Feb, 4th, 2010
Do you know how well your online marketing tactics are performing?
Most Marketing Managers would answer yes to this question. For example: “Our PPC campaign costs us $6 for every sale”. For the majority of organisations, the data would come from Google Analytics or another web analytics platform.
But what does “$6 cost per sale” really mean?
You would probably agree with me that the data provided from any of the major web analytics platforms (including Google Analytics) can actually be quite misleading. Especially if data is used from analytics software that is installed using the “default” parameters, without correct configuration and expert interpretation.
Interpreting Analytics Data
Let me give you an example: Let’s assume your website sells Sony TVs.
A searcher on Google searches for a Sony TV and finds your Google Adwords PPC ad. They click on the ad (and if your PPC account is set up correctly) they land on your website and are presented with a range of relevant Sony TVs. They find the “SONY Bravia KDL40X4500 model”, they like the look of it and decide they want to buy it. Let’s assume they do this on a Monday.
But before they purchase, let’s assume the consumer wants to do a bit more research the next day on other websites. In the end they find that the product is cheaper on your site and that your delivery and returns terms are more favourable. So they decide to buy the product from you.
However, this particular consumer didn’t bookmark the product page – and can’t remember what your web address is! But they do remember what your company is called. So they type in your company’s brand name into Google, and click the top natural listing. They navigate around your site, find the TV that they liked, and finally, purchase it.
Here’s how that process would look:
The default setup on most analytics packages would register the sale under the branded keyword clicked in the natural results. Is this fair? What about the paid search click that happened first?
By comparison, the conversion tracking option within Google Adwords would attribute that same sale to the click on the paid search ad. The data within Google Adwords and Google Analytics would fail to match up and there is a strong possibility that sales would be “over-reported” using Adwords data only. The reason behind this is due to the differences between first-click and last-click conversion attribution.
Google Adwords Conversion Tracking
Google Adwords employs first-click attribution. If Google sends a visitor to the website because the visitor clicked on that paid search ad, Google Adwords would always attribute any sale to that initial click. The cookie is not over-written and even if Google Adwords was further down the buying cycle, it would always attribute a sale to the click as long as the sale happens within 30 days of this click.
Google Adwords employs post-click tracking, which is initiated when the user clicks on a paid ad that has been assigned a tracking URL. From that moment, a cookie is stored on the user’s machine and if they visit the website’s complete page within 30 days, the Google tracking pixel meets with the cookie and reports back to Google Adwords which click brought the sale, and on what date.
Google Analytics Conversion Tracking
Google Analytics (and the majority of the other analytics packages), as a default, employ last-click attribution. This means that the cookie will always overwrite itself based upon the last referrer to the website. This means that no-matter what medium referred the latest visit to the website, the sale would always be attributed to that medium (other than direct – Google Analytics doesn’t overwrite direct traffic). This means that the website that initially referred the visitor gets no credit for the initial introduction.
This is where first-click and multi-click attribution comes in.
For a Google Analytics installation, First Rate has produced a product called “Acquisition by Referrer”. This product changes the way data appears in Google Analytics and attributes the sale to the first click the visitor made (much in the way that Google Adwords would record it). In the above case, Google Analytics would record the sale as a coming from the PPC advert, rather than a branded organic search phrase. (Note: The script records both first-click and last-click – it doesn’t stop the default nature of GA recording last-click.)
Another interesting thing to note is the cookie length set by Google Analytics. The cookie that stores where the visitor came from, what link in an email was clicked on (it is possible to tag the links to tell Google Analytics about the email campaign and each individual link), what keyword was used if they came from a search engine and also other data such as screen resolution and browser language used when the website was accessed – all of that is stored in the Google Analytics UTMZ cookie.
The UTMZ has a 6 month life-span. This is adequate for most websites, but for some websites it may be necessary for this period to be longer, or potentially shorter. First Rate can recommend a suitable cookie length to suit your company’s business model (eg. time it takes for customers to make a purchase decision or a cookie length that aligns with your own understanding of your customer’s loyalty/lifecycle).
In addition, because of the overload of data, the results will become illegible within the Google Analytics interface. The best way to make sense of the data would be to export the data into excel and then use pivot tables to analyse the data.
A reason why multi-click attribution is becoming more popular is due to the ‘weighting’ of visits to the website. How much weight should the first click have, then a potential second, and what about a third or fourth?
Of course we can go another step further – what if you have a traditional ‘bricks and mortar’ store. The TV example above is a perfect example – a visitor goes online, researches the product that has the required specification, and then goes into a store to look at it, or even a competitor’s store! The salesman tells them about the TV, gives them a demonstration that could never happen online, and then the shopper returns home, goes on the net and buys from your website. How much emphasis should be placed on the in-store visit, but more importantly – how can we measure it? How can an analytics package possibly tell that the customer went in-store? At the moment, we simply can’t.
Website Analytics Recommendations
First Rate suggests initially using first-click attribution to see how the data is affected to get a ‘clearer’ picture as to how each marketing is acquiring traffic and sales (and at what cost!)
As a Google accredited partner (GAAC), First Rate can certainly help with custom Google Analytics implementations, please do contact us if we can assist in any way.