The majority of consumers will interact with a brand or product more than once before buying. Because of this, marketers have the opportunity to track and measure these interactions and decide just how valuable they are in driving sales.
One of the ways to do this is through attribution modeling. The general idea is that, by measuring the timing and sequence with which customers find, interact with and eventually purchase a product, advertisers can better understand how various marketing investments contribute to conversions and which strategies or combination of strategies are the most effective. Using this data, marketing efforts and ROI can be optimized over time.
The Problem with Traditional Attribution Models
Historically, the problem with attribution models has been that they don’t paint the full picture. Because they segment attributes and assign value to given user actions at only the single channel level (paid search, social media, direct traffic, organic search, etc.), classical models often lack the ability to simultaneously compare various models across a multi-channel format.
For example, First Touch Attribution Models attribute a single channel — a Facebook ad, for example — as being the one that first puts a given number of leads in touch with a brand or product. Conversely, Last Touch Attribution Models give all the credit to the last channel a lead went through before converting — for example, a landing page where a customer made a purchase.
Traditional attribution models like First Touch or Last Touch focus on single channels, and there’s a significant problem with that: modern consumers use many channels and multiple devices in order to make a purchase. Furthermore, classical models and even newer multi-channel models don’t take into consideration the personas of the customers who are interacting with a brand or product. Who is visiting your digital assets? How are they different or similar to others in the marketplace?
Get the Full Picture
What marketers need today is the predictive power that comes from identifying true cause and effect. They need the ability to capture, analyze and append data and make marketing decisions based off the multi-layered interrelationships between channels. Classical models ignore these relationships.
A truly robust attribution modeling strategy should paint the full, multi-channel, multi-touch, multi-journey and multi-consumer picture.
It should answer the following questions:
Who is your customer?
Knowing who is interacting with your product is the first step in determining how to talk to them. Does your product resonate with more than one audience? Is it bringing in a different audience than you expected based off your targeting? Is your website traffic made up of males 18-35, but sales made up of males 31-35? Are SUV drivers with high household incomes coming to your site? How old are they, how much do they make, and what are they currently in the market for? These are just a few examples of what you can know about your customers with data science and all-touch attribution.
Where is your customer finding you?
Is Facebook the main source of where people are finding your brand or product? Or is it paid search? Was the retargeting email successful in getting customers back to your site? Perhaps it was a TV commercial that initially piqued their interest and a display ad that brought them back a second time. Getting a picture of the channel mix your customers are using to interact with your product lets you know what investments in which channels are worthwhile.
How is your customer interacting with you?
What part of your website is the most helpful, i.e. pushes customers further down the funnel? What part is most confusing, i.e. drives people to leave? Which particular page indicates that the customer is about to buy? What assets do your customers view and at what part of the journey? The answers to these questions help you define your typical customer’s journey to purchasing your product.
What can you do with this information?
With the insight that comes out of an all-touch, data-driven attribution model, you will be able to answer questions like, is my marketing driving enough sales to justify the investment? Should I be bidding more or less on my Facebook ads? How about Google? Am I sending the right message to the right audience? Am I missing an opportunity based on the type of person visiting my site but not purchasing anything? Should I retarget a person who has viewed a certain page on my website? What should my communication strategy be for my sales team?
An all-touch, data-driven attribution model affords deep insight into your customers’ behavior surrounding your brand or product. On a multi-touch, multi-channel and multi-persona scale, you can prioritize, test and learn where your marketing investment is worthwhile and where it falls short. Strong data collection and analysis tools are essential in this type of modeling in order to track customer web activity and identify the personas of your ideal customers.
Want to learn more about how all-touch, data-driven attribution works? Contact us today.