Most people often perform several searches and navigate through several ads before completing a valuable action or finally making a purchase on a website. Before now, all credit for the purchase or final conversion is ascribed to the last ad or touchpoint the customer interacted with. But come to think of it, was it really the last ad that made him or her take that actionable decision?
If you are looking to know more about those new and exciting technologies that are constantly gaining prominence in the ad landscape, here is what you need to know about Google’s Data-driven Attribution summed up in a nutshell for you.
What is data-driven attribution?
Depending on how people search for your business and their interest in becoming your customer, data-driven attribution can use data from your account to assign credit for conversions and even determine which campaigns, keywords, and ads are receiving the greatest impact on your business goals. Marketers can effectively make good use of this form of attribution conversions on Google Analytics and their websites from Search Network campaigns. Like other attribution models, you can only make use of data-driven attribution for Google Analytics and website conversion actions, as well as for Google.com search ads. Before considering this form of attribution model, here is what you need to know.
Understand the existing limitations
Data-driven attribution is currently the only compatible form of attribution with shopping campaigns and Google Search. As a matter of fact, Google is yet to incorporate data from other online channels and engines like email and display. Just so you know, this model does not support the use of multiple accounts even when dealing with Search and Shopping campaigns.
For those marketers housing their Search and Shopping campaigns in more than one account, it is important to know that Google would only make use of data-driven attribution for bidding within separate accounts.
While Google is trying so hard to ensure that the model works inclusively in years to come, many believe that the overall efficacy of the model can be hampered by these limitations. Nevertheless, it is expedient for every marketer to get knowledgeable about the potential capabilities of this attribution model regardless of the limitations around it. By so doing they will understand how it works and also know how to drive improved performance for their campaigns.
Calls for smarter bidding
One good thing about this model is that it can be selected as a source for bidding and reporting AdWords once the data-driven attribution is available for any conversion type. It is from here that marketers can effectively organize their bids in accordance with their account goals by setting up automated bidding in Google’s User Interface (UI).
Basically, the main objective is to drive and maximize the volume of conversions before they even hit their cost per threshold or drive more conversions at a more efficient cost per. It is best to consider this new release as a test. If you are not maximizing the volume of conversion or driving more conversions then it is likely that the model isn’t the right solution for your needs. It is advisable to seek advice from your agency or Google reps at every step of the way particularly if you are not used to it. This will help you to understand those signals that are invariably considered by Google for bidding.
Understand how it works
Unlike other attribution models, data-driven attribution works quite differently. It calculates the actual contribution of every keyword located along the conversion path by making use of the available conversion data. As you already know, people often conduct several searches particularly when it comes to converting online. This means that they are very much likely to click on several ads on a range of devices before finally converting.
First and foremost, you must understand that each model from this attribution is specific to each advertiser. All the clicks made on your Google Search ads are considered and possibly examined by AdWords. Basically, what the model will do is to identify those patterns among the clicks that lead to conversions by comparing the click paths of prospects that convert to the paths of those who don’t. Ultimately, it is important to know that there could be some steps along the way that can offer a higher probability of leading a customer to perform a complete conversion. It is after identifying these valuable clicks that more credit can be assigned to them on the customer’s path.
In a bid to determine the actual contribution of those touchpoints across the conversion path, Google will take a look at all the different paths of conversion and uses data from converting and non-converting users to find commonalities between them. It is after this has been done that fractional credit will be assigned for a conversion based on the propensity of each touchpoint to lead to a conversion. By implication, marketers will get to know the ads that possess the greatest effect on their business goals when they’re evaluating conversion data. Also, if you are looking to drive more conversions with an automated bid strategy, it is this vital information that your bidding will utilize to help acquire more conversions.
Adequate data is required
Many marketers wonder why they don’t see the data-driven attribution in AdWords. It is mostly because they do not have adequate data. Before a precise model can be created to enhance how conversions should be attributed, a certain amount of data is required to enhance data-driven attribution. Just so you know, you will not see an option to use data-driven attribution once you have enough data.
Before Google can even begin to develop a data-driven attribution model for a brand’s ad, at least it should have garnered 15,000 clicks with at least 600 conversions tied to its conversion action within the last 30 days. This is basically a general guideline essential for the availability of this model. As soon as you achieve the minimum required attribution data, the process of preparing a data-driven model will be launched. This is when your data becomes available in AdWords.
Now, you may notice that some of your Google Analytics and website conversion actions have data-driven attribution while others do not. This is mainly because the data for each conversion action determines data-driven attribution. Just so you know, any marketer employing data-driven attribution whose data falls below 400 conversions for the conversion action or 10,000 clicks for the account within the space of 30 days will definitely not be able to continue using this model.
To this end, it is important to always ensure that everything’s working properly so as not to experience an unexpected decrease in data. You can do this by checking your conversion action settings, your Conversion actions’ status, your conversion tracking tag, and other account settings available.
Get beyond the times
While many marketers are still assigning credit to the last click, it is important to note that you could be possibly missing out on a significant percent of your potential ROI with it. If you are looking to make a significant increase in performance (between 25 to 50 percent) then you must be ready to move away from the use of the Last-Click Attribution model and move towards a data-driven attribution model. Ensure to test the data-driven attribution model first and see for yourself how it affects your ROI before trying it out.
Micro-conversion may be helpful
Data-driven attribution does not just distribute credit to any channel based on its location on a conversion path. To this end, it requires sufficient data to perform an efficient modeling task. However, it might be very much reasonable to establish a micro-conversion event particularly if you are not eligible to perform data-driven modeling. This might make much sense, as it can be used to evaluate the data-driven attribution model against the others. For those marketers whose conversion actions fail to meet the criteria, it is recommended to set up a micro-conversion event as it can prove to be very helpful.