75% of people end their purchase journey on a different device they started on, so it’s becoming increasingly harder for marketers to follow their path. Thankfully, Google Attribution maps it out for us, so we can bid more intelligently with our ads.
WHY GOOGLE ATTRIBUTION?
Half of web traffic comes from mobile devices, so users can search anytime, anywhere and on anything. 75% of adults start their online activity on one device, but continue or finish on another. Whether that be another mobile, tablet, desktop or laptop. We always look at the last touch point of how they converted, but we need to look at the whole journey with the help of Google Attribution.
For example, have you ever booked a holiday in one sitting on one device? The entire process could be spread out over a number days or weeks on several devices. Google Attribution helps you connect all these interactions in any given customer journey.
When you know what works, you can invest more in those interactions.
WHAT IS GOOGLE ATTRIBUTION?
It’s all about a path. A path your customers take when making a conversion. Google’s Attribution model is the rule, or set of rules, that determines how credit is assigned to touchpoints along the conversion path, not just the final click. It makes us look at the bigger picture and how a purchase started with a simple query.
For example, if you have a total of three campaigns and a customer clicks on each before leading to a conversion, Attribution will distribute that conversion over the three conversions, not duplicate it three times.
WHICH ATTRIBUTION MODEL SHOULD I USE FOR MY CLIENT?
On average, there are five interactions between first touch and purchase. By looking at only the last step, you are only looking at 20% of the data, you are neglecting the other 80%. You wouldn’t buy a five-bedroom house after only looking in one bedroom. You need to look at everything before making any decisions.
There is a total of six attribution models, but we are only going to talk about four of them, as Last Click and First Click don’t account for multiple interactions in a customer journey.
Linear– Distributes the credit for the conversion equally across all clicks on a path.
Time Decay – Gives more credit to clicks that happened closer in time to the conversion. Credit is distributed using a 7-day half-life. In other words, a click 8 days before a conversion gets half as much credit as a click 1 day before a conversion.
Position-Based – Gives 40% of credit to both the first and last-clicked ads and corresponding keyword, with the remaining 20% spread out across the other clicks on the path.
Data Driven – Distributes credit for the conversion based on past data for the conversion action. (This is only available to accounts with a minimum of 600 conversions/15,000 clicks over a 30-day period).
Data Driven is the best model, as you track each individual interaction and build a better customer journey. This model avoids you making an assumptions and instead better educated decisions.
If you cannot meet the Data Driven minimum requirements, it may be worth adding micro-conversions to your site to boost your numbers. For example, signup forms. Make sure the micro-conversions relate to your business and a customer’s conversion journey.
If you cannot meet the requirements, then the other three models will be better suited:
As you can see from the graph above, it is better to choose either Time Decay or Position-Based models depending on your business. Linear does not give an accurate enough conversion path.
HOW TO CHANGE TO NON-LAST CLICK MODELS
If you have Attribution enabled, at the top of your AdWords dashboard you select Conversion Settings > Conversion Actions > At bottom on page you can select a model. Select Time Decay or Position-Based if Data Driven is unavailable.
When using multi-touch models, AdWords will report conversions based on earlier clicks in the conversion journey. To properly analyse performance of these touch points, allow sufficient time for conversions to register (we recommend a 30-day range) and avoid including recent days (first 30 days).
When you change the model, conversions are not only associated with the last click (which often happened on that day). Instead, the fractional conversion credit will be redistributed to previous days. There is no loss in conversion, just redistribution. Over time, conversion credit on the day of the change will normalise.
For example, many businesses will increase their ad bids on Black Friday. If you have a last click model it will show a spike in conversions and traffic, but this isn’t accurate. If you have changed your model, you will see an increase in traffic in the days prior to Black Friday, as customers are researching their purchases. You want to influence the buyer days and weeks before the event when they are researching and planning their purchases.
WHAT TO DO NEXT?
Use the insights from your new Attribution model to drive strategic decisions. Understand how long the customer journey is, how many interactions there are and then change your ads to reflect this. Consider how many conversions shift from desktop to mobile.
Multi-touch models often reveal a shift in credit to mobile devices and upper-funnel targeting. Use the upper-funnel keywords to first capture that customer. You may have several paused keywords in your campaigns because they aren’t performing as strongly as others. You had a hunch those keywords would work when you first added them, so un-pause them as they will help capture more relevant traffic. They won’t cost you much money due to low performance and are beneficial to capturing traffic.
This is where Smart-Bidding comes into play. Some people may not be interesting to you, but they are interested in you. Smart-Bidding techniques helps capture those people. It makes your ads attractive to users who are more likely to convert and lets you bid on every step of the journey, helping you return better ROI. Google sets a bid for every auction based on a user’s likelihood to convert.
Google considers 70 million variables when determining the bid- both individually and the intersection of those variables.