Marketing is important if you want to make your product or service visible. This ultimately leads to conversions. That’s the end goal of any marketing campaign or channel. If you’re a client, you would want to know the impact that the marketing touch-points have on conversions. On the other hand, if you’re the marketer you need to communicate how each channel is important to conversions. This allows you to determine whether you’re using the marketing budget to the best of your ability.
Marketing attribution helps you get to the bottom of all this. It assigns credit to the different touch-points in the customer’s conversion journey. Through attribution, marketers are able to come up with rules that determine what each touch-point contributes to the customer journey. Therefore, attribution marketing is aimed at helping people optimize their marketing strategies.
Simple Models of Attribution
Some companies or businesses have marketing to sales processes that are very simple and straightforward. They don’t go through many channels or campaigns for conversions. This type of marketing attribution is easy to come up with. The most common attribution models cater to these types of businesses.
First touch and last touch attribution models are simplistic. They’re also some of the most commonly used models. They assign credit to the first and the last touch-point respectively. Ideally, this means that they assign all of the credit to just one channel or campaign. This isn’t realistic especially in cases where there’s more than one touch-point. These simplistic models of attribution make one major assumption: that only one touch-point is important in the customer journey.
Complex Models of Attribution
The simple models of attribution, as we’ve seen, have one major flaw in the assumption. As a marketer, you cannot rule out any channel or campaign based on your intuition alone. It might work in the short run but it will end up costing you early in the long run. Without optimizing your marketing strategies, you’re definitely wasting money. The solution is to use the complex models of attribution. These are also known as multi-touch attribution models. They assign credit to various marketing touch-points but in varying percentages.
Attribution marketing today focuses on these types of models. This is because the marketing process has become more complex over the years. They are important when it comes to:
· Accommodating for all touch-points
· Identifying which touch-point contributes more or less to conversion
· Reallocating your budget according to this information
Some of the simpler models in this category are the position-based models which include the U-shaped and W-shaped models. The more complex multi-touch attribution models are the data-driven ones. These data-driven models use machine learning. You need a good number of conversions for the model to work. The machine collects data on all the touch-points in the customer journey and then assigns weights. The process continues so that the machine reassigns weights to capture the current situation and help in marketing optimization in real time.
Evolution of Attribution
Marketing attribution has come a long way: from simple first and last touch models which are largely biased to data-driven models which are accurate and scientific. Apart from the evolving nature of marketing, the need for actionable results has also influenced the changes in attribution modeling. In time we can be sure that with the increased complexity of the marketing process, the complexity of the attribution models will also increase. For example, the use of mobile search has become more significant over the past few years. Before that it was not included in most attribution models or assigned any credit.
For a long time, people were forced to do marketing with no way of knowing just how effective their channels and campaigns were. Sure, you could see an increase in the number of conversions. However, you would not be able to tell which of your ads or campaigns yielded the results. It was therefore difficult to assign credit or find out where money was being wasted. Today, you are able to assign credit, monitor, and maximize on your channels and campaigns. Technological advancements have made attribution modeling possible. One of the best methods is the multi-touch attribution model.
Single Point vs. Multi-touch Models
From the names, you can easily tell what each model entails. The single point attribution models assign credit to only one channel or campaign while the multi-touch model assigns credit to each and every channel on the customer’s journey.
The single point models include first touch and last touch models. The former attributes conversions to the first channel while the latter credits the last channel. As for multi-touch attribution models, they come in different types. These include: linear, time decay, u-shaped, w-shaped and custom attribution models. We will look into this later.
Advantages of Multi-touch Attribution
Before you decide against using the single point attribution models, you need to have a good reason for this. The benefits of using a multi-touch attribution model should outweigh those of using a single touch attribution model. One benefit of assigning credit to every touch-point is that it gives you an idea of each channel or campaign’s contribution to the customer’s journey. This then informs you on which touch-points to focus on and which ones you can do without. Additionally, this helps you in forecasting. You can plan the business’ growth by using effective marketing techniques and ensuring that they have the necessary budgeting provisions.
Examples of Multi-touch Attribution Models
As earlier mentioned, there are different types of multi-touch attribution models to choose from. These include:
· Linear Attribution:
The linear attribution model assigns equal credit to every touch-point. It does not matter at what point it is on the customer’s journey. This is an improvement from the single point attribution models. However, it’s a very simplistic view and assumes that all channels are necessary and that they all make the same contribution to conversions. In real life, this isn’t the case.
· Time Decay:
The time decay model, on the other hand, assigns more weight to channels or campaigns that are nearer to the conversions. The general thinking is that the first touch-points would have led to immediate conversions if they were that important. Therefore, these first touch-points have the least weights.
· U-Shaped Model:
Unlike time decay, this attribution model assigns more weight to the first and the last touch-points in the consumer’s journey. The channels in the middle have equal weights that are less than those at the ends, i.e., the beginning and the end of the journey.
· Custom Models:
This is one of the most advanced attribution models. It assigns custom weights to each of the channels or campaigns. The amount assigned is upon your discretion. For these types of models, the blueprint lies with the marketer and is based on the desired goal. To develop the model, you have to go through a lot of trial and error before you can get to the perfect model form.
