diff --git a/pymc_marketing/mmm/mmm.py b/pymc_marketing/mmm/mmm.py
index c5c32f441..71d4c1772 100644
--- a/pymc_marketing/mmm/mmm.py
+++ b/pymc_marketing/mmm/mmm.py
@@ -1929,9 +1929,12 @@ def sample_posterior_predictive(
             if extend_idata:
                 self.idata.extend(post_pred, join="right")  # type: ignore
 
-        posterior_predictive_samples = az.extract(
-            post_pred, "posterior_predictive", combined=combined
+        group = (
+            "predictions"
+            if sample_posterior_predictive_kwargs.get("predictions", False)
+            else "posterior_predictive"
         )
+        posterior_predictive_samples = az.extract(post_pred, group, combined=combined)
 
         if include_last_observations:
             posterior_predictive_samples = posterior_predictive_samples.isel(
diff --git a/tests/mmm/test_mmm.py b/tests/mmm/test_mmm.py
index 1df4ed0a2..7617a9fd9 100644
--- a/tests/mmm/test_mmm.py
+++ b/tests/mmm/test_mmm.py
@@ -897,6 +897,30 @@ def test_new_data_sample_posterior_predictive_method(
     )
 
 
+@pytest.mark.parametrize(
+    "predictions",
+    [True, False],
+)
+def test_sample_posterior_predictive_with_prediction_kwarg(
+    generate_data,
+    mmm_fitted,
+    predictions: bool,
+) -> None:
+    new_dates = pd.date_range("2022-01-01", "2022-03-01", freq="W-MON")
+    X_pred = generate_data(new_dates)
+
+    predictions = mmm_fitted.sample_posterior_predictive(
+        X_pred=X_pred,
+        extend_idata=False,
+        combined=True,
+        predictions=predictions,
+    )
+    pd.testing.assert_index_equal(
+        pd.DatetimeIndex(predictions.coords["date"]),
+        new_dates,
+    )
+
+
 @pytest.mark.parametrize(
     "model_name", ["mmm_fitted", "mmm_fitted_with_fourier_features"]
 )