Elder Research collaborated with a leading financial services company to develop a framework for measuring the effectiveness of their marketing email campaigns. In many sales cycles, customers engage with marketing and sales simultaneously. But how much impact does each have on the final purchase? Our framework used marketing attribution models to correlate sales with different customer touchpoints, helping marketing teams better understand engagement patterns and the effectiveness of email campaigns.
The Challenge
This client offers a wide range of financial products, nurturing leads through a variety of sales and marketing interactions. As a customer moves toward a purchase, they can engage with many different marketing materials and sales representatives. The intricacy and variety of these interactions made it difficult to estimate how each touchpoint contributed to a final sale and prevented the marketing team from obtaining a clear view into the value of their efforts.
The Solution
To tackle the complexity in sales attribution across the customer journey, Elder Research developed a framework that estimated the portions of a sale that were attributable to sales and to marketing, with the goal of helping to measure email campaign effectiveness.
A causal approach was not possible in this case, so our framework incorporated an ensemble of attribution models that cleanly correlated sales with various interactions throughout the customer journey. The solution we delivered is dynamic and customizable, allowing stakeholders to view sales estimates across different timeframes, products, and touchpoints.
The final attribution framework incorporates four distinct models, each offering a unique perspective on sales attribution.
1. Last touch: This model assigns the full sales value to the sales or marketing interaction that occurred immediately prior to the sale. It operates on the intuition that the final interaction is more likely the one that directly influenced the customer’s decision to purchase.
2. First touch: In contrast to the last-touch model, this one attributes the entire sales value to the initial sales or marketing interaction in a customer’s journey. The assumption is that the first interaction led directly to the sale, even if the customer received marketing communications and the sale occurred later.
3. Linear: This model spreads the sales value equally across all the interactions that occurred during the customer’s journey. It reflects the view that each touchpoint contributed to the final sale, and we cannot distinguish among these effects.
4. Markov chain: This model estimates the importance of each interaction by calculating the effect of removing it from the customer journey. This model encodes the basic intuition that some interactions are more likely to correlate with sales than others, so calculating a “removal effect” provides a measure of importance.
Together, these models provide a diverse and flexible way of attributing sales to various interactions. We also delivered an overall ensemble model that provides a balanced and robust attribution estimate, incorporating these four diverse perspectives.
The Results
We demonstrated this new framework for the client by applying it to measure the impact of a selected marketing email campaign, providing new estimates of the campaign’s total attributed sales in the previous calendar year from several different perspectives.
Our solution gives the client a better understanding of how various marketing and sales interactions drive sales, offering insights into the specific contributions of each touchpoint and of the portions of the sale attributable to marketing and sales efforts.
By working closely with our client to integrate their expertise with customer journeys into the attribution framework, we created a customizable solution enabling marketing stakeholders to attribute sales to future campaigns.
Example:
Consider the following set of three customer journeys leading to purchases. They mix sales and marketing interactions in different ways on the path to the purchase of a financial product:
Given the various customer journeys above that lead to a sale, we can attribute the following touchpoints to the final sale using last-touch attribution:
Similarly, we can look at first-touch attribution:
