
– Saw the Apple iPod Nano television commercial five times across two different cable channels;
– Went directly to Apple’s website, found the iPod Nano section, and took a long hard look at its technical specs and color options;
– Listened to my friend ramble on in glowing terms about how great the new iPod Nano was; and
– Played with the iPod Nano at the Apple Store and purchased it (but not before discovering that the Apple Store was out of the 8GB silver, which forced me to choose the 8GB blue one).
For those of you who caught on to the sarcasm in the previous paragraph, you know that we are nowhere near being able to call up such a report from our favorite web analytics vendor. In fact, the only piece of information from the four points listed above that you would be able to access would be:
– Went directly to Apple’s website, found the iPod Nano section, and took a long hard look at its technical specs and color options
The information on the other three points is not in Google Analytics, Oracle CRM or SalesForce. The data isn’t anywhere online, other than in this blog post and in my memory banks. As far as Apple is concerned, I made a single, very-long visit to their website without purchasing something via their online shopping system. On the surface, my visit to Apple’s website could be considered largely unsuccessful, since I never performed an important action, like making a purchase on their website. However, if it wasn’t for the Apple website, I would have never visited the brick and mortar Apple store in the Boca Raton Town Center Mall and purchased the iPod Nano in person in the first place!
To further elaborate upon this point, I wanted to show you what a possible report would look like in an imaginary world where I was Apple’s only website visitor, using the four points previously outlined as purchase influencing factors. Using an arbitrary value of 100 to score a purchase for an individual visitor (myself), a possible report listing the influencing factors for a purchase and their scores could look like this:
3 TV Commercials on Channel A: 0 pts.
2 TV Commercials on Channel B: 0 pts.
Website Visit: 0 pts.
Customer (Friend) Testimonial: 0 pts.
Apple Store in-store Demo: 0 pts.
Color / Model Availability: 0 pts.
Apple Store Purchase: 100 pts.
But in actuality, as far as my particular purchase is concerned, credit is spread across the board, as all of these factors had their own contributions in their own right. I would score my purchase influencers this way:
3 TV Commercials on Channel A: 10 pts.
2 TV Commercials on Channel B: 15 pts.
Website Visit: 35 pts.
Customer (Friend) Testimonial: 5 pts.
Apple Store in-store Demo: 30 pts.
Color / Model Availability: 5 pts.
So, how does Apple (and how do you) really know when a visit to a website is a successful one? How do you really know what source is deserving of the official credit for a purchase (or a conversion)? Unfortunately, there is no perfect answer or one-size-fits-all solution to this very complex problem that all marketers and web analysts face on a daily basis. In my single example alone, I am crediting six different factors as influencers into my purchase of a couple of weeks ago. Using the first scoring model, you would conclude that the website is doing a poor job of selling iPod Nanos. In actuality, in the second scoring model, I gave it 35 points – the biggest slice of my pie – which would indicate that the website is actually doing a great job toward the purchase of iPod Nanos!
My message to you today and the point of this blog post is for you to take a more critical, less cynical approach to the data that your web analytics program is spitting out at you – especially when you are measuring conversions, sales, or purchases. Not only are there going to be multiple influencing factors involved other than what your web analytics package is showing you, but the website that you are in charge of measuring and the online marketing campaign you are running can and will have an impact at your brick and mortar, physical store location or your offline advertising efforts. It did for me.
Last week, Google announced a new set of reports within the AdWords interface called Search Funnels, which are rolling out to all AdWords accounts within the next few weeks. With search funnels, advertisers will be able to obtain a truer sense of value and attribution for keywords that help in the conversion process.
Currently, AdWords assigns credit to a conversion to the first click’s keyword, ad, ad group, and campaign up to 30 days after that first click occurred. For example, suppose I perform a search on Google for the term brown shoes on March 29th. I click on the ad, view a couple of pages, but I don’t convert. Two weeks later, say, April 11th, I search for brown dockers shoes, click on another ad, and this time I convert. AdWords will assign the credit for the conversion to my original search term of brown shoes (provided the advertiser is bidding on that keyword).
Now, with the new search funnels report section, I’ll be able to see which AdWords keywords helped my original keyword convert, as well as a funnel (hence the name) of each keyword that led to a conversion in succession. This new search funnels report section is going to have nine different reports like “assisted conversions”, “last click analysis” and “top paths” to perform deeper conversion analysis than possible before.
Why it’s “the right step” toward proper attribution
Before this product launch, you had two possible options for assigning credit to a keyword for a conversion. You could log-in to AdWords and view the reports in the interface (first-click attribution), or, you could use Google Analytics to view the keywords that matched goals (last-click attribution). If you were an advertiser, you would, over time, start bidding more for the converting keywords that either AdWords or Analytics were displaying, and less for the “non-converting” keywords. As it turns out – and as the more experienced marketers long ago deduced – other keywords lend a big helping hand along the way, but never received the proper credit. What would happen is that advertisers would either shut down those keywords that didn’t appear to convert, or change their bidding philosophy to such an extent that these assisting keywords become irrelevant over time. The result: the number of conversions would struggle to climb, leaving advertisers scratching their heads.
Now, we can assign importance and value to those assisting keywords, and not automatically cast them off as losers or rejects. They are an integral part in the conversion cycle; keywords that assist in the conversion process should remain active and managed intelligently for optimal campaign success.
Why it’s only “a step” at this point
It’s not an end-all, be-all solution, but it’s a great start. Search funnels in AdWords does have a few limitations. First, you must import your Google Analytics goals into AdWords, which is not a big deal for an administrator, but still something that must be done.
As of now, search funnels can only report on AdWords keywords and web site visits. If a user in the conversion process accesses a site directly after previously clicking on an AdWords ad, that user’s direct visit cannot be tracked in search funnels. Not even natural / organic search engine queries are available in search funnels at this time. Also, while the new search funnels reports look like they belong in Google Analytics and not AdWords, they’re actually not available in Analytics yet.
So, true attribution – if there is even such a thing – is not yet within our grasp. But with search funnels, Google has taken a very large step toward that general direction. I recommend you log-in to your account today, import your goals, and discover which assisting keywords should be optimized for greater campaign success.