- Colour sample manufacturing industry provides an important support for the paint manufacturers. It manufactures colour samples that help original paint manufacturers to sell paint by allowing potential customers to accurately visualise a specific colour. One of the burning issues in colour sampling manufacturing is accurately predicting the tally gallons. It involves an estimation of paint volume to cover a given surface area of paper. The as-is tally gallons estimation process is rudimentary and largely depends upon the human experience. Ideally, this quantity should be enough to cover the target surface area regardless of colour appearance. However, this is not the case with existing process. Currently, lighter colours run out in the middle of the production run while darker colour appearances have significant amount of left over paint. The amount of painting waste is as high as 15% (by volume) in some cases. The objective of this paper is to present a predictive model to better estimate the tally gallons by minimising painting waste. It presents a case study of a US colour sample manufacturing company. The results show that the proposed multiple linear regression approach reduces the leftover paint significantly. Copyright 2010 Inderseienee Enterprises Ltd.