diff --git a/vizier/_src/benchmarks/analyzers/plot_utils.py b/vizier/_src/benchmarks/analyzers/plot_utils.py
index 2e4e3d143..3e4cde805 100644
--- a/vizier/_src/benchmarks/analyzers/plot_utils.py
+++ b/vizier/_src/benchmarks/analyzers/plot_utils.py
@@ -225,7 +225,6 @@ def _metadata_to_str(metadata: vz.Metadata) -> str:
               percentiles=(elem_for_metric.percentile_error_bar,),
               **kwargs,
           )
-          ax.set_xlabel('# of Trials')
         elif plot_type == 'scatter':
           plot = elem_for_metric.plot_array
           ax.scatter(
@@ -254,6 +253,7 @@ def _metadata_to_str(metadata: vz.Metadata) -> str:
           )
         else:
           raise ValueError(f'{plot_type} plot not yet supported!')
+        ax.set_xlabel(elem_for_metric.xlabel)
         ax.set_yscale(elem_for_metric.yscale)
         ax.yaxis.set_major_locator(mpl.ticker.LinearLocator(20))
         ax.yaxis.set_minor_locator(mpl.ticker.LinearLocator(100))
diff --git a/vizier/_src/benchmarks/analyzers/state_analyzer.py b/vizier/_src/benchmarks/analyzers/state_analyzer.py
index df951780b..1d6d321b7 100644
--- a/vizier/_src/benchmarks/analyzers/state_analyzer.py
+++ b/vizier/_src/benchmarks/analyzers/state_analyzer.py
@@ -54,6 +54,9 @@ class PlotElement:
       default='error-bar',
       validator=attrs.validators.in_(['error-bar', 'histogram', 'scatter']),
   )
+  xlabel: str = attrs.field(
+      default='Num Trials', validator=attrs.validators.instance_of(str)
+  )
   yscale: str = attrs.field(
       default='linear',
       validator=attrs.validators.in_(['linear', 'symlog', 'logit']),