Better response thanks to customized mailings

SAS improves marketing relevance at Overtoom

Increasing customer response rates while keeping marketing budgets under control is a challenge for any sales organization. Overtoom International, market leader in office and warehouse supplies, deployed SAS to improve their marketing efficiency. Python Predictions developed a SAS model that enables the customization of Overtoom mailings. The marketing campaigns are now 300 percent more relevant and significantly more successful, showing a higher success rate and return on investment. And the new system enables Overtoom to continuously improve its offering.

In the past, Overtoom International always sent full product catalogues to their customer base once a year. In addition, they sent a monthly leaflet with bestsellers and seasonal products to a substantial number of their customers. In 2009, they improved the profitability of their mail campaign thanks to a more fine-tuned selection of targets. This selection was based on SAS technology, deployed by preferred SAS partner Python Predictions. Overtoom Managing Director Ghislaine Duymelings explains: "Python Predictions encouraged us to be more selective and choose targets based on a large set of customer qualification variables. This was a first step, but we decided to go further. We figured that our campaigns could be even more successful if we managed to fully customize them."

Models for customizing campaigns

Python Predictions helped find the best way to customize Overtoom's campaigns. They examined three different analytical methods:

• Market basket analysis — a method which checks the customer's typical purchases and proposes other products based on associations.
• Response modeling — a more sophisticated method that analyses a customer's future propensity to buy each product category, based on historical purchase data of all customers.
• Similarity modeling — a method which compares customers and offers products that are popular with customers having a similar purchase behavior. Customer similarity is established using a vast set of purchase indicators on the customer contact level.

Similarity model increases response rates significantly

Tests revealed that market basket analysis was ineffective in this context. However, the more complex response models and similarity method both proved to be successful. The similarity model had some clear advantages over the response model. Dr. Wouter Buckinx of Python Predictions explains: "The model is customer-based, not product-based, which is very useful when it comes to customizing campaigns. It was far less time-consuming, while it includes all products and product categories, not just a sample. And finally: both the response models and the similarity model result in a large variety of different offers. This is a first indication that the offers meet different individual needs."

Python Predictions also calculated response rates as a function of the number of personalized products in the offer. It turned out that it is best to limit the offers to the most relevant products or product categories as determined by the model.

Reaching the right customer with the right products

Based on these findings, Overtoom eventually decided to reorganize their mailings. They now include a digitally printed cover for their leaflet each month. The cover highlights a limited number of customer-independent products (such as bestsellers, seasonal products, or a gift) and a selection of customer-specific products, which the model identified as having high success potential. In the near future also the weekly email will be entirely personalized. It will highlight special limited time offers within categories that are popular with the customer contact at hand.

The benefits are clear. Overtoom now has increased its revenue on mailings with more than ten percent. The mailings now include products that are up to 300 percent more relevant for the customer, compared to the relevance of the average product in the monthly leaflet. Thanks to SAS, they also have more insight into customer profiles and priorities. This in turn helps improve product assortments. Duymelings notes that the system enables Overtoom to improve continuously so that they can reach the right customer by offering the right products. In its next step, Overtoom plans to optimize the use of different marketing channels.

Copyright © SAS Institute Inc. All Rights Reserved.