Marketing Attribution: Is there really a difference?
Using big data for attribution has become increasingly important for data driven organizations to understand what’s working in their marketing lineup. Several methodologies have been used over the years to measure and give credit to marketing actions. Understanding the differences and impact of each methodology is critical to choosing the most accurate methodology for your business.
So what does that mean in real life? Let’s take an example using a $100 purchase and see how each of these methodologies would attribute the sale:
Last-Touch (Last Click): Current de-facto industry standard; this method gives full credit for the sale to the last marketing channel to “touch” the customer. Last Touch has the advantage of simplicity, but even marketing planners who employ it lack faith in the conclusions. Favors most frequent touches and late-in-the-funnel actions such as search and affiliate.
First Click: Gives full credit to the first marketing channel to touch the customer. May be valuable for identifying customer acquisition opportunities, but does not consider later funnel activity.
Double Counting: Every marketing channel that had recent contact with a customer before their purchase takes credit for all or some of the sale; in most cases the sum of sales claimed by all marketing channels exceeds 100%. This is common when credit is given using a combination of media vendor reports that have no way of understanding other customer interactions.
Uniform: Each marketing channel that has touched the customer takes equal credit for the sale. Although diplomatic, this method assumes all channels and customer touches are equally effective (which they’re not!).
Weighted (Arbitrary Parameters): Many organizations use business logic to weight first touch (customer acquisition), last touch (closing), and intermediate touches disproportionately. While this appeals to our intuition, the different weights are assigned arbitrarily.
Exponential: This method uses the importance of time to value the most recent interaction with the customer as the most influential.
Advanced Attribution: This new method uses techniques borrowed from biostatistics to determine the incremental sales attributable to each marketing channel. It uses consumer buying intelligence to account for time as well as both the relative influence and frequency of each marketing channel for each person.
In biostatistics, positive and negative risk factors include medical treatments (such as frequency and dosage of medication) and health risks, and the event being measured is illness/death. In a marketing world, the factors are marketing treatments, and the outcome being measured is customer conversion/purchase. In both cases, factored in are the types, frequency, and recency of various treatments, and identify the incremental difference in outcome by comparing the statistical buying behavior of groups that have experienced different “doses” of marketing.
Particularly helpful and unique is the ability to measure the incremental impact of marketing. In the same way that health risks are evaluated for medical treatments, so are non-marketing treatments for purchase behavior. In the Advanced Attribution example, brand loyalty is credited with $18 of the sale because this particular customer has a history of purchasing the latest movie releases on DVD each week. Quantifying the effect of brand equity helps marketers understand the actual impact of their marketing efforts and ROI.
Brand equity, or customer driven sales, can be a significant portion for established brands. Customer driven sales can be seen in everything from repeat purchases to trademarked or branded natural search, and direct load to the website. One way to think about it is that if you were to turn off all of your marketing today, you would continue to see a portion of your sales tomorrow, next week, and next month. This is your brand equity and loyal customer base.
Examples of other non-marketing treatments that should be considered are: calendar effects (holiday season), trade area (store locations), promotional effect (separate out the impact of the marketing vehicle such as an email, from the 50% discount), customer events (VIP store event).
Holy cow! Those are big differences… now what?
The actual differences will undoubtedly depend on an organization’s marketing mix, how established a brand is, type of product offering, and level of promotion. The important thing to understand is how different each method is for measuring marketing effectiveness. One of the main objectives of attributing revenue is to allow marketers to fund channels appropriately and optimize return. Using each of the methods illustrated above would lead to very different marketing budgets, and drastically different sales.
Learn more about using customer buying intelligence for advanced revenue attribution.