In our recent Marketing Monday video series, we covered a topic that is challenging modern marketers, which is marketing attribution. In this blog, I’ll provide guidance on what attribution is, why it’s important, and where you can get started.
What is marketing attribution?
Marketing attribution is the process measurement of the value of each touchpoint in a customer journey relative to the total value of the customer engagement. So whether a customer buys a $100 pair of shoes or is part of a multi-million-dollar complex sale, the goal of marketing attribution is the same: to help the company understand the value of each of the interactions through those engagements. Was it a digital ad, webinar, white paper, or an email that convinced them to purchase your product—or was it a combination of all these activities?View Transcript
Today, we are going to talk about marketing attribution. So what is digital attribution modeling? Well in marketing, we typically create journeys for customers and they have a start and an end. But it's not just a one-step journey. There are multiple touches, and one touch will hopefully end up in a purchase for your organization. What we are trying to do is give the appropriate credit back to all those touches in our digital marketing world so that we can start to really understand what’s driving customers to purchase from us and how we can speed that process up from a digital marketing channels perspective.If you take a look at the sample journey here, this is more of an e-commerce type journey. The user does a Google search for a particular product, they see a display ad that has been developed by the marketing campaigns team, they click on that display ad, they browse the products in the e-commerce store, and they add something to their cart.It looks like they have abandoned their cart over time though, and so at a certain point in time, the marketing organization sends an abandoned cart email. The user clicks on that offer and makes a purchase. So how do we assign what should get the credit for the value of that purchase? Should we give it all to the abandoned cart email because if that didn’t happen, the purchase never would have occurred? Or do we assign it all to the display ad because that’s where the user clicked, and we never would have got to the abandoned cart unless the user clicked on that display ad? Or do we share credit equally to both of those areas?But what about the Google search? The user may have never even seen the display ad if they didn’t search for that type of product, so maybe there was some branding already done or, potentially, the SEO team had a very good high ranking that really helped that user click on the display ad by affirming their understanding of your brand. So trying to work out how we assign the credit for a purchase to all of these digital marketing efforts is what attribution is all about—and there are many types of attribution.The attribution challenge that we all have is which attribution model is best for you. If you have a look at some of these models, we can see that there are single-touch models and multi-channel models.In a single-touch model, we look at simple things like the first touch in a journey (that would be the Google search), or we look at the last touch (the abandoned-cart email), or we can look at stage-based single touch, so maybe the lead creation. So how did we get that lead into our e-commerce site? Well, that was by the display ad, so we give all the credit to the lead creation. Or maybe it was the opportunity, maybe it was the abandoned-cart email.Then we have multi-touch options, things like linear attribution where everything gets equal weighting or we provide weightings in certain stages. We can also do that in a partial way or we look at time decay. We work backwards and say, “Well, the most recent things get more credit and the things further back in time get less credit.”So these are the various attribution models that we have and can use —and there are many more and many different ways—but these are the most common.When selecting an attribution type within your organization, it's really important that you choose the model that’s going to drive the behavior for your marketing team. If you are assigning a first-touch approach, then all of those top-of-funnel type activities are going to get all the credit and all the ROI, and so people are going to create more of those types of activities because that’s what you are measuring them on. Therefore, it's really important to understand the behaviors that you want to drive with your marketers and also how you are measuring them. And using attribution models that will ultimately provide marketers with the insight of really what is working, but also helping them drive the behaviors of what’s actually working for our customers, what are our customers doing with this information, and really what’s driving the pipeline. And I think it's really important that we look at those types of models and implement those so that we can really create a performance marketing engine. Thanks for stopping by and see you next Monday.Ever wondered what marketing attribution actually is? In this video, we cover some key components of marketing attribution and how to choose an attribution method that is right for your business.
There are numerous methods of marketing attribution, but they can be split into two main buckets: single-touch attribution and multi-touch attribution.
