1\begin{table}[htbp]\centering\footnotesize2\caption{Audit of the rule-based category labels (500-listing stratified sample)}\label{tab:cat_audit}3\begin{tabular}{lrr}\toprule4Rule label & Listings with annotator consensus & Precision of the rule label \\\midrule5Maple \& honey & 24 & 0.50 \\6Fine grocery & 24 & 0.42 \\7Coffee \& tea & 23 & 0.35 \\8Chocolate & 24 & 0.62 \\9Bakery & 25 & 0.40 \\10Meat \& fish & 22 & 0.32 \\11Cheese \& dairy & 24 & 0.58 \\12Beverages & 24 & 0.17 \\13Alcohol & 25 & 0.28 \\14Beauty \& care & 24 & 0.62 \\15Home \& decor & 22 & 0.36 \\16Fashion & 25 & 0.72 \\17Jewellery & 25 & 0.80 \\18Art \& craft & 25 & 0.52 \\19Children & 21 & 0.62 \\20Pets & 24 & 0.29 \\21Outdoor \& sport & 19 & 0.32 \\22Garden & 25 & 0.84 \\23Industrial & 24 & 0.42 \\24Other & 25 & 0.00 \\25\textbf{All} & 474 & \textbf{0.46} \\26\bottomrule\end{tabular}27\begin{minipage}{0.95\textwidth}\vspace{2pt}\footnotesize\textit{Notes:} 25 random listings per category annotated independently by two LLMs (claude-opus-5-5, gpt-5.5) with the platform's taxonomy; agreement rule vs A 0.45, rule vs B 0.45, A vs B 0.95; consensus exists for 95\% of listings. Precision = share of consensus labels equal to the rule label. Disagreements listed in the replication package.\end{minipage}28\end{table}29