Attribution modeling is the process of determining how credit for sales and conversions is assigned to various touch points across the marketing journey.
The challenge is figuring out how to properly attribute sales in a world filled with non-linear customer journeys. Since every buying cycle is different, attribution models aren’t a one size fits all solution. Let’s go through six major attribution models and see how to use these models in Google Analytics.
First Click Attribution Model
Definition: In the First Click attribution model, the first touch point receives 100% of the credit for the conversion.
Example: Use First Click to measure the success of reaching a new audience, like running promotional ads on Facebook to generate buzz about a new product.
Most Effective: First Click is most effective if your advertising is limited to one or two channels. Because you know a customer has limited options for finding your store. Use first click for measuring brand awareness and to help find which channels are driving new customer acquisition.
Least Effective: First Click Attribution is not effective for longer sales cycles or multi-channel marketing. This is because a customer may interact with multiple channels prior to purchasing, and First Click ignores a large portion of the marketing process.
Last Click Attribution Model
Definition: In the Last Click attribution model, the final touch point before the sale would receive 100% of the credit for the conversion.
Example: Use Last Click to measure the success of an email campaign. Say you run a campaign to your current customers with a promo code for 20% off and make $10k. All of that revenue would accurately be attributed to Email using Last Click Attribution.
Most Effective: Last Click puts emphasis on the final actions before a sale. Use this model to most effectively measure the “decision factor” that resulted in sales. Similar to First Click, Last Click works best for businesses with shorter sales cycles.
Least Effective: Last Click is the most common attribution model. (But that doesn’t make it the best.) Because Last Click adds emphasis to the end of the funnel, this model shouldn’t be used for products that involved a long research process. Think of buying a mattress online. Rarely would you see an ad and immediately make a purchase. When you are doing a large amount of research on a product, the earlier channels that first introduced your product should be properly weighted when measuring success.
Last Non-Direct Click Attribution Model
Definition: In the Last Non-Direct Click, all Direct Traffic is ignored and 100% of the credit goes to the last non-direct channel that the customer clicked through before converting.
Example: A customer clicks on a Display Ad and previews a product and then leaves. Three days later they enter your URL directly into their search bar and make a purchase. Last Non-Direct Click would attribute the conversion to your Display Ad.
Why would you want to look beyond Direct Traffic?
Unless you work for a brand that has universal recognition (like Nike), it’s safe to assume that most of your customers heard about your brand a few days or weeks before going directly to your store. Direct Traffic is when a customer enters your URL directly into the search bar and navigates to your site immediately.
The majority of the time, customers are circling back from a previous channel. This model places the credit with the channel that most likely should be attributed with a sale.
Linear Attribution Model
Definition: In the Linear attribution model, each touch point on the conversion path shares equal credit for a conversion.
Example: A customer finds your store through Organic Search, then joins your email list and clicks to your store from an email. Two months later they are retargeted by a Facebook Ad and go to your store directly. A week later they see a Display Ad and make a purchase. The Linear model would give equal credit to all five channels.
Most Effective: The Linear Attribution Model is effective to measure overall brand awareness, and to see which channels are consistently influential during a customer journey.
Least Effective: Unfortunately, the Linear model can sound better in theory than in practice. To quote a marketer we follow, “Basically everyone on the team gets a trophy.” When all channels are awarded equal weight, it is near impossible to choose a winner. And the purpose of attribution modeling is to find what is working successfully so you can continue to replicate that success.
Position Based Attribution Model
Definition: In most Position Based attribution models, 40% of the credit for a conversion goes to the first and last touch points and the remaining 20% is evenly distributed to the middle touch points. (Note that weighting is flexible and can be tweaked as you see fit.)
Example: A customer wants to buy a bicycle online. They run an Organic Search and click on a link. From your site they join your email list and click on a link in your email. After that they are retargeted on Facebook, and continue to research your store by directly going to your blog. A week later they see a Display Ad for your bikes and make a purchase. In this example Organic Search and Display Ads would receive the majority of the credit, and the middle three channels would receive partial credit.
Most effective: Position Based Attribution heavily weights the channel that first brought in a customer as well as the channel that caused a conversion, but it also credits channels in between that helped nurture and keep your customers warm. This helps you find out which channels to focus on for bringing new customers in and which channels to focus on when trying to convert at a high rate.
Least Effective: Timing has a huge effect on the buying process. Position Based Attribution can cause some arbitrary allocation of the weights to channels. For example, if the first channel visited receives a 40% weight, but six months later two more channels are visited in less than a week, the entry channel may be artificially inflated given it wasn’t the influencing channel.
Time Decay Attribution Model
Definition: The Time Decay model gives more weight to channels closer to the conversion point and less to channels earlier in the funnel.
Example: Time Decay is meant to give credit to the channels that helped your prospect reach the conclusion to buy. If you use clear calls to action and a coordinated journey, this method can help asses the effectiveness of moving customers through a buying funnel.
Most effective: Since the Time Decay model credits every touch point, it works well for stores with a large amount of repeat purchasers. These customers are exposed to different advertising and marketing methods, so using the Time Decay model can help find what is driving repeat conversions.
Least Effective: Like Position Based modeling, Time Decay can inaccurately provide too much or too little value to a channel. For instance if a customer is buying a computer online, the final conversion point could be the result of a week-long research process. These touch points (like Organic Search or Direct) would weight much higher where in reality a touch point earlier on influences the sale just as much as later touch points.