Single-touch marketing attribution
Single-touch attribution only focuses on one of the touches in the customer journey; this means it’s less credible, but also the easiest to implement. Most CRM and marketing automation systems have single-touch attribution embedded within them, so it is a very common place for many marketers to start.Common single-touch methods include:
Hi and welcome back to Marketing Mondays. Today we're going to be talking about single-touch attribution. If you remember from one of the previous videos, attribution is about measuring a journey and the various touch points along that journey of events. Typically we have a start and an end, and the end is typically a touch.What we're going to be talking about today is what we call single-touch attribution, where we only measure one of the touches along that journey. Some common examples include first-touch attribution—this is where we assign all of the credit to the very first thing that happens in the journey as the first touch. So in this example, if there's a purchase for $10,000, then all of that credit gets attributed to the first touch. This is a really great, simple model for understanding what awareness activities are working really well.Another common example is last-touch attribution. Again, just like first touch, we're assigning all of the credit to the last thing that happens. And this is really good for understanding where dual conversion occurs. So, again, all of the $10,000 credit is attributed to this last touch. Some other examples of single-touch attribution include stage-based attribution—this is where we divide the journey up into stages, and there can be as many stages as you have within your process. And what we do is we look to assign credit to an important stage or conversion in those stages of the journey. Examples include lead creation attribution—what we do here is we assign all of the credit to the stage that created the lead; in this case, that was the last touch prior to stage one. Other examples include opportunity creation—and again, this is where we assign all of the credit to the last touch within this example prior to stage two; so prior to the opportunity being created in a more B2B example.Now, using these methodologies, you're becoming more advanced in first and last touch, and you're actually really measuring the goal of the marketing organization, which might be creating leads or creating opportunities. So what are some of the benefits and some of the drawbacks of single-touch attributions? Well, the benefits include ease of implementation; often, they are already built within your marketing automation or CRM solutions. So it's an easier measurement to start with, and it creates a focus on a key metric. And that's really what the lead creation and opportunity creation models are focused on—those key metrics the marketing team is measured on.Some of the challenges though: It’s not really representative of the buyer’s journey. We have lots of touch points in the buyer’s journey, so it can often be very misleading for those that don’t understand attribution if you're only using a single-touch methodology. For example, if you're only measuring the first touch, people might focus all of their investments in the thing that generates awareness. Or last touch, they only focused their investments on the thing that converts that journey. And so that's often misleading when you're presenting this information to your sales counterparts and other people within your organization as to why you're making other investments if the data doesn’t point that way. So it's really important to make sure that when you're using a single-touch methodology, you don’t forget about all of the other touches in your journey. They're great because they do give you some representation of what's working, but there's lots of other touch points that really should be measured. Thanks for stopping by, and see you soon.
Multi-touch marketing attribution
Multi-touch attribution focuses on touchpoints throughout the journey based on the methodology you are using. The multiple touchpoints ensures the data is more credible and gives marketers more accurate insight into how your marketing touches are contributing to the customer journey.Common multi-touch methodologies include:
Hi and welcome back to marketing Mondays. Today, we're going to be talking about multi-touch attribution. In our last video, we talked about single-touch attribution and the various methods of measuring single-touch attribution. Today, we're really going to cover some multi-touch attribution models, including linear attribution models multi-touch, weighted and partial weighted attribution, as well as time decay and even algorithm-based attribution.Multi-touch attribution is more complex than single-touch attribution—it gives you a better picture of what's going on across your customer journey, and it's also a really effective way of helping you understand that journey and what is having a bigger impact. Now, obviously, these models are different in their complexity, so let’s take you though them.So when we're talking about linear multi-touch attribution, what we're looking at here is the customer journey of a start to a purchase, and you can see this is broken up into a few stages. And with linear attribution model, all we're simply doing is giving every touch an equal weighting. In this example, there are 10 touches, so each of those touches get 10 percent of the credit of whatever that purchase may have been—that works out to be $1,000 per touch. So this is a simplistic way of looking at an end-to-end journey; it's the most simple version of multi-touch attribution and an extension from looking at single touch, such as first or last touch, it's a much more comprehensive model. It often has different names such as blended attribution or mixed attribution or even blended-mixed attribution.Now if we take a step further into multi-touch attribution, we can look at what we call weighted attribution. So what we have here—we entered in journey with those few stages again and what we simply do is start to apply weightings to each of these various stages. So in this example, I'm going to apply a 30 percent weighting to the start and the end of the journey. This particular journey seems to be heavily favored in terms of we're going to overweight things, in terms of how we get people into the funnel, and then we're also