What is attribution modeling? What is attribution in digital marketing? What are the different attribution models? Learn more about how it impacts Google Ads and PPC strategy.
The first statement I make is to know your objectives. What is my target CPA or ROAS are the main questions. These are highly influenced by the type of attribution you use and who impacts your source mediums, as well as the budgets you put behind them to grow your business. Consider what channels you are using, source mediums in Google Analytics and the value you place on each source/medium.
This all brings us back to the customer journey and when tracking and reporting on performance becomes a problem for Managers and marketers alike, for if you are not reporting on the same metrics that matter it can be easy to neglect keywords that contribute to the customer’s journey is different, perhaps less obvious, ways.
There are a series of other attribution models, set out below:
1. First Click
2. Last Click
3. Linear Attribution
4. Time Decay
5. Data-Driven
6. Position Based
7. Custom Attribution
As simple as it sounds and the Google Analytics default. This means the Last Click before the sale wins and all the value for that transaction is taking by that source medium. The main problem for a marketer is even though simple and still, what clients use, it weights most decisions towards brand and brand+ keywords.
Therefore, when making decisions you are missing the steps as part of the customer journey that people take and the top funnel keywords and sources including paid social and display are undervalued within a standard Google Analytics view.
Again nice and simple and the reverse of the Last Click. In this case, the First Click in the chain takes the value of the transaction. The advantages are it rewards top of funnel conversions paths and supporting channels and keywords are missing.
You can see the impact on the First Click within the Multi-Channel reporting section on Google Analytics.
From reviewing both First Click and Last Click transactions a linear approach could make sense. This is where each source/medium or keyword the user has touched on there journey to conversion takes an equal part of the value. This option has now also become available in Google Ads and is seen as the fairest option. The easiest way to explain a liner approach is if there was an order for £100 and there were 4 points evolved before the user bought. Each section would be given £25.00
The main issue with this attribution model is that, in reality, not all touchpoints in a conversion path are of equal value. This is therefore given weight to supporting mix elements and budget decisions are not weighted correctly.
Time decay seems a better option then again, it weights the value of the sale towards the last touchpoint. Therefore this diminishing value attributed to each previous touchpoint in the conversion path.
The model is great for an understanding, but most marketers tend not to use it as it again neglects previous touchpoints and can overvalue the brand.
Direct from Google in the explanation about Data-driven Analytics “Data-driven attribution gives credit for conversions based on how people search for your business and decide to become your customers. It uses data from your account to determine which ads, keywords, and campaigns have the greatest impact on your business goal”.
Data-driven analytics uses machine learning technology to distribute credit between clicks, however, this is only more relevant to google ads then analytics and cross source tracking is more of an issue then cross keyword tracking.
Custom attribution is if you have an advanced tracking solution you can set your own parameters and weighting of value.
This is great unless working with 3rd party’s they are unlikely to report to your specification.
Each attribution is personal and will have its own justification and model. What I can recommend is to use the multi-channel funnels in Google Analytics. There is a great tool with the model comparison tool in Google Analytics. The tools allow you to compare different models.
For example, your analytics is set to last click, which on average is the standard. However, you can run a comparison between, first, click and last click and see the comparison. This allows you to compare and make informed decisions if you do not have the ability to move to a data-driven strategy.