loader
Untitled design - 2024-07-02T170341.575 pattents-img pattents-img

Retail & Consumer Packaged Goods

pattern
white-dice-img

We understand Retail & CPG

Data and AI are propelling advancements in the Retail and Consumer Packaged Goods (CPG) sectors, yet retailers continue to face substantial challenges. Outdated systems fail to meet real-time demands and offer subpar reporting, resulting in diminished accuracy and mistrust in critical data.

Limited compatibility with diverse data types and costly, complex data-sharing agreements further complicate the situation.

Retailers resort to implementing fragmented data systems to address these issues individually. Our data platforms enable our clients to unlock the potential of Data and AI effectively.

 

Take a look at some of our customer stories:

Group 1
bottom-right-pattern-1
bottom-right-pattern-2
Untitled design - 2024-07-02T164831.009
pattents-img pattents-img pattents-img pattents-img

The Hershey Company

Hershey approached Advancing Analytics to help them deliver a transformational project around how they process highly sensitive data at an organisational level. With any significant technical shift, Hershey needed support and guidance. Advancing Analytics were asked to help Hershey's with the design, implementation and education of their team, to tackle new challenges and make the most of a modern data platform.

Read the whole Hershey case study now. 

Untitled design - 2024-07-02T165352.979
pattents-img pattents-img pattents-img pattents-img

Sunbelt Rentals (Ashtead)

To drive commercial growth, Sunbelt Rentals needed to understand, predict, and automate their depot stock levels. This ensures the right assets are available at the right time, avoiding both depletion and overstocking. Their manual demand and utilisation forecasting was limited to a few days, posing a logistical challenge.

Understanding demand impacts every part of Sunbelt’s business, from supply chain and purchasing to the rentals, delivery, and repairs teams. Recognising this, Advancing Analytics collaborated with Sunbelt to design and build a data platform in Azure. By leveraging Advancing Analytics’ Hydr8, they rapidly onboarded data from various internal and external sources. Azure Databricks facilitated the development of advanced forecasting algorithms considering factors like seasonality, locality, weather, trends, and popularity, ensuring high accuracy even during the pandemic. The deployment was accelerated with one of Advancing Analytics' AI accelerators.

Together, they established a cloud-native unified analytics platform that efficiently forecasts equipment demand using machine learning. Advancing Analytics developed a scalable ML model that predicts stock levels for specific locations with 90% accuracy. This enabled Sunbelt to understand which marketing initiatives to pursue, manage inventory, prevent stock losses, and improve prediction performance by 50%. This ML-based model for demand forecasting enables Sunbelt to stay ahead of the curve.

Untitled design - 2024-07-02T165748.042
pattents-img pattents-img pattents-img pattents-img

Lootcakes

90% of mobile game revenue comes from free-to-play games. Lootcakes, a platform that incentivises in-app purchases and drives upsell in existing games, faced a challenge: they didn't know which game to recommend next. To solve this, Lootcakes enlisted Advancing Analytics to design and create a Data Lakehouse platform for player-game recommendations, aiming to better understand their customers and create a personalised gaming environment.

Advancing Analytics collaborated with Lootcakes to develop a data platform in Azure. Once the platform was established, they built an ensemble of models to predict which games a player was most likely to purchase. Making accurate recommendations was complex due to the popularity bias—playing a game like Candy Crush doesn't necessarily mean the player prefers puzzle games. Addressing this bias was crucial for precise recommendations.

Advancing Analytics created an ensemble model deployed at scale in Elasticsearch to tackle Lootcakes' challenges. To predict a user’s propensity to spend, they analysed the types of games played, spending habits, and engagement patterns, ensuring incentives effectively drove actions. Each prediction needed to be resolved in under 400 milliseconds to avoid disrupting the user experience.

The deployed platform enabled Lootcakes to enhance customer engagement. Lootcakes v2 is now fully AI-driven, with the model integrated into the new user journey, leading to better customer interactions and increased in-app purchases.

_Ñëîé_1