OTIS EMEA - Leveraging Data Management to Better Identify Customer Needs and Improve Service Quality
Project developed with external providers: Eulidia and Ipso Senso
Eulidia: Alexis Declercq Clément Grenier Lead by Fabien Jongelen and Cedric Jacques
Ipso Senso: Clément Brun Thomas Fossurier Nina Lallier Lead by Regis Fagianelli
Title: EMEA, Director Business and Analytics projects
Organization: OTIS EMEA
Otis is the world’s leading elevator and escalator manufacturing, installation, and service company. We move 2 billion people a day and maintain approximately 2.2 million customer units worldwide, the industry’s largest Service portfolio. Headquartered in Connecticut, USA, Otis is 69,000 people strong, including 41,000 field professionals, all committed to meeting the diverse needs of our customers and passengers in more than 200 countries and territories worldwide. For more information, visit www.otis.com and follow us on LinkedIn, Instagram, Facebook and Twitter @OtisElevatorCo.
Best MLOps Use Case
Best Partner Acceleration Use Case
Otis is the world’s leading elevator and escalator manufacturing, installation, and service company.
Through a service contract signed with our customers, our service activity consists of ensuring the smooth operation of our solutions in the safest possible conditions for our customers and passengers.
In this increasingly competitive and digitalized environment, Otis EMEA (Europe, Middle East, and Africa) teams had to define and deploy a robust strategy to analyze and anticipate the behaviors and expectations of their service customers in order to better serve them.
With this approach, we decided to improve our data management to better identify customers’ needs and develop appropriate actions to reinforce our relationships with them.
We have built a project team, sponsored by Otis EMEA Sales Executives, and managed by an EMEA Project Director, to develop a data science solution for our Operations teams (Sales Engineers and Sales Managers).
The objective was to provide clear information on contracts (especially contracts with potential cancellation risk) and guide the actions of our teams on the field.
The requirements were:
A stand-alone solution to be deployed quickly and with a limited budget across EMEA countries.
To be used by a population of more than 1,000 Sales engineers and managers.
Based on data science models customized by country.
Action-oriented with enhanced UI/UX.
The project manager decided to outsource the development of this solution to a service provider specialized in data science and with the skills to develop web applications.
Otis finally took over the Dataiku license used by the external data science team to maximize the work of data cleaning, formatting, contract scoring, etc.… and ensure that each country could benefit from a robust analytical model customized to the local business environment.
Collecting and harmonizing data between countries using different CRM systems was a real challenge. Moreover, data accuracy had to be continuously improved and managed to support the model efficiency. For some countries, this meant reworking their data and establishing a data culture.
Although the end-user application is a custom-developed tool, the data preparation, the contracts clustering, and the churn risk assessment were successfully performed through Dataiku. Dataiku was key to the success of the solution:
Dataiku has been used to prepare, harmonize and clean the data despite the different levels of data maturity in the countries involved in the current deployment.
The Auto ML tools have been used in the project to:
Asses the churn risks (Random forests including a lot of features).
Consolidate the contracts according to specific criteria (prices/value/size).
The data is refreshed each and every month for 11 countries (and counting) thanks to the scenarization framework offered by Dataiku, allowing us to upload each model successfully at the same time, taking into account specific countries' data.
The analytics tools are used to develop and improve the solution.
Since its launch almost three years ago, the Dataiku License has been deployed in 11 countries. This project has changed the way Sales Engineers and Branch Managers focus on their portfolio and customers.
We have significantly improved our service data management culture since integrating them into data science models. The analysis of our data is more advanced. We now have information that we didn't have before, enabling us to improve the quality of our service activity and consequently increase our customer satisfaction.
Operations teams can also prioritize their actions more easily and be more proactive by anticipating and scheduling service interventions, thus better managing potential technical challenges.
Each Sales Engineer has access to his or her lists of actions that they access through a dedicated app — which is regularly updated — with detailed information on service and sales data. Thanks to data management, they can easily track the performance expected by our customers and their managers.
Business Area: Marketing/Sales/Customer Relationship Management
Use Case Stage: In Production
Contract classification through more accurate data management enables Otis EMEA local entities to improve their service quality. The teams can now proactively manage their contracts portfolios by anticipating and customizing the service activities according to the customer and, therefore, better-managing expectations. Thanks to the availability of new data and improved data management, sales teams are now empowered to plan the right actions at the right time.
For Otis EMEA entities that have been using the solution for a year or more, this has resulted in a significant decrease in contract cancellations as well as better efficiency in extending current contracts for the following years.
We also noted that Sales Engineers were increasing their number of interactions and visits with customers.
A side (and very positive) effect of this project has been the improvement in data quality and accuracy, which helps to improve the model efficiency.
Value Brought by Dataiku:
The specific value brought by Dataiku was on various levels:
The speed and agility provided by the visual recipes allowed us to:
Work on different approaches at the beginning of the project. The POC carried out on the data from the first country enabled us to define a generic data treatment to prepare it for usage in the final application.
Adapt to the different countries where the solution has been deployed, regardless of their maturity, by easily taking into account local specificities.
With the AutoML features, we were able to compare various churn risk models during the POC phase and easily group similar contracts based on these criteria.
Analytics features add flexibility and capacity for continuous improvement when deploying in new countries.
The scenarization tools and interdependencies of the projects allow us to unify data from different places to get all of the data available for the tool after getting some valuable information.
Finally, Python and SQL scripting allows us to customize our treatments when needed.