The 5 most common attribution mistakes
Written by Sarah Geoffrion
16th January, 2018
Attributing conversions to their original source can be tricky, but it's also an important part of understanding ad effectiveness.
With a number of different attribution models, there’s no ‘one model to rule them all’. We take a look at 5 of the most common mistakes when tackling attribution and how you can avoid them.
1. Waiting for perfect data
It is almost impossible to track the entire user journey, as much as every marketer wants to measure each advertising touch point, no one can keep tabs on which tube ad someone sees before they Google something.
This inability to have complete 360 degree vision of a consumer’s buying process leads some marketers to abandon any efforts to get a clearer view on their Attribution and refer to only measuring their click through rate.
It is important not to wait for this ‘perfect data’ as in reality, it isn’t going to pop up anytime soon and there are other things you can do other than resort to only measuring click throughs. Look at the things you can measure, including offline measurements. For example, where are your highest conversion rates geographically? Do your customers buy online or offline more often? Use the data you can measure effectively, rather than worrying about what you can’t.
2. Using bad data
If you think of your data as an input to your model which provides an output, it doesn’t make sense to input bad data.
Consider how your data could be unreliable, for example, does your location data consider that the user could be using a proxy? Or could you be counting touch-points multiple times whilst trying to match up data from different sources?
Incorrect data skews your attribution model negatively; even if your data model is precise if the data going in isn’t correct you will get bad analytics coming out so it is important to input as accurate data as possible.
3. Losing sight of the wider customer journey
To get a true picture of the customer journey, marketers must take into account the user’s experiences with other brands. For example, if a customer has a negative experience with a competitor it can affect the way they interact with your brand.
Customers know that if a brand doesn’t interact with them on their terms then another one will, so it is imperative that marketers run a multi-channel strategy to ensure they are reaching customers in different ways and catering to all users.
Comparison is also increasingly a part of a customer journey, so it is important for marketers to acknowledge this and make appropriate changes. For example, including your competitors’ names in your keyword search criteria so even if a customer searches a competitive brand, you appear as an alternative. Or searching for competitor keywords in interests on Facebook, if they are available, and targeting users who have listed competing brands as an interest.
4. Ignoring cross-device tracking
Often attribution breaks when people change devices, for example, moving from a desktop to a mobile device. Cross device tracking helps you to see a fuller picture of the customer journey and compare levels of interaction across devices.
It is really important to be able to attribute conversions to different clicks and impressions on specific devices as you may find that people convert better on mobile or tablet and then you know when to focus budget etc.
Although it’s not all about where the customer converts, cross device tracking can help you understand the customer journey and in turn help you decide where to invest your budget. For example, customers may convert on desktop but they could have all viewed an ad on mobile first.
5. Ignoring the role of offline ads and offline stores
Over 90% of all retail purchases are made offline, so for marketers it is important to understand which online activities contribute to this offline conversions. This is where data onboarding comes in – this is the process of putting offline data online to help marketers match customers in both places and hopefully see a fuller view of the user journey.