Our offline conversion trap

Our offline conversion trap

Problem

Most stores cannot link internet traffic and physical offline purchases. After a customer makes an offline purchase, ordinary analytics counters cannot determine the traffic channel - where exactly the customer came from. We know a case where half of the clients firstly looked at an item in an online store and then came to see and buy the same item in a physical offline store. The business owner was convinced that the online and the offline stores are not connected in any way :)

Solution

We always turn on cookies on the website in order to track the actions of each user. We also set up events - any positive actions - clicking on the phone, visiting the contact page, adding to the cart, going to social networks, scrolling to the bottom of the page, etc. All this is done for deeper traffic analysis and creation of a machine learning model that can predict the purchase and give out the factors that most influence the purchase. Knowing these factors, we can significantly increase the level of marketing and page conversion. You can see how we did it in 2019 here.

Since we are writing down cookies, we know the unique number of the client's browser. Thus, we just need to make it so that offline buyers click on a special link that will be given only to them. After that, our end-to-end analytics system will determine its first connection and associate it with the traffic channel. This link can be sent via SMS with any offer - get a discount, confirm your bonus account, here's a gift for you, etc.

Instagram and other social networks should be noted separately in this story. They use their own browser inside the application, so this method may not work. But they have internal services that allow you to track offline sales. For them, it is enough to send the customer's phone number or email, and they will associate it with the advertising campaigns inside. Our marketing server can connect to your CRM system and send this data to social networks automatically.

Recommended cases