going to reward the touches that close that journey.We'll then also give 15 percent of the credit to the items in the middle of the journey and 10 percent to the very center of that journey. So what's the going to look like now is, as we apply that purchase amount, we're going to apply that to the various stages. So you can see that in the first part, we're looking at $3,000, and in the end part, we're looking at another $3,000 dollars. And then we just divide that across the touches. So quite simply, we apply that in terms of, “okay, we have $1,500 for each touch in the first stage of that journey.” That's how weighted attribution works—we divide the journey into stages, we apply a weighting to each of those stages, and then we credit those touch points.You could also start to do things like overweight a particular single touch within that journey where you may, for example, overweight the last touch in the first stage—that's getting a little bit more complex with weighted attribution. We then look at partially weighted attribution; this is where you apply a similar weightings. And what we then do though is remove a stage in a journey because maybe that shouldn’t get any credit for whatever reason—marketing is involved or shouldn’t be involved pass that point. So what we do here is we just apply again—I'm going to apply 40 percent credit to the start, 40 percent to the second stage, and 10 and 10 to the second two stages, and obviously we've used our 100 percent so that's going to leave zero value. And so there's no credit coming those last two touch points.So when we start to divide that up again, we simply see the $4000 being applied to the first two stages, and $1000 to the subsequent second two stages. And finally, we end up with an amount where we have $2000 for each touch point and $5000 for the subsequent touch points in those second two stages. This is where we do partially weighted attribution, and it's a way of looking at a funnel slightly differently and applying those credits.We then also have another model called time-decay attribution. In this example, what we're looking at is over time and decay—which obviously means that it's fading away. If we think about a purchase, most of the credit will go to the last thing that happened in time decay attribution, and it will decrease its attribution overtime. You can do certain things with time decay attribution, like saying “stop giving credit to something at a certain time pass.” So in a B2B sale, that might be saying “don’t give any credit to anything before the last 90 days or three months.” In a B2C, that may even be a couple of days or a week you might not give any credit to anything that happens a week prior. And that's a really important way of looking at time decay attribution, and you can set various methodologies into how those percentages drain overtime.Now the most advanced version of attribution is looking at algorithm-based attribution. This is where many organizations have developed algorithms that look at attribution look at all of the behaviors that occur across their buyers’ journeys and all of their various marketing campaigns. And they create models to give these touches these certain weightings. And with many machine learning and advanced predictive capabilities, what those models can then start to do is start to predict future journeys and also learn over time that, as your attribution changes, the model will change over time as well.A simple example of that could be where we're going to really heavily weight the item before the start of stage two because that's a huge indicator that were likely to get to a purchase. So anytime that happens, we're going to really give a lot of credit to that particular activity, and then subsequently in this example, we're seeing a big reward also for the thing that brings the lead in; you know, bring something into the top of the funnel, and then also the item prior to stage three gets a fair bit of credit.So you can see that this algorithm-based [attribution channel], that's very complex. Some companies are there, but many of the challenges associated with that a really about getting the data. When using multi-touch attribution, the simple answer is it is more complex than single touch but you do get a more accurate picture of what is driving customer behavior and where you should really start to be applying credit—and therefore, helping you understand what touch points you should be really focusing on in terms of your campaign and promotions spend, and what types of content is really working. So that's a quick introduction into multi-touch attribution marketing channels. Thanks for stopping by.
Where should you start?
This depends on your organization’s maturity and access to data. Most companies start with single-touch attribution as it’s the easiest and also requires less complexity from a data perspective. More advanced marketing organizations will have moved on from single-touch and will be trying to advance their capabilities with more sophisticated multi-touch attribution methodologies.Some marketers are even running simultaneous models to answer questions from many perspectives or to view the data from different angles—for example, if you want to know which activities are driving customers into your funnel, then you can view first-touch attribution. In contrast, if you’re making a decision about content that you want to be evergreen throughout the customer journey, then you could view linear multi-touch attribution to identify where you can get the most bang for your buck.Anaplan’s Marketing Performance Management solution can help maximize the ROI of your marketing department by helping manage your budgets and resources, while also providing attribution capabilities to determine how well your activities are performing through multiple attribution methodologies. To learn more, sign up for one of our live demos to see it in action.
Rowan Tonkin, Head of Sales and Marketing Solutions, Anaplan Rowan joined Anaplan after spending the last 10 years helping marketing teams deliver great marketing campaigns through better planning and operations management. He has experience across many industries, including working with clients such as Shell, PayPal, Telefonica O2, SportsDirect.com and many more. Rowan has a broad background working across the customer lifecycle from sales through to implementation, consulting, and customer success.