How Consumer Packaged Goods (CPGs) can optimise pre-allocation of inventory by sensing demand to ensure high fill rates for retailers. David Hawkings, SVP at antuit.ai, introduces Intelligent Order Promising.
While some of the sharpest spikes in demand and supply seen during the pandemic have eased, it is widely accepted that variability is here to stay, given the global nature of sourcing and supply chains and their vulnerability to global events.
Learning from the pandemic so they can build more resilient demand ecosystems across sales, planning and fulfilment, CPG companies are investigating a new capability enabled by artificial intelligence (AI) called Intelligent Order Processing (IOP). As a supply chain expert, Sarah Pfaff, explains: “In the next five to 20 years, technology will allow supply chains to go from being an end-to-end efficient linear machine, where if one thing stops the whole thing stops, to an ecosystem with an incredible amount of data and AI-driven predictability.”
Put as simply as possible, Intelligent Order Processing (IOP) is the application of sophisticated algorithms to near-term consumption sensing data, guided by pre-determined strategic business objectives and customer segmentation, to produce allocation recommendations across all channels that ensure optimized business performance despite sudden gaps between demand and supply. And in order to deliver value over the long term, machine learning creates recommendations that are guided by flexible business rules for different business conditions.
IOP was developed to deal with the limitations of manual, high touch order fulfilment processes; a lack of consistency in process across the cycle as a result of decentralised decision making and multiple data sets; and the current first come, first serve response to customers that have led to high OTIF fines when stock cannot be fulfilled. IOP also addresses the limitations of ERP-based Available to Promise (ATP) and Order Promising systems that have limited capability and typically require three to four months to recalibrate.
The CPG companies neatly summarise their own needs in the words of one senior executive: “Typically, we just bend to retailers with higher OTIF penalties or compliance rates…I really wish we had a more strategic way of managing to our customers’ needs.” So this is not just about more responsive planning and fulfilment, but better financial management through fewer fines and management of cash flow impact from inventory disruption.
And it is about collaboration around shared views of common data that will give real-time visibility about who is going to suffer shipment delays on the current allocation plan, and even the ability to juggle order fulfilment in a way that reduces OTIF fines. This in contrast to moving the blame for disruptions from one department to another. For instance, Sales can prioritise strategic accounts, giving them their fair share of inventory to keep both consumer loyalty and retail shelf space through improved distribution of constrained inventory. Then moving through to fulfilment, decisions can be made based on the current market situation, strategic goals and customer priorities, as well as a near-term signals of consumer demand.
IOP leverages data and models to help CPGs see through the noise and confusion of short-term demand volatility and react quickly by automating the prioritisation of order fulfilment. It uses a combination of demand sensing, to anticipate the near-term demand from each retailer’s ship-to location, and AI/ML modelling, that incorporates customer attributes, including service level targets, profitability, volume growth, and OTIF fines levied. These attributes define the significance of each retailer partner to the CPG company, providing the ability to optimize for both allocated and unallocated inventory.
CPG firms that have deployed IOP are seeing $3m-$5m annual reduction in OTIF fines, a 4-5% improvement in fill rate for key customers, $500,000 annual reduction in planning & administrative costs and a 10x ROI. And the long-term benefits are an ability to fill based on customer segments to support growth and profitability.
Based on a study with a large European consumer electronics manufacturer, comparing conventional ATP with IOP, the latter approach used machine learning-based capability and a simple user-driven rules engine with sophisticated features to calculate allocations based on fill-rate target, fill-rate, and customer priority (risk-based modelling). An early warning capability could identify future fill-rate risks based on forecast error, past fill-rates, inventory trends, etc., and recommend product on allocation. A predictive process enabled dynamic reserves to ensure strategic customers are given priority immaterial of order sequencing, whilst still enabling advanced cost-based substitution and alternative location logic. And consensus on allocation recommendations and inventory/ supply allocations to the orders is built in using UI driven exception management.
Our recommendations for getting started are, organise a SWAT team of key stakeholders to identify the key attributes for defining the customer segmentation drivers and the outcomes to be achieved with the new process. Identify the potential integration points between legacy systems and the processes for IOP. Then benchmark ‘what we did’ versus ‘what IOP would have done’ to validate the output against the past, and to extrapolate the impact on the future.
Then move to pilot by running IOP in parallel for one distribution centre to validate the findings in a real environment, perform acceptance testing, and battle test the entire process under pressure. And lastly, go live, cut-over to production and rollout across the enterprise, making sure to measure the business impact.
The imperatives for investing in IOP are all commercial :
- Higher availability on the shelf boosts sales;
- Increased allocation in the category for the CPG’s products;
- Lower costs in the supply chain once the need for top ups is reduced;
- Competitive advantage over CPGs unable to reach higher than benchmark availability;
- Reduced OTIF fines.
