Client

Lorenz Polska

Date

Services

Data Engineering, Machine Learning, Data Science, Cloud Solutions

Technologies

Google Cloud Platform, BigQuery / BigQuery ML, Python (pandas, scikit‑learn, LightGBM, PuLP), Vertex AI, SQL, Jupyter / Vertex AI Notebooks

Challenge

Lorenz Polska operates at the heart of the B2B snack market, serving major retail chains while prioritizing modern trade channels.

The company’s strategic goals were to:

• Build world‑class analytical capabilities within commercial teams

• Automate complex assortment preparation for diverse retailers

• Deeply understand consumer purchasing behaviour

• Maximise gross margin contribution across the portfolio

• Reduce assortment management costs through automation

• Implement a robust, data‑driven promotion effectiveness measurement framework

Key considerations along Lorenz’s journey:

• Organizational focus: Analytics expertise was still developing internally, with teams balancing several parallel initiatives.

• Data landscape: Inputs varied in granularity and format, offering opportunities to harmonize datasets and enrich SKU‑level and promotional insights.

• Right-sized scope: Operating at gigabyte rather than terabyte scale, Lorenz’s needs were best met with tailored, efficient solutions rather than heavy big‑data infrastructure.

Our approach

We designed a modular, scalable, and pragmatic analytics ecosystem on Google Cloud Platform (GCP), carefully avoiding unnecessary complexity while ensuring future scalability.

Core solution components:

• Customer and product segmentation (K‑Means, DBSCAN, hierarchical clustering)

• SKU‑level demand forecasting considering seasonality and region (ARIMA, LightGBM)

• Price elasticity modelling (Bayesian GLM)

• Detection of product substitutes and complements via correlation and semantic analysis

• SKU portfolio optimisation using linear and quadratic programming

• Automated data integration (ETL in BigQuery) with interactive dashboards for business users

Implementation emphasized knowledge transfer and co‑creation. Each analytical module was paired with hands‑on workshops, enabling Lorenz’s team to independently enhance and expand the platform. Prototypes were built to be scalable and extendable, allowing future applications such as shelf analytics, integration with Nielsen data, and advanced pricing strategies.

The outcome

• Internal capability building delivers lasting business value

• Modular architecture accelerates delivery and simplifies scaling

• Pragmatic technology selection ensures cost‑effective, impactful results

• LLM‑driven semantic enrichment maximizes existing data value

• Agile, iterative delivery paves the way for new business insights  

Our partnership with Lorenz Polska illustrates how advanced analytics, delivered with a consultative and category‑savvy approach, can transform assortment planning, pricing strategy, and promotional analysis for FMCG manufacturers.

By focusing on upskilling, modularity, and pragmatic innovation, Lorenz achieved measurable operational benefits and secured the in‑house capabilities required to lead Poland’s snack category in an era defined by data‑driven excellence.

Business Impact

• Superior sales forecasts. Machine learning models consistently outperformed manual sales estimates, improving demand planning accuracy.

• Promotional data enrichment. Integration of LLMs unlocked value from previously siloed promotion datasets.

• Automated assortments change detection deployed in selected retail chain stores, enabling faster market response.

• Sustained analytical capabilities: Lorenz’s team now independently develops and adapts platform features.

"Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat."

Name Surname
Position, Company name

Lass uns reden und zusammenarbeiten

An Werktagen (Montag — Freitag, 9.00 — 17.00 Uhr MEZ) melden wir uns innerhalb von 4 Stunden bei Ihnen.

Data engineering for cloud-based data processing and storage.
Dominik Radwanski
Partner für Servicebereitstellung
Der Verantwortliche für Ihre personenbezogenen Daten ist DS Stream sp. z o.o. mit Sitz in Warschau (03-840), ul. Grochowska 306/308. Ihre persönlichen Daten werden verarbeitet, um die Frage zu beantworten und das Formular zu archivieren. Weitere Informationen zur Verarbeitung Ihrer personenbezogenen Daten finden Sie in der Datenschutzrichtlinie.
Danke! Deine Einreichung ist eingegangen!
Hoppla! Beim Absenden des Formulars ist etwas schief gelaufen.