New customers sales
The essence of machine learning methods is the possibility of automatic processing of large amounts of data as well as identifying patterns and dependencies.
These are quite powerful tools, which are supposed to really change the world for the better in a number of areas.
However, the hype around this topic somewhat spoils the impression of the technology on the whole.
Usually the reasons are:
- artificial intelligence solves problems that can be solved with multiple Excel spreadsheets;
- simple classical algorithms are called artificial intelligence;
- machine learning is used without analyzing business processes and therefore provides inaccurate results.
All of these reduce the credibility for the technology, and its specialists are often viewed with great skepticism.
But true machine learning tools, used on time, can significantly improve performance and reduce errors.
Let’s see who can benefit from ML in sales process?
An experienced specialist “feels” the right decision: a client to give more attention or the one to offer a discount, and the one who’s a “difficult” case.
Usually, we call it intuition. In reality it’s just experience that shows up in better deals. The problem is experience’s hard to describe in simple words, clear rules or algorithms.
A mentor can't be assigned to each young specialist for a long time it's too expensive.
But if each “fresher” has a mentor with experience in the entire history of transactions both successful and unsuccessful he/she has got experience of the entire company. It means making less mistakes and taking decisions faster.
A machine learning model can become such mentor.
After processing the data about all transactions made in the company during its history, including information about clients, circumstances and product parameters, the model will be able to recommend the best possible solution to a new employee and indicate which transaction parameters should be paid special attention to.
Sounds complicated? Here’s an example.
One of our clients (a large construction equipment dealer) implemented a recommendation system for its sales department to rank clients according to the degree of readiness to order.
For data, we used information about the company's existing customers, previously made offers (positive and rejected), information about potential customers (from open sources and internal company data).
Previously, employees called in an order close to alphabetical, that is, almost randomly. Since the list consisted of several thousand potential customers, the effectiveness of calls was very low (only 20% of companies managed to call), there was a high negative, from the receiving side of the calls, there was a rapid burnout of employees. After the introduction of the recommendation system, the nature of the work changed.
Each employee received a ranked list of companies to call and the likelihood that the client would buy construction equipment. They also saw why this customer was more likely to complete the deal than another. This solved three problems:
1. Employees knew what to offer, that is, they spent less time on selecting an offer, and quickly found contact with a potential client;
2. With the same number of contacts, the first calls were made to those who were more likely to make a deal. That is, the companies remained untreated, which most likely would not buy anything;
3. Learned faster, because they understood what to pay attention to when communicating with clients.
In continuation of the project, this system was extended to sales of service and spare parts to current customers. There are much more opportunities there, but we will talk about this in one of the following articles.
In b2c segments, such systems can also work effectively. If we are talking about online trading, then the signs by which the user will be evaluated (whether he needs to be given a discount and what product to offer in addition) can be viewed products or the number of returns to the site, the phone model from which he entered the site, the city, in which it is located, etc.
For experienced professionals
The market changes over time and what worked before may not be as effective in the future. That is, many years of experience is not always a guarantee of success.
For experienced employees, a tool that would show complex or implicit patterns and help generate and test hypotheses in a semi-automatic mode would be useful.
In particular, the above system for assistance in sales of construction equipment in a few clicks made it possible to test hypotheses about the impact of the cost of raw materials on sales of equipment, to make forecasts on sales by region, depending on aggregated indicators by region. Sales managers received a tool to forecast sales by region and industry, as well as identify the reasons (not always obvious) deviations from the plan.
In this article, we've looked at how AI can help you sell goods and services to new customers.
The mechanism lies in the fact that the system, having data on previous successful and unsuccessful offers, helps managers to make a personalized commercial offer.
For ordinary employees, it is an assistant that saves time and resources.
For experienced professionals and managers - a deep analytics tool that helps you back your decisions with big data insights.