Towards a learning curve theory for batch production

Consumer attitudes and fashion trends often impose severe constraints in manufacturing. When market needs for varied product designs lead to product competition for a place in the production schedule, batch size and available production time are depressed. Due to frequent set-ups and retooling, quality issues arise that hold back the line from a full ramp up. Consequently, the major advantage of quality through learning in mass production is often compromised. In this paper, we stipulate that in batch production the process of industrial learning becomes more complex and thus the archetypal exponential or power curve is not sufficiently rich to describe it. Specifically, we demonstrate that the learning phenomenon appears in double guise: both within and between batches. Actual production data from the fashion-conscious ceramic tile industry provide sufficient support for our premise about the dynamic nature of learning in batch production.