One of our insurance sector customers recently hosted the annual “Focus Days” Hackathon. In the context of the recent successfully concluded SAP Hybris marketing implementation and integration project, in the course of which classic campaign channels such as e-mail, CRM leads and CRM CallTasks were provided, one topic of perspective should be carefully examined. “The integration of SAP Hybris marketing “Facebook campaigns” supported by predictive analytics.”
The aim initially was to identify customer “lookalikes” within the customer database who already demonstrated successful sale closures for a specific product. These “lookalikes” have a similar profile or behavior. Based on this information it is possible to predict that they are highly likely to purchase similar products. Existing functions on Facebook were not used since the approach involved systematically targeting pre-existing insurance customers. The next step involved approaching the contacts identified using SAP Hybris marketing-controlled Facebook campaigns.
The tasks here have been split up among two teams:
- Set-up of SAP predictive analytics to identify “lookalikes” for a Facebook campaign using manually-triggered processes
- Set-up of Facebook integration and execution of a SAP Hybris marketing “Facebook campaign” based on the target groups identified with predictive analytics
- Advertisement of the Facebook campaign on a Facebook account
In order to comply with the timeframe of 1.5 working days, some preparatory measures were implemented in the run-up:
- Creation of a Facebook campaign including content
- Opening up of technical access between SAP Hybris Marketing and Facebook
- Provision of users and authorizations
The team consisted of marketeers, data modelers, SAP Basis specialists and a SAP Hybris Marketing specialist. The result was surprisingly successful for all of us. Even within the first few days we were able to put together and implement a prediction model, thus providing a target group with potential “lookalikes”. These were then used in the course of a periodic SAP Hybris marketing campaign: The channel Facebook was used to address the target group, clearly assigned to a pre-prepared Facebook campaign.
40 percent of the potential interested customers identified were already known to Facebook on the basis of the phone numbers or e-mail addresses supplied, thus demonstrating the key role played by social media in the insurance sector too. Whether and to what extent advertising via a channel like Facebook in a serious sector like insurance is rewarded by customers, was not dealt with in this step.
For the target group, Facebook requires a critical size of more than 100 to ensure a certain measure of protection against the traceability of communications. This must be taken into account within the context of the forty percent hit rate mentioned above. Facebook does not disclose which users were actually targeted by the Facebook campaign. However, if the post is actively opened, this information is transmitted to SAP Hybris Marketing directly via a landing page for example and made available for follow-on campaigns.
One small added obstacle was: Facebook prioritizes the execution of a campaign over the value (cost) per post. In our case, the minimum amount was selected and the minimum target group size chosen – this resulted in the campaign not being executed for a few hours.
And on the second day, before the end of the Hackathon, our own Facebook test customers had the marketing post on their mobiles.
The topic of privacy and data security were not dealt with any further during the focus days. Since it is an interface based on SAP Hybris Marketing accessing Facebook, the initial assessment was that this could be classed as less critical. The approach to this topic and the motivated team involved turned this into a success and delivered great results. Based on the knowledge obtained, we would certainly recommend this function to others.