How We Track Sales
Retailers can already identify their customers through email, phone, address, customer ID and/or order ID. If a customer logs into the retailer’s website or app from various devices, the retailer can tie the customer to each of those devices. If the customer never logs in from a particular device, a connection will never be made and the retailer won’t see the buyer’s path to purchase. But if a device cannot be directly (“deterministically”) linked to its owner, retailers can use additional data to “probabilistically” make the connection and tie devices to the appropriate buyer’s journey. SourceKnowledge uses probabilistic attribution to determine if a sale occurred.
Probabilistic Attribution For Tracking Sales
SourceKnowledge uses probabilistic attribution to determine if a sale occurred in our shopping, awareness and retargeting campaigns by finding a match between IP addresses and user agents. SourceKnowledge uses publicly available information to establish probabilistic attribution. For example, when accessing the web, devices routinely communicate with web servers, wi-fi and other machines. In doing so, certain public information becomes available (user agent & IP):
- IP address
- Device make & model
- Operating system
- Browser type & version
- Current and previous URL
- Settings such as language, time zone & geolocation
- Apps and plugins
How We Decide Where to Run Ads
SourceKnowledge uses contextual placements to run ads for Shopping and Domain Targeting campaigns. Our network is divided into three contextual verticals:
- Social to Commerce
- Content to Commerce
- Search to Commerce
We do not use deterministic or probabilistic targeting to place ads for Shopping or Domain Targeting campaigns. We only use probabilistic attribution to determine if a sale occurred.
How we Run Retargeting and Awareness
We use a third party DSP to run retargeting ad awareness campaigns.