09/15/2026
Introducing Predictive Analytics Cycle Counting for Inventory Management
GEODIS believes we can improve on standard inventory counting with an upgrade to a new model for counting, checking, reconciling, and correcting inventory: Predictive Analytics Cycle Counting.
Maintaining accurate stock levels is critical to customer satisfaction, and good inventory management ensures that products are available for ordering and can be sent to customers in a timely manner. GEODIS believes that we can improve on standard cycle counting and proposes an evolution to a new model for counting, checking, reconciling, and correcting inventory: Predictive Analytics Cycle Counting (PACC). This is a novel and unique approach that drives more accurate inventory management compared to standard cycle counting methods.
Key takeaways
- Predictive Analytics Cycle Counting is a novel and unique approach that drives more accurate inventory management compared to standard cycle counting methods.
- Accurate inventories prevent overstocking and out-of-stock items while meeting customer needs, improving demand forecasting, preventing loss, and enhancing satisfaction.
- PACC uses a mathematical model to predict cycle count locations that have a high probability of mismatch in a warehouse, based on historical cycle counting data and other contributing factors.
- The benefits of PACC include enhanced root cause analysis and remediation to prevent inventory errors, redeployment of labor to revenue-generating activities, and a more engaged workforce that benefits the client.
The PACC pilot demonstrated:
- Indirect labor hours were reduced by more than 30% and redirected to client revenue-generating activities.
- An increase in the units per hour processed through GEODIS facilities.
- An increase in productivity performance from 95% to 110%.
- A reduction in per-unit costs for picking and packing.
- An increase in the permanent GEODIS workforce servicing the pilot client, compared to temporary labor.
Using PACC with a live client demonstrated:
- A significant boost to top-line productivity through more efficient picking, packing, and replenishment activities, boosting productivity by 17% over four months.
- 28% reduction in non-value-added, indirect labor for our client.
- A matching increase to direct, revenue-generating activities for our client.
- Cost savings of approximately $57,000 a year, or one FTE of effort.
- PACC is an optional approach that can replace or be used alongside existing cycle counting methods.
- PACC is available to all new and existing clients for GEODIS Americas.
Clients rely on in-house and third-party logistics (3PL) providers to receive, pick, pack, and distribute products to end customers on the client's behalf. Maintaining accurate stock levels is critical to customer satisfaction, and good inventory management ensures that products are available for ordering and can be sent to customers in a timely manner.
Logistics providers ensure that inventory levels are accurate by physically counting the individual products stored in the warehouse. They then match this physical count against the expected number recorded in a warehouse management system (WMS).
3PLs and others use several approaches to count and reconcile inventory levels. Industry best practice recommends a "cycle counting" approach. Inventory control (IC) staff are responsible for counting subsets of a client's inventory on a regular basis and reconciling that physical product count against levels in a WMS. This helps the warehouse provider identify discrepancies, investigate and resolve issues, and keep shrinkage within acceptable service levels.
Although standard cycle counting is a reliable way to check and maintain inventory accuracy, it does have several drawbacks. These include substantial workforce effort, delays in identifying and resolving errors, and fast-moving inventory not being checked often enough.
GEODIS believes that we can improve on standard cycle counting and proposes an evolution to a new model for counting, checking, reconciling, and correcting inventory: Predictive Analytics Cycle Counting (PACC).
The GEODIS PACC approach analyzes a client's historic inventory data. A neural network model then uses algorithms to predict specific products and warehouse locations that should be counted next, based on their likelihood of being incorrect. This allows us to concentrate our IC teams on potentially problematic inventory locations and SKUs. Pilot studies have shown that this reduces IC effort, allows for earlier identification of root causes, maintains a more accurate inventory, and delivers significant benefits for GEODIS and our clients.
Predictive Analytics Cycle Counting is now available for new and existing GEODIS clients. In this article series, we explore:
- The issues logistics providers face when counting inventory.
- How a predictive analytics approach can solve those problems.
- The benefits of predictive analytics cycle counting.
- How GEODIS developed a predictive analytics mathematical model.
- Results of the GEODIS PACC pilot.
- How you can take advantage of GEODIS predictive analytics cycle counts.
Take advantage of PACC to optimize your stock levels. Get in touch with GEODIS.
An overview of predictive analytics cycle counting
The aims of predictive analytics cycle counting
PACC is an innovative method that GEODIS is using to improve inventory management cycle counting in our facilities and operations. We have successfully completed the GEODIS PACC project to define, implement, test, and refine a more effective and efficient cycle counting model, based on data analysis and problem prediction. The aims of this project included:
- Building a mathematical, predictive model that collects data and uses algorithms to identify potentially problematic inventory areas, SKUs, and cycle count locations.
- Training the model to learn from historic data and external information.
