Business Analytics for United States Retail Channel Planning Under Structural Change

Md Jubayar Hossain (St. Francis College, Brooklyn, New York, USA)
Semanto Saha Utsab (Montclair State University, Montclair, New Jersey, United States)
Sadia Afrin Jarir (International University of America, Los Angeles, California 90010)
Evha Rozario (International University of America, Los Angeles, California 90010)
Zerin Akter Tanni (St. Francis College, Brooklyn, New York, USA)
Tushar Roy (St. Francis College, Brooklyn, New York, USA)

The Eastasouth Journal of Information System and Computer Science
Vol. 3 No. 01 (2025): The Eastasouth Journal of Information System and Computer Science (ESISCS)

Retail channel planning requires forecasts that remain useful when purchasing patterns change abruptly. This study examines the quarterly e-commerce share of United States retail sales using 99 observations from 2000 Q1 through 2024 Q3 in the data vintage available on December 31, 2024. Census Bureau estimates distributed through ALFRED provide the sole empirical data source. Six transparent specifications are compared at 59 successive one-quarter forecast origins using chronological estimation. Evaluation distinguishes 2010–2019, 2020–2021, and 2022–2024 Q3 and combines absolute error, squared error, directional bias, and asymmetric planning-loss sensitivity. The lowest mean absolute errors in these periods are 0.0755, 0.9375, and 0.1215 percentage points, respectively. Method rankings depend on the evaluation period and criterion. The findings support period-specific assessment, simple benchmarks, and explicit attention to planning consequences. Results use a fixed historical data vintage; they do not establish causal pandemic effects or realized firm-level savings.