Analysis of different attribution models and which is the best for my business
Google Analytics has had a report on attribution models for many years.
All marketers in the world have heard about attribution models, understand them, know they are or will be necessary, but few really know how it works. Honestly, it’s really a mess of noses.
Google Analytics says this:
and then he shows you this:
Well, now, go ahead, optimize and sell!
Let’s give a little light 🙂
As we have already said, the attribution is in charge of distributing in each touchpoint (click) the prisoner of 1 sale based on some rules. According to the first image, the one with the longest bar takes more “weight”.
Last Click Attribution Model:
Let’s remember one thing that is unknown: How does Google Analytics treat Direct Traffic? We do not do that in Adinton. In this case we measure well and each thing we call it by its name, if it is Direct it is Direct and if we do not know what we call it (not set).
From here, this model assigns the conversion to the last click. Therefore everything that the user has done before is not valid.
Sources favored to this attribution model:
- Closing sources: Normally SEO-Brand and Direct.
If the journey of your users is very direct, that is, the user arrives and buys or contacts, it is the model that makes the most sense.
- All those that do not close. That is, almost all! Years ago it was pausible that a user did not leave google and between SEO and PPC, the sale is made. But nowadays and increasingly, users are looking for, comparing, documenting, … trips are getting longer in number of clicks but shorter temporarily.
“It is increasingly common for customer journeys to be longer in numbers of clicks but shorter in time” username=”adinton”
First Click Attribution Model:
The weight is given to the first click. This model especially I like for those companies or projects where the brand is not known or the product is innovative or new.
Imagine, nobody knows you or nobody knows the product you are marketing; therefore it is the first click that has initiated the purchase process, has been the one that created the need.
Sources favored to this attribution model:
- Initiating Sources: Normally paid.
No one knows us, you have to pay to get to know us, whether buying traffic, posts or investing time (money) to promote blogs, influencers, …
- All those who are not Initiators. I say they are prejudiced because something should be assigned to them, because their job is not to initiate the need, but the user does not escape.
Linear Attribution Model:
This is the one I like least, I find it a vague model, it is the typical one that is chosen when you know which is the best for your business. Since I do not know which one is the best, I choose all of them.
Sources favored and harmed to this attribution model?
Everyone benefits or they will be harmed depending on the customer journey standard of your business.
If your business is “new” like the one we explained in the first click, the damaged source will be the first one and all the others will benefit.
On the contrary, if your model is “last click”, the source damaged will be the closing and the rest will benefit
Time Decay Attribution Model:
It may have this sense or inverse. The key whether to adopt the growing model (such as the image) or decreasing, is to know our users, if we have a brand / products known and users need to be “nourished” to close the deal, this model is appropriate.
Sources damaged or favored, none, if the model reflects the reality of our customer journeys.
Customized Attribution Model.
This model is very “good” (ironic). Here they sell it to you as we analyze the customer journey of your users of the last 30, 60, 90, 1 year, … I do not care and with that analysis we give you a bespoke and customized attribution model for your business.
Bravo!!!! It makes sense and a lot! but until you ask yourself:
- The 60 – 90 days analyzed is high or low season?
- The prices and products that I have now are the same as we had during the analysis?
- The prices and products of my competition are now the same as we had during the analysis?
- Difference between new and recurring customers?
- Difference between clients under 30 and over 40?
The model is still valid if …
- My competition goes up / down prices?
- Does my competition broaden / reduce catalog?
- Does my competition improve SEO and / or CRO and / or PPC and / or ….?
- dem if the three points above apply to my own business.
No Standard Attribution Model is valid.
I’m sorry to be a party pooper and more after having arrived here reading. But imagine for a moment that in the second click of a user’s trip, that user queries 1,000 page views, downloads all the pdfs, adds all the products to the cart, … but buys in the 4th click. Absolutely no model would give more weight to the second click.
Data Driven Attribution Model.
This is the model that I trust after 4 years analyzing customer journeys. You are going to tell me, “Bravo, if you use your little machine, it’s the one you want to sell me.” Put it backwards, we have been analyzing customer journeys for so long, that we have decided to throw away all the models and keep this one. Basically because the rest of models are noise. They only serve to distract and NOT make decisions based on them.
The data driven model consists of assigning the weight of a conversion to each click dynamically, taking into account variables such as:
- Order of the click or not. It depends on each company.
- Bounce Rate
- Duration of the visit
- Temporary distance between clicks
- Conduct Microconversions (pdf downloads, add to cart, subscribe to newsletter, …)
- Differentiate between new vs. recurrent customer
And this attribution is calculated individually, journey to journey and in real time, this is how we do it in Adinton.
If you do not want to take the step to a Data Driven Model, I pass this table so that you can choose the one that best reflects the reality of your base customer journey.
Attribution marketing is one of the most popular things for marketers right now. With the right attribution models, you can produce actionable results from your campaigns and marketing channels. With that you can tweak the budget and therefore optimize your marketing process.