- Predicting the subsections of a client's inventory that are most likely to have errors and discrepancies between the inventory level captured in the WMS and the actual inventory count.
- Producing, prioritizing, and ranking a representative sample of products and locations for inventory counting.
- Allowing for the continuous counting of locations flagged by the PACC model.
- Working alongside existing, best practice cycle count methods.
- Supporting early root cause analysis and remediation to reduce and prevent inaccuracies.
- Avoiding static assumptions about any part of inventory counting and adopting an objective, data-driven approach.
- Using a self-learning, self-correcting predictive analytics mathematical model.
- Decreasing rework, missed orders, or skipped picks due to incorrect inventory levels.
- Improving client service levels and satisfaction.
- Redeploying inventory counting workforce effort to more value-added and revenue-generating activities for the client.
The project has been piloted and has delivered successful, repeatable results. We believe that this new approach to inventory management is ready for wider release to GEODIS clients, and that the improvements it delivers have significant implications for the wider logistics industry.
Why GEODIS is introducing a new type of cycle counting
Inventory counting is critical because it allows 3PL providers and their clients to know how much inventory is in the warehouse at any moment in time. This is fundamental to:
- Preventing overstocking or out-of-stock items.
- Meeting customer needs for product availability and delivery speed.
- Optimizing demand forecasting and inventory replenishment lead times.
- Identifying and preventing loss.
- Managing stock that is past its shelf life.
- Providing clarity throughout the supply chain.
It's important to reduce inventory inaccuracies, also known as discrepancies or variances, as much as possible. These inventory record inaccuracies happen when the physical count of a product in a warehouse location does not match with the expected inventory level recorded in the WMS.
Causes of inventory inaccuracies
Inaccuracies are a persistent and recurring issue for logistics businesses and their clients. Errors with stock levels can incur significant direct and indirect costs, driven by product shrinkage, capital tied up in excess inventory, unnecessary workforce effort, customer dissatisfaction, service level breaches, and missed opportunities.
Research shows that the main causes for inventory inaccuracies can be broadly classified into four groups:
- Transaction issues including shipment errors, delivery errors, scanning errors, and incorrect identification of items.
- Shrinkage issues that cause losses of products ready for sale.
- Inaccessible inventory issues including products that are not in the correct location and are not available for customers.
- Supply issues including errors caused by product quality or yield efficiency.
Warehousing companies minimize inventory record inaccuracies by regularly counting physical products in the warehouse and comparing these counts against electronic records of stock levels. These inventory checks identify and rectify discrepancies between data records and physical product counts.
One of the biggest challenges for warehouse operators is that inventory checks are extremely time-consuming activities, especially when all items need to be counted. This results in significant workforce effort from cycle count teams. To reduce this effort, logistics businesses can adopt inventory cycle counting methods that check a representative sample of items.
How predictive analytics solves inventory inaccuracies
Predictive analytics cycle counting is a GEODIS initiative to identify potentially inaccurate inventory levels using historical inventory checking data. This allows for more efficient cycle counting aimed at identifying and resolving inventory discrepancies in a quick and effective way.
PACC discovers correlations between various item and storage properties (for example, storage location, frequency of movement, previous counting discrepancies, value) and inaccuracies. These correlations are then used to predict items and locations where stock levels are most likely to be inaccurate, so inventory counting can focus on checking and reconciling those items.
The PACC model is based on and trained by actual data from warehousing operations. It classifies cycle count locations that have a high probability of mismatch in a warehouse. These predictions are based on client inventory data, previous cycle counts, and other contributing factors including location, pick front activity, SKU order volumes, item retail value, and similar metrics.
GEODIS has piloted the PACC model with excellent results:
- A significant reduction in IC workforce effort allowing for labor to be redeployed to revenue-generating activities for clients.
- A strong improvement in unit-per-hour throughput for processing client products through our facilities and distribution to end customers.
- A boost in SKU-handling productivity, for faster turnaround and distribution of orders.
- An increase in workforce engagement and knowledge retention.
It's helpful to understand the context for cycle counting and its use within 3PL operations.
Take advantage of our inventory management. Get in touch with GEODIS.
Defining existing cycle counting methods
Physical inventory counting approaches
Prior to the introduction of cycle counting, warehousing and 3PL businesses used a "physical inventory" counting approach. In a physical inventory approach, an organization would count the entirety of a client's inventory on at least an annual basis. Unfortunately, physical inventory counts suffer from several disadvantages:
- Substantial labor requirements, often requiring extra staff.
- Costly to perform due to increased workforce size and overtime.
- Requires facility downtime, resulting in delays processing incoming stock and customer orders.
- Lacks root cause analysis, making remediation and prevention difficult.
As a result of these drawbacks, GEODIS, together with many other 3PLs, uses a cycle count methodology.
