مدل عرضه و تقاضا در یک بازار B2B-بالادست جدید با استفاده از فرایند به روز رسانی دانش

Business-to-Business (B2B) services companies invest heavily in acquiring very expensive assets that they hire out to serve their clients (e.g. UPS buys huge warehouses and hires them out to companies), and hence they engage in careful long-term planning and forecasting, especially when it concerns a new market. It is interesting to note that the client-firms, on the other hand, decide to hire those assets based mostly on the prevailing short-term market forces. Hence, it is important for the companies which provide the assets for hire to also build the prevailing short-term market trends into their long-term forecasting and planning. In this paper, we develop a model for tracking these two simultaneously evolving and interacting patterns, namely the asset-availability (i.e. supply) and utilization (i.e. demand) patterns, in order to better understand the underlying processes, and thereby provide a basis for better forecasting. We test our models using three sets of data collected from the oil drilling industry, and find the proposed model to provide a good fit and forecasting efficiency.

Trying to understand the dynamics that underlie the demand process of a new product has always been of a primary interest for marketing researchers. However, this is actually of much more practical interest to the firms that supply the new product to the market, because they need to plan well in advance their production capacity, etc. Hence, forecasting becomes a critical factor for firms that seek to supply a new product. For established product categories, firms use techniques such as time series analysis to derive the long term market demand for the products. However, with new products such information is not available. Furthermore, new product sales cycles typically exhibit a diffusion pattern, starting off with very low numbers during the introduction stage, gaining momentum rapidly to far greater numbers during the growth stage, and reducing to a different range of sales numbers again when the cycle slows down. The great variation in the sales numbers during this cycle presents a big challenge to the suppliers. Fortunately, marketing researchers have shown that the demand for new products does have a systematic pattern of growth. Starting from Rogers (1962) and Bass (1969), many marketing researchers have advanced models which have explained the growth of new product demand rather successfully, and these models can be used by suppliers as forecasting tools. However, the impact of those sales growth models on the supply side dynamics has not been paid sufficient attention in the marketing research area. In fact, it did not get any attention until 1991, when Jain, Mahajan, and Muller (1991) extended the Bass model to explain the new product sales growth that is constrained by supply restrictions. Since then, the difference between “demand” and “sales” has become increasingly significant; the former denotes the expressed need for the product in the market, while the latter term denotes what is actually sold to the market. Jain et al. (1991) modeled the demand and sales growth of new telephone installations in Israel, which faced restricted supply. In their model, the supply was assumed to be such that only a certain constant fraction of the demand was able to be met by the supply at any one time. This in turn assumed a supply growth process that mimicked the demand growth, of course lagged by a few periods. The authors’ objective was to explain the observed demand pattern and estimate the supply-to-demand restriction constant. In a more recent paper, Ho, Savin, and Terwiesch (2002) used a similar framework normatively to analyze the way in which a supplier can devise an optimal production capacity in order to meet the demand growth of a new product that is expected to experience a typical diffusion pattern. Assuming a fixed production capacity, the authors set out to find out the optimal fixed capacity and the optimal product launch time, which is made possible through building up enough stock before launching the product. The authors point out that high-tech firms use this process of a pre-launch production build-up. Both of these models rely critically on modeling the demand under the condition of a restricted supply. While Jain et al. (1991) studied the demand pattern facing a constant supply-to-demand restriction, Ho et al. (2002) derived the optimal fixed production capacity, taking into account the fact that such a fixed capacity will introduce supply restrictions which will affect the demand pattern. The underlying assumption in these studies is that without such supply restrictions, the demand will follow a typical Bass-type diffusion pattern. Consider now a different product category which we explain briefly. In Fig. 1 and Fig. 2, we show the demand and the corresponding supply of this product over a period of time. There are three aspects of this case that differentiate the focal product’s supply and demand patterns from those captured by Jain et al. (1991) and Ho et al. (2002). During the period of observation, the demand never faced any supply restrictions; except perhaps for a brief time around January 1986, there was always ample supply in the market place to meet the demand. Secondly, the demand does not have a smooth diffusion-like pattern, even though it was not constrained by the supply. Thirdly, the supply curve has a smooth pattern that seems to capture the demand dynamics with some lag effect. For this product, it is difficult to apply the models developed in the extant literature, and hence, we need a new model to capture the supply and demand patterns observed in this product category.

We focused our research on modeling the hiring pattern of B2B services in a new market. B2B service companies are those firms that provide comprehensive services such as logistics for firms like HP or oil-well drilling services for oil firms like BP and Shell. An important component of the services is the renting of very expensive capital assets by the B2B service companies to their business clients. We argued that it is the service providers who plan with a long-term strategic perspective on how to make available their assets for hire, while their clients decide on the actual day-to-day hiring of these services based on many short-term fluctuations in the downstream market. We took the case of rig-services in the oil industry for our specific analysis. We found that the drilling rig companies, anticipating and estimating rig hiring, position their rigs in the oil field with a long-term strategic view in mind, while the oil firms base their day-to-day hiring decisions on the changing short-term market conditions including prevailing crude oil price and current hiring rate. We developed two models, one for each process, and tested them with three data sets and found the models to provide a good fit to all of the data sets, in turn suggesting that the proposed models are clearly one way of analyzing this supply and hiring pattern of capital assets in the B2B services sector. Our main contribution is three-fold. First, for the oil and gas industry, our model will come to be very useful because we not only use the commonly available data but also develop straight-forward modeling techniques to bring some new insights in the industry. Further, new markets (i.e. B2B service markets) are opening up across the world, especially in the huge developing economies such as China and India, and hence there is huge scope for the application of the proposed model. Secondly, for the marketing literature, we believe that our approach is the first one to analyze the service product in a B2B setting. Our two-stage model that enables one to track the supply side dynamics and the demand side dynamics within a unified framework is an important contribution to the literature, especially to the service literature. Thirdly, the empirical results are very insightful and could be used by the industry. The three key results, namely, the evolving capacity of oil fields, the impact of oil price on rig hiring and the insignificant impact of day-rate on rig hiring, and the significant impact of the ‘experience’ factor on capacity updating and well drilling, hold good across all of the three data sets. Noting that our research is just a first step in understanding this booming B2B services sector, we recommend the following as directions for further improvement. First, although we believe that the rationale behind our model would apply to other categories in this sector, it is necessary to understand the impact of the unique characteristics of each category on the model. Secondly, the rig supply and hiring pattern is only one important aspect of this industry. There are other key aspects such as contract length and day-rate determination. Future research should look into the determinants of these two key elements of the rig hiring process, and how these affect the rig supply and hiring patterns. Thirdly, the rig owners invest substantial amounts of money in building their rigs and it will be worthwhile to see whether one can apply the proposed model and make some recommendations to the firms. Fourthly, one interesting outcome of our research is that the estimation of the proposed model was able to throw some light on how the oil industry keeps updating the capacity of an oil field. One could do further research on this topic to understand the other key factors that affect the capacity estimation. This is a very sensitive issue, however, since an oil-firm’s stock price is partly determined by the ‘oil reserve’ they claim to maintain. For example, in 2004, the revelation from Shell that the company had overstated its proved oil and gas reserves by nearly 25% resulted in a sharp decrease in the stock price of Shell.21 This can also be useful for the various governments that lease out oil tracts. Finally, one could analyze whether the companies in the B2C service sector (car-rentals, hotels etc.) also face a similar problem when they enter a new and growing market.