In business-to-business (B2B) pricing, a seller seeks to maximize revenue/profit obtained from high-volume transactions involving a wide variety of buyers, products, and other characteristics. Buyer response is highly uncertain, and the seller only observes whether buyers accept or reject the offered prices. These deals are also subject to high opportunity cost, since revenue is zero if the price is rejected. The seller must adapt to this uncertain environment and learn quickly from new deals as they take place. We propose an approximate Bayesian statistical model for the win/loss probability as a function of offered price, which has the ability to measure and update the seller's uncertainty about the demand curve based on new deals. This model employs optimal learning and stochastic gradient search to update parameter estimates in the logistic regression based on binary (win/loss) data. We also consider an approach for recommending target prices based on the approximate Bayesian model, thus integrating uncertainty into decision-making. We test the statistical model and the target price recommendation strategy on both simulated data and real data from a leading consulting firm specializing in B2B pricing.