For many marketers and business owners relying on paid advertising and driving a lot of their activity through tools like Google AdWords, this can be tricky. As pay-per-click (PPC) advertising is driven by numbers, it’s vital that each and every sale can be tracked and attributed accordingly, in order to ensure advertising budgets are being spent on the activities that are most likely to convert.
In this blog post we’re going to discuss why using different attribution models in AdWords and other tools is useful for marketers, and how it can help you make better decisions about your marketing.
AdWords attribution models
Before we begin, let’s do a quick recap on which attribution models are available in Google AdWords. It’s also worth noting that attribution modelling is only available for campaigns on the Search Network (this includes Shopping ads). It’s not currently available for Display Network campaigns.
|When is this model useful?
|Last click attribution
|The last click and corresponding keyword before conversion gets all of the credit.
|If you want to maximise efficiency and customer recycling.
|First click attribution
|The first click and corresponding keyword gets all of the credit.
|If you a place a lot of emphasis on outward growth and new customer acquisition is your number one priority.
|Distributed equally across all clicks in the path.
|If you value each and every touch point you have with the customer equally and you want to assign value to reflect this.
|More credit is given to clicks that happened closer to the conversion, using a 7-day half-life. So a click that happens 8 days before a conversion gets half as much credit as a click that happens 1 day before.
|Maximise efficiency while still giving some credit to top of the funnel interactions.
|40% of the credit goes to both the first and last clicked ads, and the remaining 20% is spread out among other clicks along the path.
|If you want to emphasise key touch points (i.e. the first and last ads clicked) but also want some middle of the funnel efforts to be acknowledged.
|Distributes credit based on past data for this conversion.
|If you have a very active account and want to save time by letting Google do the heavy lifting for you. Google recommends you have at least 20,000 clicks and 800 conversions a month to get the most out of this model.
Why is using different AdWords attribution models useful?
By default, the attribution model used in tools like AdWords and Analytics is last click. But as the customer becomes smarter, and businesses want to have more touch-points with them throughout the sales cycle, this rather simplistic view is becoming less and less useful.
There are 3 key reasons why you might want to explore using a different attribution model in AdWords:
Identify more opportunities to reach customers earlier on in the sales cycle
Using more top-of-funnel focused attribution models like first click and position-based allow you to identify what the most successful initial touch-points are, and then use this insight to maximise growth. Conversely, you can also look at which first interactions are not as successful, and look at addressing this in order to move more customers through the sales funnel.
Find a model that suits you and your business
If your particular product or service has a longer sales cycle, such as a software-as-a-service product which offers a free trial as part of the sales process, you’ll probably want to use a different attribution model to understand how customers interact with you during their trial, and how this influences their decision.
Exploring different types of attribution model allows you to judge better how your ads are performing, and how they contribute to the different stage of the cycle. For example, you could discover using the time decay model that certain ads or keywords perform better as the customer gets closer to converting – and you can adjust your bids on these keywords accordingly in order to ensure you’re getting in front of these customers as much as possible.
Obviously, for many marketers, AdWords isn’t the be-all and end-all. In an increasingly multi-channel world of marketing, potential customers can have so many more interactions with your brand, in so many different ways. Think social, organic search engine results, email, text, and even good old print and TV advertising. How does this all fit together? Using attribution modelling within the AdWords tool is a great way to understand what’s happening in your PPC campaigns, but it doesn’t give you visibility of how a potential customer is interacting with you across all of your marketing efforts. As we’ve already mentioned, AdWords doesn’t even allow you to use different modelling across all of its own networks.
The bottom line is, even if you’re running an amazingly efficient and effective AdWords campaign, focusing solely on that and the data it gives you doesn’t give you a full picture of your customer’s journey and their touch points with your brand. Being able to see that they clicked on an ad 14 days before purchasing and then 7 days before purchasing and assigning value accordingly may be useful, but what happens if they received an email from you in between? How, if at all, did that affect their decision? How do you assign value to that interaction?
Remember – the attribution models that AdWords uses aren’t exclusive to the tool – they’re just ways of thinking about how your marketing efforts influence a sale and the value they add. These models can be adapted and applied across all of your channels using integration tools to give you granular detail on how all your marketing, from print to PPC, is performing in bringing in customers.
In an age where the consumer is getting smarter and smarter, using the default last click attribution model is becoming less and less useful. In order to be a smarter marketer, you need to be thinking about the whole picture – when during the sales cycle is the customer interacting with you? What does this mean for the final sale? The best thing to do is to decide on what exactly your goals are, log in to AdWords and experiment with which attribution model best serves you in achieving this goal.