The birth of ecommerce, alongside the rapid uptake of technology such as smartphones and tablets, has significantly changed how consumers browse, consider and make purchases. This has made marketing more challenging for marketers in many ways; people now have much more choice over what and who they engage with.
However, for many marketers, it's also made the profession more of a fine art. Now, you're able to get information about your target audience and design campaigns that speak to them directly. For each campaign, you can also collect data to find out more about the people engaging with your brand and adapt future efforts. Attribution modeling is just one of the ways you can gather information about the users engaging with your marketing.
What is attribution modeling?
Attribution modeling is the process of assigning a value to various user touch-points that lead to a conversion or purchase. This allows marketers to understand more about the visitors coming to their website and look at the finer details of what may have encouraged or discouraged them.
It also means they can optimize certain elements in paid advertising that are proving to be more effective to make the most of their investment. Not only does this help make sense of what device consumers are using to engage with brands but it gives information on the channels they use too.
People rarely go directly to a website and immediately make a purchase. The user journey is much more complicated than that and attribution tools allow marketers to get a much clearer sense of what is influencing audiences. More importantly, they also indicate what has led to the decision to make a purchase.
How does it work?
If you're already using Google Adwords for paid search, it will automatically give you the option to view Last-click attribution. This can be misleading because it gives all the credit to the last thing the user did, rather than all the micro-conversions that led to that point.
There are many different approaches to marketing attribution, which range from basic models that focus on a single factor to much more advanced practices that incorporate a number of different ones. The most appropriate solution for any particular business will depend on budget and how much you are willing to invest but it also depends on what your marketing objectives are.
For example, if you want to know whether social media is leading visitors to your website or if your blog content is encouraging readers to make a purchase, then one of the more basic options will suffice. However, if you care about how effective any single advertisement is being, regardless of the channel, then you are likely to need a more advanced solution.
Here are some of the most commonly used:
Single Source Attribution
Much like Google Adwords, single source models attribute all credit to a single action. This is usually the first or last touchpoint a user takes before converting. For many marketers, it doesn't give enough information to make an accurate judgement and influence campaigns. Like any method, it has its advantages and disadvantages:
This is one of the easiest to implement but can be misleading because it doesn't take into account any subsequent actions taken by the consumer. For example, they may enter your website via Twitter or download a whitepaper from your site but they then go on to talk with a Live Chat agent about their options. First-touch attribution would give all the credit for the purchase to the first touch, regardless of the impact of the conversation on Live Chat.
In this case, it's the last action that gets the credit. Whether it's a CTA at the bottom of a blog post or a call with your sales team, this still doesn't give you the whole picture of what led to the purchase. They may have visited your site a number of times previously or they saw your brand via a social influencer.
There are now a variety of ways that brands can engage with consumers on social media alone, which is the reason why many companies are turning away from single source attribution in search of something a little more detailed.
Each channel that leads to a conversion is given credit in multi-touch attribution (MTA), giving marketers a much clearer view of the overall consumer journey, compared to just a snapshot at the beginning or end. So whether someone sees you via a webinar or a Twitter ad, you'll be able to clearly see where your engagement started.
However, MTA has its own cons too. With a variety of different models, it can be complicated for even an experienced marketer to set up and implement properly. It also doesn't assign any proportional value to the touchpoints, making it difficult to see which of your marketing efforts are engaging the best with consumers.
The spectrum of MTA models available means that brands can find a solution that suits them best, but can make it more time-consuming to introduce this method and get real-time data.
Linear models are the simplest but give equal revenue credit to all touchpoints the user has made along their journey, regardless of how much time they spent on each. A U-shaped or W-shaped method, on the other hand, both give credit to the first touch and lead creation. In a U-shaped model 40% of the revenue is attributed to first touch and lead creation, while the remaining 20% is divided between any other touches. A W-shaped method adds a third touchpoint - opportunity creation - and gives 30% to each of them, with the leftover 10% being attributed to others.
Brands also have the option to adopt full path attribution, giving most of the credit to the major touchpoints in the journey and the remaining credit is assigned to the other touches. Custom models can be created to allow companies to decide exactly how much credit they want to give each touchpoint. This is the most accurate but can be very time consuming and expensive for many businesses.
Weighted Multi-Source Attribution
A weighted multi-source attribution model assigns the most credit to the touchpoint that had the biggest impact on the eventual sale. Again, this gives an accurate picture of the consumer journey in principle but can be much harder to implement and get to a point where it works for you.
To get the most out of weighted multi-source attribution you need to already have a good understanding of your target audience and how they interact with your brand. Applying the weight to certain interactions over others won't prove successful if you don't already have an intermediate knowledge of how your most valuable customers are engaging with you. Of course, you can achieve this through trial-and-error but it can be extremely frustrating and expensive.
How do I find the right method?
Unfortunately there's no foolproof recipe for success when it comes to attribution modeling but whichever option you go for needs to align with your business and marketing objectives. It can be a difficult process finding the right model that works for you but there are clear advantages if you can implement one that complements your wider goals and is attainable for your business.
With consumers increasingly using different channels and devices to engage with brands at different stages of their buyer cycle, there's a real benefit to having some sort of model in place.
The trick to success
Although the more advanced models try and give the most credit to the touchpoints that deserve it, this is always a predefined point and so can't be 100% accurate for each visitor. Social media can also suffer in an attribution model as it's often part of the user journey but rarely the first or last step before conversion. This could be misleading and, as such, attribution modeling shouldn't be judged in a vacuum and should be an element in a wider data-driven marketing campaign.