Business Magazine

Why Customer-Level Data is Key to Attributing Revenue

Posted on the 02 February 2012 by Tchu @UpStreamMPM

attribution_optimizationMarketers face the challenge of understanding how spending on marketing affects conversion.
Techniques for attributing revenue, such as marketing mix models, came from the CPG world as a way to measure marketing effectiveness. As the number of marketing channels increases, traditional methodologies no longer work. The blockbuster limitation: they are built on aggregate data.

Why is aggregate data a problem for attributing revenue?

By definition, using event data (ie. opening an email, receiving a catalog, purchase, browsing the website) requires the ability to capture all of the events that occur for each person. By aggregating the data, you can’t see who received which marketing treatments.

For example, let’s assume that these events occurred last month:

2 million catalogs

1.3 million emails

600K paid searches

100K display impressions

50K affiliate clicks

By aggregating the data, you won’t know that John received a catalog, 2 emails, 3 display impressions and then made a purchase, while Jennifer went directly to the website and purchased without any marketing touches. Aggregate data would assume an average of the marketing treatments for John and Jennifer. Conclusions drawn from aggregate data can be incorrect.

What are the benefits for using customer-level data for revenue attribution?

Two main benefits:

1 - Customer-level revenue attribution is accurate

A holistic view of a customer incorporates each customer’s previous purchase behavior, each marketing treatment a person is exposed to, the decay of a treatment’s impact over time, the person’s digital footprint, and the customer’s proximity to the trade area. By creating this view, you can quantify the value of a single treatment among multiple marketing campaigns.

2 – Customer-level revenue attribution is actionable

Identifying and attributing the revenue drivers is helpful, but applying that information and making it actionable is what marketers crave. Customer-level data allows for response models which predict each customer’s likelihood to purchase as well as the sensitivity of a customer to the next marketing treatment. Does this customer need the catalog to purchase? Optimizing each customer in this way – identifying the most profitable channels for each customer and the most profitable customers for each channel – can revolutionize how marketers think about planning and ROI.

Take the next step and learn more. 


Back to Featured Articles on Logo Paperblog