A mid-size book printing company has been expanding its offerings to become more service oriented. The company wanted to help its customers to order the optimum number of books to be printed. Books can be printed on analog presses with low incremental cost per book but high setup costs and long turn-around time. Books supply exceeding the demand is a waste while insufficient supply leads to lost revenues. Books can also be printed on digital presses with high incremental cost but quick response to fluctuating demand. What should be the optimal order quantity and the optimal mix between analog and digital prints that will maximize profit for the customer?
After identifying the drivers of book demand with the client, Pandoras Analytics has developed a causal model to predict future book demand based on factors including reputation of the author, book genre, and economy. We have then quantified the historic variability of demand as probability distributions and designed a Monte Carlo simulation model to predict a range of possible book sales. All this was embedded into a profit optimization tool that identifies the optimal book order quantity and analog-digital mix based on margin per book, cost of analog prints, and cost of digital prints.
The book printing company invested in capturing historic data from its own systems as well as data from its customers. The company’s marketing director spent few days with Pandoras Analytics brainstorming potential sales drivers. The overall consulting engagement lasted two weeks and cost 15,000 EUR. The company’s customers can now meet additional demand leading to average 7,000 EUR higher revenues per printing order while reducing scrap expenses by an average of 4,000 EUR per order. The book printing company has now been able to improve its customer retention rate and attract new customers.
An IT company is developing and introducing new products to increase profit. Financial feasibility of these new products depends on development costs, future unit sales, expected prices, cost of sales, and operating expenses. All these factors can change over time - even before the product is introduced - and improve or worsen profitability. The client wanted to understand the range of possible outcomes before even starting product development and at several checkpoints throughout the development stages. This information was going to be used to optimize R&D investments by canceling unprofitable R&D projects early in the development cycle.
Together with the client, Pandoras Analytics identified the key levers determining the financial feasibility of a new product. We have then developed a robust net present value (NPV) Excel model to capture these levers as well as a risk matrix to identify the appropriate discount rate for the NPV analysis. The model was enhanced with Monte Carlo simulation to simulate the variability of key drivers. Probability distributions needed for the simulation were defined with the client based on client’s expertise. The tool enables now the IT company to assess the risk of negative project NPV early on and cancel projects before wasting R&D investments.
The client engagement took four weeks to complete and required client’s representatives from R&D, Marketing, and Finance to provide key parameters and review progress throughout the engagement. The NPV simulation model was developed for $30,000 including detailed model documentation. The IT company is spending about $3M on R&D every year and has now been able to reduce its annual R&D waste by $50,000 with the new tool.
A small sports apparel company has been expanding its customer base and wanted to better understand customer behavior and trends. It specifically wanted to know customers’ transaction volumes and values and their purchase patterns. It also wanted to be alerted in real time if a customer changes his or her purchase behavior significantly, especially if there is a risk of losing that customer.
Pandoras Analytics has analyzed the company’s historic sales data with data mining tools to visualize patterns, trends, and outliers. We have clustered customers into few distinctive groups based on their purchase frequency, types of preferred clothing, and spend. The client knows now its high value customers as well as products and designs that are more attractive. Together with the client we have also identified volume and value thresholds for an alert system, which is now part of regular sales reports.
The apparel company provided access to its sales data and cleansed some of the data that was corrupt or mislabeled. The firm owner spent few hours with Pandoras Analytics to review patterns and trends and to identify alert thresholds. The project took four days to complete and cost 6,000 EUR. Using the customer clusters to target its marketing spend, the apparel company has now been able to increase revenues from high value customers by 2,000 EUR per month.
How can we help you to improve sales, reduce cost, or optimize investments?