47 CHAPTER V CONCLUSION AND RECOMMENDATION 5.1 Conclusion After analyzing the data, here are the conclusions to answer the questions of this research. PT Idola Cahaya Semesta had already developed a software program that help the company determine when to order (called Minimum stock target) and how much to order (called Suggested order Quantity) for each item of each category, though the company still face under-stock and overstock problems. The result of the software however is not always fully implemented because the company sometimes has to adapt to meet sudden rise or fall in demand. After comparing the total cost for the year 2015 between the existing condition, TSM fully implemented, EOQ, and Hybrid policies, the proposed EOQ policy on how much to order and the level of inventory point to be replenished is the best policy to use because of the lowest total inventory cost and the insensitivity of cost towards the change in demand. The optimum parameter resulted from the projection of the EOQ method, is the long term reliability from the high profit probability ratio. Compared to the existing condition, the current TSM policy fully implemented, and the Hybrid policy, the total cost was reduced by IDR 3,544,761 or a 55.7% improvement, IDR 3,143,214 or a 49.4% improvement and IDR 1,313,180 or a 20.6% improvement. The proposed EOQ policy is less sensitive against the rise of demand, making it more reliable in the long term than using the existing TSM policy. Using seasonal index demand forecasting, the forecast is reliable for most items. The company’s existing business process, especially in the retail store has a potential problem, which is human error and data input mistakes, so the researcher suggested a manual inventory checking on paper, in addition to their already 5.2 Recommendation The company should use the EOQ (Economic Order Quantity) policy as it produced the lowest inventory cost and deviation of quantity. The implementation plan is simple, because the company has its own third party developer for the inventory management software, the company should contact the software engineer to tweak the software’s formula, without buying new software, and it is very quickly can be implemented. 48 This research aims to give an inventory solution for a retail company operating with non- food products, though this research is very limited to one particular store in Bintaro and one particular category of items which is stationary items. This outlet in Bintaro has a number of schools, offices and religious places that is potential to Some recommendations for future researches are to research on different categories of items, retail stores of a different location because environmental factors are highly influential to the demand levels. For example an outlet with more schools and offices nearby has a chance of higher stationary demands, while an outlet that has more households nearby has a chance of higher house and baby supplies demands. For small and medium retail enterprises that doesn’t have the software program like the company does, this method of inventory management and forecasting is also applicable as long as there are historical data of each items, so keeping a historical data of demands, costs and other data is essential to run a retail store. To summarize, a retail company must have a good historical data, a good inventory management policy of how much to order a certain item and when to order it. Retail companies has to also be aware of the environmental because it is impactful if nearby are the company’s target markets, especially in demand and branding. With both of that variables, a prediction of how many sales there is going to be (demand) can be anticipated with the forecasting methods. Marketing is also huge factor in determining the number of sales. The best retail stores keeps their inventory as minimum as possible while keeping up with customers demand as a balance in their work operations. For future research that is analyzing this similar topic, it would be better to analyze more than one outlet because it would be more representable if there are more data samples.