Standard cycle counting approaches
Prior to PACC, current GEODIS cycle counting uses a combination of best practice approaches for checking and reconciling inventory levels. These approaches are defined and certified by the Association for Supply Chain Management (ASCM) through the APICS framework (originally the American Production and Inventory Control Society).
"According to the APICS Dictionary, cycle counting is: 'An inventory accuracy audit technique where inventory is counted on a cyclic schedule rather than once a year. A cycle inventory count is usually taken on a regular, defined basis (often more frequently for high-value or fast-moving items and less frequently for low-value or slow-moving items). Most effective cycle counting systems require the counting of a certain number of items every workday with each item counted at a prescribed frequency. The key purpose of cycle counting is to identify items in error, thus triggering research, identification, and elimination of the cause of the errors.'" — Cycle Counting by the Probabilities, Association of Supply Chain Management
GEODIS carries out cycle counting to reduce inventory shrinkage and to monitor, track, and meet our service level agreements. Broadly, GEODIS uses two main cycle-counting techniques:
- Geographic location cycle count.
- ABC prioritization cycle count.
Geographic location cycle count method
The geographic location method splits the warehouse up into a number of discrete physical locations. Our IC workforce then counts all items stored in those locations and compares them against inventory levels in our WMS. Each location is counted one or more times a year.
All locations within GEODIS warehouses are counted, this includes regular inventory locations, pick faces, empty locations, staging areas, and out-of-service locations.
ABC prioritization cycle count method
The ABC count method categorizes individual items and SKUs into priorities, based on certain criteria such as velocity or retail value. It then counts each priority level at a different frequency. For example, in a velocity-based ABC count approach:
- A client sells 10,000 different SKUs.
- The 2,000 fastest moving make up 70% of the volume and are placed in priority category "A" and are counted on a monthly basis.
- The 2,001-4,000 fastest moving make up the next 15% of the volume and are placed in priority category "B" and are counted on a quarterly basis.
- The 4,001-6,000 fastest moving make up the next 10% of the volume and are placed in priority category "C" and are counted every six months.
- The remaining SKUs (6,001-10,000) make up the final 5% of the volume and are placed in priority category "D" and are counted once a year.
Clients can choose to use either or both of these inventory count methods. The new PACC approach can replace either of these methods or be used alongside them, depending on client needs.
Advantages of standard inventory cycle counting
Cycle counting, whether standard or PACC, provides several advantages over physical inventory count methods:
- Count frequency: Cycle counting means that inventory items can be counted and reconciled more often.
- Live count: These cycle count methods can be carried out during normal operations and do not require quiet periods or operational downtime.
- Blind count: The user who is performing a count does not know how many items should be in an area according to the WMS, instead they purely count the actual items present.
- Automation: Counts are performed through a scan with an RF gun, allowing for quick cycle counting and removing misidentification errors.
- Service level agreements: Discrepancies are calculated against total inventory to calculate shrinkage rates and to ensure GEODIS meets its client Service Level Agreement (SLA) requirements.
- Root cause analysis: Any inventory count discrepancies are recorded and flagged for investigation as part of root cause analysis (RCA).
- Reason codes: Following RCA investigation, count discrepancies have a "reason code" attached that identifies the problem that caused the variance, which allows for future analysis and preventative measures.
- Reduced variance: Cycle counting reduces the number and severity of inventory discrepancies.
- Guaranteed annual count: GEODIS counts all locations and every item in a client's inventory at least once a year.
Cycle counting delivers several benefits for clients, including:
- Higher product availability, order fulfillment rates, and sales.
- Better customer service levels and increased customer satisfaction.
- More accurate inventory assessments and fewer errors.
- Fewer inventory write-offs and obsolete inventory.
- More efficient warehouse operations.
Decreased audit fees and employee overtime costs.
Issues with standard inventory cycle counting approaches
Although standard cycle counting provides major benefits to clients, it does suffer from some drawbacks. The typical, industry-standard approach is to cycle count each location and product in a warehouse at least once a year. Because warehouses operate on limited resources (the number of employee hours available to complete operational activities), this presents several challenges:
- Wasted effort counting locations with extremely slow-moving products, as the inventory level does not change between counts.
- Some locations remain empty for extended periods, yet these locations still require cycle counting.
- Delays in identifying and resolving issues with fast-moving items, as it may take several months to uncover an error in stock levels.
- Undiscovered errors in stock levels can be compounded and compromised further the longer they remain, making root cause investigation, analysis, and remediation much more difficult.
Fortunately, we can resolve these issues with a PACC approach, covered in Part 2 of this series.
Paul Maplesden
Former Lead Content Strategist
This is Part 1 of a three-part series. Part 2 covers how the PACC model works and the benefits it delivers.