Things Machines Do Not Do

Experts have predicted the automation of the online advertising market on several occasions in the past 20 years. However, it has not happened yet.

If we compare the self-service cash registers of bigger stores with the advertising market, we can notice several oddities: in IKEA the register of products and the payment are automatic – you read the bar code, then pay by card -, whereas the monitoring is personal: there are cameras and store personnel watching. In the stores of Tesco on the other hand, there is no need for human control, since an in-built scale monitors whether the weight of the basket is the same before and after payment. It’s a great idea, but those who tried it already know that doing your groceries takes more time this way and it is rather inconvenient as opposed to traditional queuing.

Anita Király, the Media Director of Café Communications used the example above at the Evolution Conference in March to demonstrate the fact that the efficiency of automation differs from field to field.

You can build automation in two ways: you either insert the user data based on the experiences and the consumer habits of previous years, or you trust the results and the profiling of the system, that is based on KPI-related settings. Which one is better? It depends from the product and the service type as well as your objective.

Anita presented the knowhow of profiling people in order to reach them with automated advertisements and promotions via the example of personal loans. Borrowing is typically an impulsive product; it usually takes place out of blue. To set an example: people don’t necessarily search for car components on the Internet, because their car broke down. With that in mind, it is fair to say that coincidence plays a bigger role than it could be profiled. In terms of consumer habits, it is important to consider the time of borrowing. The tendency of people to borrow money changes by season and even by month. It is not a surprise that financial institutions receive more loan requests at the end of the month and it is also important whether it is a new or a private client, because the later can be reached via other channels as well.

Still, how can you make automation work?’ you might wonder. Anita believes that you have to test convictions and offline experiences and if they turn out to be true, you can set automation. In the case of borrowing the semi-automated mode works the best, provided that you keep yourself to the following advice: you follow users till the signing of the contract instead of a lead; you target them manually, and you launch numerous campaigns with different strategies, aiming at various target groups.