The complex geology, geomechanics and flow mechanisms make it difficult to predict the drainage area of shale wells, and thus the estimated ultimate recovery (EUR) from these reservoirs. A large uncertainty in predicting the drainage area and EUR can result in a sub-optimal development plan - a key factor driving the economics of such reservoirs. To increase our confidence in the drainage area and EUR predictions in the Haynesville, we applied an integrated production analysis approach using multiple data types from a single multi-frac horizontal well with a rich dataset. The analysis integrates Flowing Material Balance (FMB), Pressure Transient Analysis (PTA), Rate Transient Analysis (RTA), Decline Curve Analysis (DCA), and analytical/numerical reservoir simulation. Using detailed well completion and production logging data, we created nine alternative models that honored the historical production and flowing pressure data. These alternative forecasts were generated using readily available commercial software that includes the Multi-frac Horizontal Composite (MFC) model, the Enhanced Fracture Region (EFR) model and the General Multi-frac (GMF) model. In this paper, we demonstrate how certain models used routinely in the industry may be too simplistic, and how these models can be progressively improved as more production and flowing pressure data become available with time. Our results demonstrate the following: • Permanent down-hole pressure gauges, hourly metered production rates and production logs provide high frequency historical data that enable more insightful PTA, RTA and detailed non-uniform fractured reservoir simulation models; • Inclusion of higher frequency data and geomechanical (i.e. pressure dependency) effects are critical for both PTA and RTA in shale wells. This is especially true where permeability dependence on pressure has been documented, such as in the Haynesville. PTA can be useful in constraining the history-match but the results are still non-unique; • Routinely used uniform fracture half-length/evenly-spaced fracture analytical models may be too idealized for optimizing well spacing and completion design. Numerical reservoir simulation models with more detailed and realistic reservoir characterization and fracture dimensions are therefore key in planning appropriate well spacing; This work provides valuable guidelines for data acquisition and prediction of shale gas EUR and optimal well spacing.