NEC Networks & System Integration Corporation - Supporting Sales With Process Mining and Cluster An

JUNKI_TOYOKAWA Dataiku DSS Core Designer, Dataiku DSS ML Practitioner, Dataiku DSS Adv Designer, Registered Posts: 1 ✭✭

Team members:

  • Junki Toyokawo
  • Ryotaro Suzuki
  • Ichikawa Ginji
  • Takuya Minegishi
  • Yumi Miyai
  • Yueming Zhu

Country: Japan

Organization: NEC Networks & System Integration Corporation

NEC Networks & System Integration provides a wide range of services, from ICT and networking to installation and maintenance, to a variety of customers, including corporations, telecommunications carriers, government agencies, and social infrastructure providers.

Awards Categories:

  • Best Acceleration Use Case
  • Best Moonshot Use Case

Business Challenge:

Our company's main business is the design, construction, and installation of PBX and networks for both private and public customers in Japan. In recent years, we have also focused on SaaS resale, SI, and operation business, including the first resale of Zoom, a web conferencing tool, in Japan.

As there have recently been developments in various industry areas with SaaS, our sales organization structure is industry-specific. Our company is engaged in a variety of sales activities, including Zoom, SaaS resale, SI, and operations, but the complexity of sales methods and flows for the products and organizations makes it difficult to make quantitative recommendations because sales methods are not standardized.

The organization is subdivided, and each employee is in charge conducts business negotiations for specialized industries, so it is necessary to recommend the most appropriate business negotiation method for each person or to assign staff accordingly. Under such circumstances, it is now urgent for us to visualize the negotiation process in the sales business area, while reexamining the allocation of human resources and formulating GAP measures for personnel evaluation in a logical manner.

In the past, each sales representative in our sales department had to conduct business negotiations with customers independently, even though there were a wide variety of customers and price ranges.

However, these presented several challenges:

  • The process of acquiring sales from customers is not standardized.
  • Failure to evaluate personnel quantitatively on sales activities.
  • Staffing is not optimized according to sales representatives.

To address these issues, we utilized the following models:

  • Business transition process mining.
  • Cluster model by number of business talks by value.
  • Reports utilizing process mining and cluster models.

Business Solution:

Dataiku enables data preprocessing, feature engineering, machine learning, and model monitoring in a single application. Data preprocessing was handled by Dataiku recipes, while Business Solutions was used for the learning model's predictive output method and process mining.

The project was undertaken by 10 people, including two recipe creators and a service planning advisor (including the sales department). In addition, we worked with the sales department, the end user, to address the following two issues:

1. Process mining of business meetings registered in SFA

Process mining is performed by creating transition pattern items according to time series from the business meeting classification entered in SFA. The 10 most common process patterns were used as indicators. The first process mining took only about a month to create. Simply presenting the results of process mining could not lead to the end user's next action, so we made several innovations.

  • Display model process user results

Visualization of the process was conducted by interviewing sales representatives regarding the personnel who should be used as models and narrowing down the target audience. This will clarify the negotiation process that should be used as a reference. In particular, we found that sales activities tend to receive orders through a process that involves a lot of movement and negotiation of amounts and delivery dates.

  • Display process mining results by sales team

Because of the wide variety of customer industries that each sales team negotiates with, it will be possible to recommend a negotiation process that is tailored to each industry. Since we found that there is a bias in the sales organization depending on the process type, we have clarified the process that is suitable for each organization.

2. Cluster analysis of deals registered with SFDC

Since business negotiations include high-value projects and low-value projects, the amount, range, and number of projects handled varies greatly depending on the person in charge. Cluster to separate levels of variability.

  • Create a BIN and map and visualize the amount of revenue of the person in charge, the number of cases, and the number of customers handled

This will inform how busy the sales representative is, the price range and number of projects that can be handled.

  • View the distribution of cluster results by organization

By displaying the information for each organization, it can be used as reference information when arranging for the person in charge to be more active. Nearly 30 recipes and 1 model were designed into the design node after data cleansing in Snowflake.

Day-to-day Change:

Benefits for the sales department include:

  • Being able to visually check the business negotiation flow according to the industry as a reference standard flow. This can be used as reference information by confirming the negotiation process of the person in charge of the model.
  • Being able to check the standing position of the person in charge according to the price range of the project, the number of projects, and the number of customers. It is now possible to use clusters as reference information as one of the means of arranging for each sales team.
  • As evaluation indices, clusters and process results can now be used as one of the quantitative evaluation means in addition to reference information.

Business Area: Internal Operations

Use Case Stage: Proof of Concept

Value Generated:

This solution makes our sales activities, personnel evaluations and staffing smarter and more efficient. By using sales activities in the education of new employees and mid-career employees, I was able to have an image of the process as a guideline. In personnel evaluations, it is now possible to use the invisible part as a reference for evaluation information by showing the hard work of sales activities in terms of numbers. Personnel allocation can now be used as reference information for assigning personnel to organizations that are more suitable for their activities and process methods.

Value Brought by Dataiku:

  • Even without having an advanced data science or programming knowledge, you can quickly build a process mining flow with Business Solutions.
  • Ability to create interpretable cluster models by partial dependency graphs, important feature graphs, and correlation analysis.
  • It is possible to easily work on the latest analysis methods from various business solutions.

Value Type:

  • Improve customer/employee satisfaction
  • Increase revenue
  • Reduce cost
  • Reduce risk
  • Save time
  • Increase trust
Setup Info
      Help me…