TOCICO

Dr. Rakesh Sinha examines a specific challenge in TOC Distribution: if consumer demand is the ideal constraint and acts as the drum, can inventory buffers respond to demand changes quickly enough to exploit that constraint? The context is consumer packaged goods and omnichannel retail, where demand is volatile, seasonal and difficult to forecast, while distribution must support physical stores, direct-to-consumer, e-commerce and quick-commerce channels. Sinha argues that the buffer-size decision is the leverage point because availability, shortages and excess inventory are driven at the granular SKU-node level. The current reality described in the presentation uses buffers that may stay stable for extended periods and are resized from penetration history. Adjustments for promotions, seasonality, festivals, pricing and distribution expansion are often manual. The resulting undesirable effects include delayed response to demand shifts, lumpy changes that stress supply, blindness to forthcoming demand spikes, stockouts during surges and obsolescence for short-shelf-life items. This raises a set of questions: can maximum demand during replenishment lead time be assessed from past consumption alone, are demand changes really unpredictable noise, and does buffer resizing need to remain reactive? The proposed injection is Demand Sensing at the level where inventory buffers are required. AI/ML self-learning models predict near-term demand at a granular level, incorporate known demand drivers, and calculate a buffer using replenishment-lead-time demand plus protection for prediction error and supply reliability. Predictions and buffer sizes can refresh daily or hourly as new demand data arrives. This changes buffer management from reactive to proactive, with smaller and more frequent adjustments intended to respond quickly without creating supply instability. The deck frames this as resolving the apparent Dynamic Buffer Management conflict between frequent buffer changes for responsiveness and stable buffers for reliable supply. Sinha reports that Dynamic Buffer Sizing has been implemented at more than 50 CPG and retail companies, with typical outcomes of roughly 80% fewer stockouts and 35% lower inventory. Five examples illustrate the range: a quick-commerce retailer improved SKU-store availability from about 79% to 94%, reduced inventory days by about 30% and cut expired fresh stock by about 50%; a bakery reduced returns from 6% to 3% and stockouts from 6% to 0.3%; a large CPG company reduced unfulfilled retail orders from 4% to 0.5%; a medium-sized CPG company improved availability from 77% to 91%; and a fish-net manufacturer reduced dealer stockouts from 5% to 1%. The core learning is that frequent, demand-driven target revisions need not create chaos when they are granular, automated and supported by fast feedback.