This paper introduces a minimal warranty database where the claims are recorded based on the month of production instead of the month of sale. It models the warranty claims to estimate the parameters, the age-based expected number of claims and the probability of sales lag, where the dates of sale of the products are unknown. The parameters of this model are not estimable uniquely. Then it is proposed that a model based on follow-up information on the dates of sale, provides unique solutions for the parameters even for a small number of followed-up products. The properties of the estimators are investigated through asymptotic variances.