Contact

Trending

Favorisez l'engagement clients

Développez un modèle de machine learning permettant de prédire les futurs churners, afin de mettre en place un plan d’actions marketing anti-churn ciblant vos clients à risque.

Approche

La mise en place du modèle de scoring de la construction jusqu’à l’industrialisation s’est faite en six étapes :

  1. Data Engineering sur les bases de données CRM
    1. Analyse descriptive des variables issues des différentes bases : qualité, exhaustivité, cohérence, répartition, corrélations
    2. Nettoyage et retraitement. Consolidation des données et préparation pour l’analyse
  2. Analyse exploratoire des données clients
    1. Identification de premiers profils de churners
    2. Identification de premiers schémas de comportements (patterns)
    3. Détection des variables discriminantes
  3. Entrainement du modèle de scoring 
    1. Entraînement et calibration du modèle supervisé
    2. Tests des résultats et analyse de performance
  4. Validation du modèle de scoring
    1. Présentation du modèle aux équipes marketing et service client
    2. Vérification de l’adéquation avec les besoins métiers
    3. Ajustements et itérations
  5. Industrialisation du modèle de score dans l’infrastructure informatique de notre client
    1. Livraison et déploiement d’un script Python industrialisé pour la prédiction hebdomadaire sur les prochaines échéances
  6. Construction d’un programme marketing anti-churn avec nos consultants Marketing et customer expérience et les équipes internes intégrant le score de churn

Our approach begins by defining CLV within the context of our client's group, tailored to meet their specific needs and objectives. We meticulously analyze the historical CLV of past customers, drawing insights from their past interactions and transactions to inform future strategies. Furthermore, we explore and benchmark various methods to estimate future CLVs, ensuring the accuracy and relevance of our projections. By estimating the future CLV of active clients, we provide invaluable insights into their potential long-term value, enabling tailored offerings and services. Through promoting the added value of CLV within the client's organization, we aim to cultivate a culture centered around customer-centricity and data-driven decision-making, ultimately driving success and innovation across all business areas.

Our comprehensive approach began with estimating the size of the market (in 2 asian countries), focusing on local residents who consume personal luxury goods. Through meticulous research and analysis, we identified key opportunities to expand our client's local client base. This included profiling key client segments, estimating population sizes, and assessing the dollar value of potential markets, all while staying abreast of current market trends. Additionally, we delved deep into understanding the motivations and barriers of our target client profiles, enabling us to develop actionable strategies to engage with priority clientele effectively. By providing these insights and actionable ideas, we empower our client to navigate the dynamic landscape of the luxury market and unlock new avenues for growth and success.

We drove a project centered around enhancing customer experience through personalized journeys. We focused on the development of robust customer knowledge, empowering us to tailor every interaction based on unique customer profiles and product affinities. Through the implementation of a predictive program, coupled with rigorous performance measurement, we ensured that every engagement was not only personalized but also optimized for effectiveness. Furthermore, we provided comprehensive support and training to all stakeholders, including sales, market, and marketing teams, on leveraging the insights and tools generated by the program. By prioritizing future customer-centric initiatives based on value, feasibility, and costs, we set the stage for continued innovation and success, reaffirming our commitment to delivering unparalleled experiences that resonate with all customers.