Most companies these days market to customers via a multitude of media and campaigns. For many of these multichannel marketers, customers can potentially place orders from two or three strong channels: phone, Web and retail. On top of that, there often are several marketing campaigns still active at the purchase date. The challenge is attributing the revenue to the marketing campaign that triggered the purchase. Which marketing campaign/channel gets credit for the sale?
Why is it so challenging?
• Every business has multiple marketing campaigns and purchasing channels that overlap, thus confusing the attribution process.
• Every business prioritizes a simple, easy ordering process over the extra complexity of collecting source information at point of sale.
• Many households with two or more buyers have split-channel personalities.
• Many consumers change their channel preferences based upon season, product or other factors.
The result is a growing volume of “unknown source” orders.
Why is this important?
Marketing measurement disciplines begin crumbling with the loss of revenue attribution precision. Incorrect revenue attribution leads to suboptimal circulation decisions—predictive models miss the target audience, and marketing campaigns receive incorrect credit for revenue.
The following 10 recommendations can improve your revenue attribution success. These best practices begin with some fundamental ideas and progressively get more complex, but all will lead to significant benefits.
1. Ask for the marketing source from the buyer at point of sale wherever and however possible. It is the only perfect information you will get.
2. Segment your customers based upon their purchasing channel preferences. While not as effective as marketing channel preferences, it is a good surrogate to begin using in modeling, planning and reporting as it separates your multichannel and single-channel customers.
3. Conduct holdout tests and customer surveys to measure incremental revenue of marketing campaigns. This gives you quantifiable evidence of revenue impacts across marketing channels.
4. Compare your incremental revenue metrics to industry benchmarks. Ask your network or listen to speakers at conferences, since many will freely provide these high-level tidbits of information. For example, what are their companies’ percentages of Internet revenue attributable to direct mail? Each business is unique, but knowing where you compare with others is often insightful.
5. Utilize the results of the holdout tests and customer surveys to make bottom-line revenue adjustments in your marketing reports. This will begin to give proper credit to the triggering marketing channel or contact.
6. Define your own revenue attribution algorithm to credit revenue where it likely was sourced. Begin with commonsense business rules that include contact history, product offered history, shelf-life of marketing contacts and purchase dates.
7. Store the attributed marketing source with the customer order right in your database. This provides processing efficiencies (you only attribute the revenue once) and consistency in modeling, planning and reporting.
8. Incorporate attributed orders and revenue into your campaign predictive models. Response and average order size models should include the attributed orders and revenue in the dependent variable for the model to accurately predict who will purchase from that campaign.
9. Enhance your revenue attribution algorithm with statistical inference models. You can model the known source orders, then apply the model to the unknowns. The inference models will assign a probability that can credit orders fully or partially to marketing campaigns.
10. Store multiple marketing sources with their fractional revenue attributions in your database. Since many of the unknown purchases potentially have several source candidates, proportionally allocate the orders and revenue to those candidates. Again, storage will ensure consistency in modeling, planning and reporting.
Following these best practices will improve your revenue attribution, and subsequently yield better business reporting and circulation decisions. You will want to refine methodologies continually as your business changes.
Remember: While it is impossible to be exact with and difficult to do revenue attribution, the effort is certainly worthwhile.
Randy Erdahl is president and co-founder of Minnetonka, Minn.-based Decision Intelligence, a consulting, analytics and database marketing solutions company. He can be reached at firstname.lastname@example.org.