Choosing an appropriate attribution model is integral to distributing marketing budgets and understanding the results of your marketing campaigns. There are many different attribution models to choose from, so how do you know which ones are best for your business? Your data and user purchase behavior gives insight into which model is right for you. We will cover typical scenarios of each attribution path.
This model attributes the complete conversion value to the last marketing channel the customer used before completing the conversion.
What’s the Last Interaction Model Good For?
Last Interaction is best for sales cycles that are very short, where users don’t have a research or consideration phase. If your products have high price points then last click usually isn’t the best model. This is because users (hopefully) are taking time before spending large amounts of money. Last interaction is better for transactional campaigns versus lead generation. With transactional campaigns, you want users to purchase as many products as possible, but with lead generation the lifetime value of a customer is higher, therefore you need less conversions.
If your customer’s journey doesn’t have many touch points, that is they don’t interact with various marketing platforms – social, email, PPC, etc – then Last Interaction should work just fine. When this is the case, use Last Interaction to qualify your leads or transactions. Which platform yields the best results? Look into what types of ads that platform is running and try to find out why it’s successful.
Pro Tip: Last Interaction is also called Last Click and Last Touch.
Last Non-Direct Click
This model is the same as Last Interaction, except for if the last interaction before a conversion was considered Direct traffic, the channel touched before Direct will receive 100% of the conversion value.
The Customer Journey Explained
In this example, the user found the site through Search and saw a PPC ad that they clicked through. Later, while browsing on Twitter they saw another ad and clicked through to the site, but still didn’t purchase. Remembering your company, they typed your domain directly into the address bar and made a purchase. With Last Non-Direct attribution modeling, Twitter would get 100% of the conversion credit since Direct was the last click.
What’s the Last Non-Direct Click Model Good For?
This is the default model for Google Analytics when attributing conversions. The exception to this is when you’re using the Multi-Channel Funnels reports. Since Direct traffic is really just a catch-all for any traffic Google can’t identify – improperly tagged campaigns, HTTPS to HTTP, users typing your domain in the address bar, etc. – it can be beneficial to leave it out of your reporting entirely, instead of trying to guess.
It’s more likely there was incorrect tagging of campaigns than a user typing your domain in the address bar (still possible, but probably doesn’t account for most of your Direct traffic). So if you decide that you want to use Last Interaction as your attribution model, use Last Non-Direct Click since it’s impossible to attribute more marketing budget to a Direct channel.
Last AdWords Click
This model attributes 100% of conversion value to the last AdWords ad a user clicked on before converting.
The Customer Journey Explained
In this example, a user found your site organically and was then served an ad while they were browsing Facebook. Later, they searched for your website and a PPC ad was shown that they clicked on. While on Twitter, they were shown an ad which ultimately lead them to convert. The PPC ad that was clicked on before the final Twitter ad would receive this conversion credit. If the user clicked on multiple PPC ads, the ad that was closest to the conversion would receive the conversion credit.
What’s the Last AdWords Click Model Good For?
AdWords. It’s not very beneficial to any other channels besides PPC. If you want to see holistically which AdWords ad is directing users to the most conversions, use this model. This model will give you a good idea of the type of ad content users respond to.
First Interaction Model
This model attributes conversions to the first marketing channel the user interacted with before converting.
What’s the First Interaction Model Good For?
This model is great for monitoring impression and awareness campaigns. If you’re running a campaign designed to introduce users to your brand, using the first interaction model will show you the most successful exposure channel.
Using a First Interaction model for brand lift campaigns, helps you understand where your customers are responding to your ads. Maybe more of your potential customers are active on Facebook than Twitter so they interact with your ads much more on the Facebook platform. In these scenarios, First Interaction can be very beneficial for branding.
Pro Tip: The First Interaction Model is also called First Touch and First Click.
The Linear model gives equal credit to each touch point on the path to a conversion.
What’s the Linear Model Good For?
The Linear model gives the least amount of insight into your marketing efforts. It’s appropriate if your buying cycle is incredibly long, which is mainly B2B. If leads take months to close it’s important to consistently maintain contact with the potential customer, which is what the Linear model is used for.
This model gives more credit to the marketing channels closest to the conversion.
What’s the Time Decay Model For?
Time Decay and the next model, Position Based, are usually the preferred methods of attribution if your campaigns have multiple touch points. The theory behind Time Decay is the closer to the conversion, the more impactful the marketing channel was. The problem with this is it can discount the “research” phase, if you have a long buying cycle. It’s safe to use Time Decay if you don’t have exceptionally long buying cycles or if you’re running a short promotion. When running a short promo, you might want to give more credit to the channels that were touched during the promo as opposed to a week before.
The Position Based model combines the First and Last Interaction models, splitting the most credit between First and Last, while leaving smaller portions of credit to the middle channels. A common way to divide conversion credit is: 40% for the first and last interactions and leave 20% for the middle channels.
What’s the Positioned Based Model For?
The Position Based model helps you understand which channel brings users to your brand and which channel drives them to convert, so it gives you the benefits of both First Interaction and Last Interaction. It can undervalue the middle touch points, which can be a down side for long buying cycles. For the most part though, Position Based gives you the most holistic view of your marketing campaign and it provides the most flexibility. In Google Analytics, you’re able to change the percentage of conversion value given to each channel so if you want to give more or less than 40% to the first and last channel, you can. Position Based models give you a template to work with while tweaking it to fit your business model and customer journey.
It’s possible to create completely custom attribution models based on your data. Need help? Contact Us!