Using attribution models
- Former user (Deleted)
- Martin Heymann
What is an attribution model?
As we have seen in the article on customer journeys, customers use various devices and touch on various channels and ad spaces before a conversion on an advertiser’s website takes place: they read an article on a consumer blog, look for products in a search engine, compare products on a price comparison site, look for vouchers and type in the name of the advertiser’s shop in their browser.
When different channels are part of the customer journey, it can become difficult to evaluate your marketing and assign costs and commission properly. Therefore, you can apply a so called attribution model. In an attribution model, you define different levels of priority to different channels – or, as we shall see, on category or ad space level. The touch point that is related to the channel with the highest priority will be defined as the winner of the customer journey. That means that the costs and commission of the conversion will be attributed to this touch point.
How attribution models work
Last touch point wins?
In the early days of online marketing, most marketing managers applied the ‘last touch point wins’ model when it came to defining winners. This meant that the last touch point involved in the customer journey was always considered to be the winner. For various reasons, this principle is highly problematic.
For one thing, the last touch point does not necessarily represent the decisive point in the customer journey. To give a simple example: a customer reads a product review blog of one of your partners and visits your advertiser’s website after clicking on an ad media item integrated on the blog. The next day, the customer remembers the advertiser, directly types in the URL of the advertiser in his browser and buys the product the partner was writing about. Clearly, the partner that wrote the blog generated this conversion; however, with the last touch point wins principle in place, the direct type-in will be regarded as the decisive channel.
There is even a more structural edge to the issue. Some types of channels are mostly touched on in the early stage of the customer journey, others tend to be touched on in the last phase: most customers will read consumer blogs before they start looking for vouchers, to name but one example.
Priority based on property
The technology on the platform allows you to configure an attribution model that does not primarily look at the chronological order of the touch points in the customer journey. Instead, it defines priority based on the properties of the channel.
The first property the channel looks at is the so called media type of the channel, i.e. whether the traffic generated via the channel is considered earned, owned or paid:
Earned traffic is traffic that is generated because potential customers are already familiar with one of your advertisers or find them looking for products the advertiser is selling. Examples of earned traffic are direct type-ins of the URL of an advertiser’s website in a browser, organic search engine results but also non-sponsored references to advertisers and/or their products on social media.
Owned traffic is traffic which is generated by marketing channels you or your advertiser possess: a customer newsletter, for example, or traffic coming from the product blog on an advertiser’s website. It could also be traffic coming from an advertiser’s own Twitter account.
Paid traffic is traffic generated by channels in which you directly invest money, such as affiliate networks, sponsored social media posts and display marketing.
In most cases, you will give the lowest priority to your earned traffic. Earned traffic is not unimportant, but it can only be there because another channel has already delivered. Paid traffic usually gets a high priority, because it involves channels that are proactively marketing your advertiser’s brand and products. In case of owned traffic, you are promoting your advertiser yourself and therefore you will usually give an average priority to the marketing channels involved.
Furthermore, you can differentiate between individual channels. This is particularly important for channels generating paid traffic. You could, for example, argue that a product blog in the content channel delivers a more substantial contribution than a website distributing your advertiser’s vouchers. The same would go for the email marketing and the display channel: because partners dispatching newsletters are capable of specifically targeting their addressees, the traffic coming from this channel will have a much higher quality than the traffic coming from the display channel.
An advanced attribution model would thus look more or less like this:
Show 102550100 entries
Search:
Tracking channel | Media type | Priority level |
|---|---|---|
Direct type-in | Earned | 1 |
Organic social media | Earned | 1 |
Organic video | Earned | 1 |
CRM | Owned | 2 |
Customer communities | Owned | 2 |
Call center | Owned | 2 |
Content | Paid | 5 |
Paid search generic | Paid | 4 |
Paid search brand | Paid | 3 |
Paid social media | Paid | 4 |
Email marketing | Paid | 5 |
RTB | Paid | 4 |
Retargeting | Paid | 4 |
Offline advertising | Paid | 3 |
Affiliate networks | Paid | 4 |
Price comparison | Paid | 4 |
Recommendation marketing | Paid | 4 |
Showing 1 to 17 of 17 entries
PreviousNext
Defining priorities on different levels
The tracking channels give you much room for differentiation, but you you might want to differentiate even further. You might want to give a specific content blog priority over all others, or give a lower priority to a specific type of retargeting. This is possible. Please remember that every ad space on your platform can be assigned to a specific subcategory, which are assigned to a category. The categories for their part are allocated to channels. Thus, you have four different levels on which you can assign priority in your attribution model:
Defining priorities based on touch point type
Basically, there are two types of touch points: clicks and ad impressions, the latter also known as views. In most cases, clicks will be regarded as far more valuable than ad impressions, because they directly lead customers to the advertiser. Therefore, you can also give a higher priority to clicks in your attribution model.
Furthermore, you can define a different attribution window in relation to the touch point type. The attribution window defines how long a touch point will be taken into consideration for attribution. Because of the higher weight clicks usually have, their attribution window will be longer (30 days is common) than the attribution window used for ad impressions, which is normally limited to a couple of days.
Multiple attribution
In some cases, there might be a good reason to attribute costs or commission to multiple touch points within the customer journey. For example, when you are working with display partners, you will probably be somewhat reluctant to work on a CPM or CPC model. However, if you attribute only to a single touch point in the customer journey, you will most likely not do justice to the contribution of the display channel. Touch points related to this channel will usually be situated at the beginning at the customer journey and most of them will be ad impressions.
To solve the problem, you could give the display channel a higher priority, but then you run the risk of disadvantaging other channels. A better way to solve the issue is creating a so called attribution group. An attribution group consists of various channels, categories, subcategories or ad spaces to which conversions will always be attributed when touch points related to them are part of the user journey.
Read more about: Multi-Touch Attribution (MTA)