Case Study

Predictive Analytics for E-commerce Daily Planning

Higher Availability and Lower Inventory Investment

Read More

Fast Growth + Manual Processes = Availability Problems




The KPIs have been measured and archived since the go-live date for each category. In short, our client achieved significant overall improvements in availability, first-day fulfillment and unproductive stock:

The savings run into the tens of millions of dollars. In contrast, when we buy outside the system, our free cash flow gets worse. ( Forbes, Jan 2019)

Executive Board Member

Predictive Analytics and Machine Learning in Action for E-commerce Inventory Management

Rapid growth is commonly associated with e-commerce. However, fast and commercially viable growth in e-commerce requires a strong technological back-end and the mastery of the supply chain: the technical and physical to support flexibility in downstream operations, and the process automation to manage millions of SKUs in the portfolio. Furthermore, supply chain processes and technology need to recognize the unconventional structure of an e-commerce company, as it handles massive number of SKUs, sells them 24/7/365, has the flexibility to change prices many times a day, fed by many diverse data sources including a customer’s propensity to buy, real-time signals from channel competitors, conversion (e.g., web searches, click rates, conversion), fulfills millions of individual orders and offers frictionless return from the comfort of your home. That means, conventional retailer solutions simply do not fit the bill.

The Problem

Since its founding in 1998, our client kept its leader position in e-commerce, currently with 2M+ products in over 30 categories listed at the e-commerce site at any point and serving 50M+ visitors a month. 18 months ago, in preparation for rapid growth in product portfolio and the introduction of the marketplace services, our client chose Solvoyo as its solution partner to address specific supply chain issues:

Situation pre Solvoyo:

  • An outdated purchasing process that is too dependent on Excel spreadsheets and personal biases
  • Excess workload on planners due to lack of decision automation
  • Higher nonperforming stock than the industry benchmark
  • Lack of universal view of data to drive fast decision-making
  • Availability problems leading into customer complaints

Our client management understood that fast growth would put a huge pressure on operations and profitability if these issues were left unresolved. Manual and reactionary processes supported by heroic execution that fueled the company’s early growth could potentially be a barrier to reach the goal of becoming a world-class ecommerce company.

The Solution

Our client’s initial objective was to get rid of Excel as the primary planning tool and establish a common planning process and performance metrics across all its product categories. Furthermore, the vast amount of data (e.g., basket transactions, web searches, click rates, conversion, competitor prices, marketplace signals) in disparate systems in varying formats were virtually useless as separate reports that ended up in the planners’ e-mail box but never used effectively.

Within 3 months of the project start, Solvoyo established a common planning process and KPIs, gathered disparate data in electronic format under one data model and automated the purchasing process through shortterm sales forecasts for over a million SKUs and inventory optimization for the 50K SKUs kept in the DC at the time. Initial deployment helped 3 pioneering product categories with immediate measurable results. Within the next 6 months, after small refinements to accommodate the special needs of each additional category, Solvoyo platform went live for the entire company.

After the platform initiation, we extended our client’s solution to leverage a wider variety of data sources consisting of searches on our client’s web site and third party sites like Google, syndicated data from vendors like Nielsen, and social media data related to the comments made about our client’s products. Data from different sources were fed into forecast engine with machine learning capabilities to detect trends, forecasts and category sales for 6 to 12 months out. Mid-term management capability also included a portfolio analysis tool which performs product segmentation based on hierarchy & attributes and identified groups of products/SKUs that have similar performance trends. Based on the mid-term forecast provided for these product segments, our client’s procurement teams developed inventory investment strategies. As stated in Steve Banker’s Forbes article, the AI project is teaching the company how to use new data sets, and how to add more data over time.

We believe we will have better optimization with AI than without it. ( Forbes, Jan 2019)

Executive Board Member

Our client now employs the Solvoyo platform to support its world-class process and the analytics tools to support its decisions:

About Solvoyo

Solvoyo is the next generation planning and optimization platform built for the digital revolution in supply chain. Solvoyo helps companies close the gap between planning and execution. Our all-in-one platform forecasts demand, optimizes inventory, plans production, replenishes networks and concurrently optimizes transportation plans. Solvoyo is the only planning platform that allows companies to plan strategic, tactical and operational supply chain actions in one tool. The scalability, analytical capabilities and short implementation cycles of our true cloud platform help our clients achieve dramatic improvements in performance. We are headquartered in Boston, MA with our R&D center in Istanbul, Turkey. For more information, please visit