As a supplier of industrial-grade AIoT products and solutions, Guangzhou Tongwei Smart Technology Co., Ltd. (Tongwei Smart) also faces a concrete question in the production and delivery of IoT equipment: which products and materials should be prepared in advance, and which work should start only after demand is clear?
This is the question that push production and pull production have to answer together. Push production schedules production according to a forecast. Pull production triggers production and replenishment according to customer orders or actual downstream consumption. The two are not an either-or choice. What matters is how a company responds to changes in demand while also covering delivery, inventory, and cash turnover.
Push production and pull production each have conditions where they fit
Push production makes goods ahead of time according to a sales forecast or a plan. Its advantage is that batch production is easier to organize, capacity can be used, and delivery time may be shortened. If the forecast is inaccurate, product may sit in the warehouse. Pull production uses orders or actual consumption as the signal. It can reduce unnecessary finished-goods inventory, but it has requirements for supply lead time, production response speed, and planning coordination.
Take a company that produces industrial controllers as an example, and assume it has 10 models. If every model is produced once at 500 units according to the forecast, popular models may sell quickly, while slow-moving models may occupy inventory and cash for a long time. After production is arranged according to orders and inventory consumption, models in regular use can keep a suitable stock, and models with unstable demand are produced to order as far as possible.
It should be noted that pull production does not mean that nothing is started until an order arrives. More mature pull management uses downstream order and consumption signals to arrange upstream replenishment, and at the same time controls the quantity of work in process, so the production pace stays closer to actual demand.
Why small and medium companies should not pursue only one method
Looking only at output and equipment utilization makes it easy to overlook the cash tied up by inventory. Batch production may lower the production cost per unit, but if the product does not sell, the cost saved can easily fail to offset the cash occupied by unsold inventory and the later handling cost. Production efficiency has to be read together with delivery, sales, and inventory turnover.
On the other hand, pull production also has practical limits. Customer orders may swing high and low, suppliers have a minimum purchase quantity, some electronic components have a long purchasing lead time, and customers may set a very tight delivery requirement. If every material and process is started only after the order is confirmed, the company may not be able to deliver on time. Inventory pressed too low also increases the risk of material shortage and stoppage.
Arrange production by product, material, and process layer
For small and medium companies, a more practical approach is to manage different products, materials, and processes separately. When Tongwei Smart arranges production and stock preparation for IoT equipment, the work can be organized along the following lines:
- Standard products with stable sales: schedule a suitable production quantity according to historical sales and the order trend, and set a reasonable safety stock.
- Products with larger sales swings: reduce finished-goods stock preparation, and arrange production according to customer orders as far as possible.
- Common materials and special materials: common materials such as chips, PCBs, and enclosures can be stocked in moderation together with the purchasing lead time; special materials are purchased according to demand as far as possible.
- Processes with different production cycle times: arrange in advance the steps with a long purchasing lead time and a long preparation time; complete configuration, program burning, final assembly, and testing after the order is clear, as far as possible.
- Shop-floor work in process: set a reasonable upper limit for key processes, so that each process does not only chase completion of its own output and work in process keeps piling up.
For example, when producing AIoT products such as industrial gateways, common chips, PCBs, enclosures, and a suitable quantity of semi-finished goods can be prepared in advance. After the customer order is confirmed, the communication module is configured according to the specific requirement, the program is burned, and assembly and testing are completed. This can shorten part of the delivery preparation time and also reduce the chance that finished goods of different models accumulate.
Safety stock has no fixed number that applies to every company. When it is set, combine product sales, demand fluctuation, purchasing lead time, the supplier’s minimum purchase quantity, delivery requirements, and the company’s cash position, and review and adjust it regularly.
Start from simple inventory and planning management
Improving the production plan does not have to begin by putting ERP or MES in place. A company can first do three things: sort out sales volume and inventory age for each model; define the minimum stock, the maximum stock, and the replenishment trigger; and let the production plan refer at the same time to orders, inventory consumption, material lead time, and actual capacity, rather than looking only at shop-floor utilization.
A clear Excel sheet or a shop-floor board can serve as the starting point. First make the inventory data and the production process run accurately and steadily, then introduce system tools step by step according to management needs. The effect is then easier to judge.
The goal of production management is not to keep the shop floor busy all the time. It is to have the company produce the right product at the right time and complete delivery as promised. Push production deals with preparation in advance. Pull production helps reduce production that has left real demand. For small and medium manufacturers, what really matters is to judge, from product characteristics, supply conditions, and the order rhythm, what should be done in advance and what should wait until demand is clear.
Tongwei Smart will continue to share practical observations from the development, production, and delivery of industrial IoT products. Specific questions about production planning, material stock preparation, and product delivery are welcome.