Keynote speech at the bevh (German e-commerce & mail order association): Artificial Intelligence in the service of B2B new customer acquisition
In mid-November 2017, Dr. Parsis Dastani showed members of the bevh B2B working group how they can employ predictions based on AI (Artificial Intelligence) to identify lucrative prospective customers while cutting acquisition and sales costs. Here are the approaches in detail.
How can we grow and how can we best exploit the market? These questions occupy the entire online business, including the B2B area. Sales and marketing functions are confronted with demanding challenges: When acquiring new customers, they should concentrate on those companies with the highest affinity and the highest potential. Existing customers need to be developed. For this to succeed, the market potential must be determined, the market segmented correctly and market participants individually addressed. We have developed applications that “take over” these tasks, increasing both sales effectiveness and efficiency.
Target Group Predict
This process takes advantage of the fact that almost all companies have an online presence, for example through their own website, in market overviews, on platforms and in social media. An AI application learns which words are sector-relevant in existing customers’ websites and are important for purchasing affinity. A neural network then evaluates the words and determines the likelihood that a company is actually a customer. Web crawling and the identification of the “best” buzzwords works for all types of products – from forklift trucks to cars and vans for the company fleet or tools and building materials. With Target Group Predict, the sales function can not only determine revenue potential in percent, but also which product suits the customer best – ideal conditions for successfully addressing new customers.
Mobile field sales contro
This also works on smartphones and tablets. Analysing a tool and building supplies customer, we examined approximately 1.2 million customer visits using Artificial Intelligence. This enabled us to answer the following questions:
• Which company will purchase?
• What is the purchasing probability?
• In which product or products is the company most interested?
The sales representatives receive this information via a location-based app so they can address the prospective customers with the highest sales potential in their sales area.
For the same client, we have ranked the entire customer base by assigning an individual revenue potential to each customer. Subsequently, these were segmented into A, B and C customers in order to assign them to a sales channel (key account, account, telesales). Thus, the AI analysis enabled us to correct the previously existing scheme of allocation.
Existing customer ratings
Using the same method for a steel distributor, we analysed the information available to us from Creditreform (German credit agency) market data and the German Federal Gazette as well as the proven process of website crawling. The guiding question was: Which customers in the mass of C customers show a high acquisition potential based on their overall potential? This method significantly expanded the proportion of A customers with a corresponding reduction of C customers.
Next Best Offer
Taking buyers of particular forklift truck models, we investigated which product they will be buying next and with which probability. With this project we were also able to significantly improve the way in which the customers were addressed. The client was able to deploy his marketing and sales budget in a more targeted way, resulting in increased sales.