Why custom attribution modelling is a must for online advertisers?
So, before diving deeper into our topic, let’s learn what is meant by an attribution model. As the attribution model plays a pivotal role in the calculation of our ROI for a campaign.
“An attribution model gives out information about various methods that are used by digital marketers to break down and assign credits regarding conversions to miscellaneous channels, in the case where a user reaches the final conversion through multiple website visits or multiple channels.”
An attribution model is a must for an advertiser, and without an attribution model, your campaigns will be anchored to a sinking ship. Google Analytics provides you with two types of attribution models.
- Default Attribution Model
- Custom Attribution Model
As mentioned, we will be talking today about Custom Attribution Model. And how this particular attribution model will be the right prognostication tool for your investment. Before, doing so, let’s learn about the loopholes present in other attribution models.
There are eight models that fall under the category of default attribution models. So, let’s delve deeper and learn about each one of them.
- Last Interaction Model:
Also known as the last touch attribution model. In this method, 100% credit is rewarded to the previous touch point in the conversion funnel.
Generally used by an e-commerce business
- First Interaction Model
In this model, the 100% credit is given to the first interaction or the first contact point in a conversion path.
Usually, used for brand awareness and also on demand generation.
- Linear Attribution Model
In the case of linear attribution model, the credit is equally divided among all the touch points in the conversion path.
- Time Decay Attribution Model
Over here, the highest weightage of credit is given to the most closest interaction in time during the conversion.
- Position Based Attribution Model
In this model, the first interaction is given 40% credit, the middle one is credited with 20% and the last point of interaction is assigned with 40% credit.
- Last non-direct Click Attribution Model
This attribution model is adapted by Google Analytics for the creation of multi-channel Funnel reports.
In this method, 100% credit is assigned to the last non-direct click in the conversion path
- Last Ad Click Attribution Model
Here, the whole credit is subjugated to the last Google Ads click during the path of conversion.
Data-Driven Attribution Model
This attribution model is only available to the Google Analytics 360 and this is a premium feature. In this attribution model, a specialised algorithm has been designed to assign credits for several touchpoints during the path for conversion.
So, it was a short description of about the several attribution models. Now, moving towards the bigger question.
“Which attribution model I shall adhere to?”
Still Confused!! Talk to an expert now.
Since everyone does not fit in the same shoe, a single attribution model can’t cater to all your needs. And if you are doing the same then maybe you are at a significant loss.
So, what’s the catch. How to get rid of the problems caused by the default attribution models?
Well here’s what anyone shall be ready to embrace
Markov Attribution Model:
This model is somewhat superior to the default attribution models in terms of the insights provided. So, let’s learn what a Markov model is and what insights does it gives.
In this model, a dataset of customer journeys is being used to allot the probabilities to the network of these stipulated channels. After getting the network of marketing channels attached to the probabilities, we will tend to inspect the effect of removal of a particular channel on the conversions.
So, by the removal of one channel a particular time and estimating the value of the change of conversions, we can encompass the complete picture of how every channel contributed to the conversion process.
This model indeed helps us to entail the complete picture about how a particular campaign performs and what is the weight of the same in generating conversions.
“The right performers are the ones that catalyze particular custom attribution models for their particular campaign.”
Yes, even though Markov attribution models gives a complete picture of all your campaigns, every campaign has a particular goal, and the righteous heir of the attribution models shall map out that specific goal and pinpoint where it will fail if it fails.
Below are examples of how not resorting to custom attribution models will cost you conversions.
- Let’s take a case study of one of our client Mr X. Mr X has a business to buy and sell old cars. His company can be categorised as Lead Generation. Prior to coming to us, Mr X used to follow First Click Attribution model and however was more worried as he used to run ads across different channels. And, Facebook uses to show ten conversions but Mr X, when increased the budget on Facebook ads, later got into the loss.
So, where does the conversion go?
Let a user saw an ad via Facebook, and he gave it a thought to buy a used car and landed upon the webpage of Mr X. At that point of time, the idea, however, didn’t materialise and he felt famished about the notion of buying a car and went back to his work. Later, that day, his wife taunted him about the neighbour’s car, and he got reminded about his hunt for used cars. Then he made a google search and again landed up on the web page of Mr X through Google Ads. But, at this point of time he also he didn’t buy.
Some days later, he got a referral for the same from one of his colleague. This, however, persuaded the user to carry on with the sale. The final deal was generated due to the referral from an organic search but not from Facebook. Mr X was thrilled and then enhanced the budget on Facebook ads and reduced from other channels.
After some days, there wasn’t any conversion from facebook, and since the budget was scaled down from other channels. The point which was the main reason that helped in the conversion got abandoned. Hence, the conversion rates got down.
After coming to us, we tailored a custom attribution model according to the needs of Mr X, and since then he was not only able to generate leads but also got a complete picture of the performance of their campaigns.
“Don’t be like Mr X. Take the correct decision”
- Well, what would have happened if you would have opted for last click attribution model?
This question will inevitably pop in your mind. Let’s take the case and see how the last click will also fail.
Well, the factor that was obligated to conducting the sale was the referral. The referral forced the user to perform the purchase, and hence, Mr X should bid more on the referrals. But if we would have opted for the last click attribution model, then 100% credit would have given to the organic search. And instead of focusing more on referrals, our focus will be riveted on organic search.
And again the default attribution model failed.
The main aim of an advertiser is to be cynic enough to ascertain from where the conversion came and what was the factor that drove the sale. Which audience responds well and what are the factors which can inflate my cost per conversion or the cost per lead.
All these questions can not be answered by a single attribution model, and for every campaign, a tailor-made attribution model will help to generate the answers.
Still confused and taken aback from the dark side of attribution models?
Talk to an expert now and get a tailor-made attribution model for your business.