en

SUCCESS STORY

How to qualify half of your database in a few clicks?

Segmentation of the active database on the clicked areas

The challenge

PURINA® is a leading pet nutrition company on the French market, with several brands such as PURINA ONE®, FELIX®, FRISKIES®, DIFO®, PROPLAN®, GOURMET®…

Over the years, Purina has succeeded in unifying consumers that can be contacted by email within a qualitative database.

Purina’s challenge in CRM is to develop engagement and maintain the interest of its customers by sending the right content to the right person at the right frequency.

The solution

Personal preference data collected on customers’ can be complex to collect and may not cover 100% of the database. Numberly wanted to supplement PII data with behavioral data, in order to segment the database by preferences of customers.

Site navigation data, opens, clicks, and click zones are also rich behavioral data that can be activated by a brand. By leveraging this type of data, it allows advertisers to collect reliable data on engaged consumers: a simple, efficient and reliable methodology.

The strategy had to meet several criteria in order to best meet the needs of the Purina teams, it needed to be:

  • Simple: a comprehensive understanding of the themes studied for a simple reading of the results
  • Efficient and cost-effective: a manual (human) method allowing a perfect understanding of the segmentation rules with less expensive development than an algorithm
  • Flexible and scalable: the possibility of evolving the themes of interest and the variables taken into account
  • Streamlined Process: limiting the time spent by Purina teams on a key project such as segmentation

The operational deployment of this clustering took place in 5 steps:

  • Categorization of emails: subject line and click zones
  • Eligibility of preferences by interest: determine preferences for which there are enough content and a significant audience volume
  • Definition of appetence rules: analysis and modeling of rules defining an appetence to a theme according to opening and clicking behaviors on a theme. Declarative data is used to refine and validate the relevance of the appetence rules.
  • Application of attribution rules
  • Personalization of emailings according to preference clusters

Results

As a result of this preference analysis, the Numberly teams were able to define 13 distinct themes.

They were able to qualify over 50% of Purina’s active database. Today, this allows Numberly to refine the targeting of its email campaigns and make them more qualitative thanks to highly personalized and targeted communications that respond to consumers’ detected interests.

An incremental measurement protocol was put in place to monitor the performance of Purina’s new relationship strategy. Using test and control populations, Numberly measured the impact of segmented communications vs. non-segmented communications.

The impact on performance is clear:
Up to +20pt in open rate on segmented targets and +2pts in reactivity. It also allowed for inactive customers to be re-targeted and once again be engaged by Purina.

The next step is to develop clustering using the data gathered (through use of coupons) on customers and their purchasing of various Purina brands.

Key figures

50 %

of Purina's active database is qualified

+20 points

of average opening rate on email campaigns

+2 points

of average reactivity on email campaigns