Improving and adapting the upstream supply chain of an i-grocer during fast growth

Abstract

Since online shopping was invented in 1979 there has been a major shift in the way people shop. More than 40% of the clothing and footwear and more than 35% of the home electronics are bought online in the Netherlands. Grocery shopping in the e-commerce market. This, and the fact that over 92% of the Dutch population has access to internet implies that there is a huge market gat for e-grocers. Predictions are made that the market share of e-grocers will grow from the current 2% to at least 5% in 2025.
At the moment the e-grocery market is dominated by traditional supermarket who also delivery online shopping as a service to the client, the so called click-and-mortar concept. Currently 55% percent of the market is owned by AH followed but not yet threatened by other supermarkets like Jumbo and Spar. In August 2015 however an online-only supermarket, a so called pure player, opened named Picnic. Because of the fact that they are a pure player gives them confidence that they can compete in this market. Since the opening of Picnic has grown to just above 9000 orders per week in January 2017 when the maximum capacity of the first fulfilment centre was reached. The aim of Picnic is to continue a with a growth of around 5% per week which comes with a lot of challenges, starting with the fact that a new fulfilment centre had to be opened in February 2017. The ambition of Picnic is to deliver high service to the customer with lowest price guarantee which means the processes at the fulfilment centres should run as efficient and as smooth as possible.
In literature there is a lot to be found about supply chains in general and more specific about e-grocer supply chains. It has become clear that there are several pit falls for e-grocers to recon with. First of all there last-mile delivery is relatively very expensive. Furthermore, over investment in automation, weak negotiation power, customer acquisition and low ordering frequency are reasons why online grocers have failed in the past. Unfortunately, there is little known about the effect of fast growth on the supply chain, let alone the effect on the supply chain of an e-grocer.
In order to give a well-founded advice in how to improve the upstream supply chain of Picnic while adapting to the fast predicted growth, the following research question is defined:
How can the additional costs due to disturbances of an upstream supply chain of an e-grocer be
reduced while accommodating expected rapid future growth?
To answer this question a research methodology is developed consisting out of five steps: literature research, analysis of the current process, design (synthesis of the two previous steps) , assessment and conclusions and recommendations.
From the literature research and the analysis it can be concluded that the process in the FC consist out of seven steps: the inbound, receiving, replenishing, picking, storing, consolidating and shipping. In a happy flow all of these processes work without disturbances ergo, all the items are delivered at the FC and are on the shelves as expected. This would mean all the order lines would be shipped completely. When something goes wrong extra costs are involved and it has been researched what can be done to reduce these additional costs. It has also been found that the rush order process is highly inefficient and store picking is relatively expensive.
It has been found that there are eight root causes which lead to unexpected shortages in the FC: un-orderable products, shortage at supplier, delivery failure, receiving’s mistake, unexpected clearance, non-fifo picking, quality buffer and stock mismatch. These unexpected shortages have either no consequence (other than a stock adjustment) or a rush order, store pick, substitute or a cancelled order line, of which respectively no consequence is the most desirable and cancelled order line is the least desirable effect. After an extensive data analysis it was found that delivery failures had the most negative impact on the processes in the FC, followed by stock mismatches and non fifo picking. Also, the rush order process appeared to be highly inefficient since it is designed for picking much larger amount of items and the store pick procedure is relatively expensive.
The results from the analysis phase have led the creation of five design alternatives: reduction of delivery failures (RDF), improvement of insight in the FC (IWI), direct delivery (DD), different store pick procedure (DSPP) and an improvement of the rush order process (IRP). Here the reduction of delivery failures is chosen because of the clear advantages for both the supplier and the e-grocer. Improvement of the insight in the FC is wanted since the consequences of an stock mismatch are relatively often a substitute or a cancelled order line which are highly unwanted. Since the performance of the supplier is so essential for the e-grocer, the alternative of (partly) direct delivery instead of using the DC of Boni as a buffer, is also taken into account. Improving the rush order processes and changing the store pick procedure are chosen as alternatives since both processes were labelled as inefficient so there is room for improvement.
These alternatives are subsequently assessed using a Monte Carlo simulation with as input the root cause distribution of the different alternatives. This input resulted in a distribution of effects which were subsequently the input for an extra costs calculator. These extra costs per day (and per order) were used as a key performance indicator to compare the alternatives. Furthermore as an input the predicted growth scenario 5% per week was chosen.
First of all it can be seen in both figures that all alternatives initially lead to more extra costs. This is due to investments needed to enable the different alternatives. In Figure 1 the influence of the performance of the supplier can be clearly seen. A reduction of delivery failures leads ultimately to relatively lower costs in contradiction to the worse performance of direct delivery. Changing the store pick procedure results in extra substitutes and cancelled order lines which that leads to extra costs.
The influence of improvement of insight in the FC is better shown in Figure 2 where at low order amounts the extra costs to enable the extra insight outweigh the advantages. At higher order amounts above 3500, extra costs are saved compared to the current situation. The same goes for improvement of the rush order process.
It can be concluded from the research that the reliability of the deliveries are essential for the performance of the FC. Therefore it is recommended that both the supplier (Boni) as Picnic look for solutions in reducing the amount of delivery failures. Furthermore, when direct delivery is necessary because of the limited capacity of Boni, clear performance deals should be made with the suppliers which ought to be checked, automatically if possible. Here the delivery performance of Boni can be used as benchmark.