diff --git a/rally/benchmark/runners/base.py b/rally/benchmark/runners/base.py
index 922c881ce7..8b7a478d93 100644
--- a/rally/benchmark/runners/base.py
+++ b/rally/benchmark/runners/base.py
@@ -55,11 +55,11 @@ def _run_scenario_once(args):
             clients=osclients.Clients(context["user"]["endpoint"]))
 
     error = []
-    scenario_output = {}
+    scenario_output = {"errors": "", "data": {}}
     try:
         with rutils.Timer() as timer:
             scenario_output = getattr(scenario,
-                                      method_name)(**kwargs) or {}
+                                      method_name)(**kwargs) or scenario_output
     except Exception as e:
         error = utils.format_exc(e)
         if cfg.CONF.debug:
diff --git a/rally/cmd/commands/task.py b/rally/cmd/commands/task.py
index f4cf13512c..a1baafefce 100644
--- a/rally/cmd/commands/task.py
+++ b/rally/cmd/commands/task.py
@@ -235,11 +235,8 @@ class TaskCommands(object):
             # NOTE(hughsaunders): ssrs=scenario specific results
             ssrs = []
             for result in raw:
-                try:
+                if result['scenario_output']['data']:
                     ssrs.append(result['scenario_output']['data'])
-                except (KeyError, TypeError):
-                    # No SSRs in this result
-                    pass
             if ssrs:
                 keys = set()
                 for ssr in ssrs:
diff --git a/tests/benchmark/runners/test_base.py b/tests/benchmark/runners/test_base.py
index c7166c3233..50ac015b28 100644
--- a/tests/benchmark/runners/test_base.py
+++ b/tests/benchmark/runners/test_base.py
@@ -75,14 +75,14 @@ class ScenarioHelpersTestCase(test.TestCase):
         args = (1, fakes.FakeScenario, "do_it", context, {})
         result = base._run_scenario_once(args)
 
-        expected_reuslt = {
+        expected_result = {
             "duration": fakes.FakeTimer().duration(),
             "idle_duration": 0,
             "error": [],
-            "scenario_output": {},
+            "scenario_output": {"errors": "", "data": {}},
             "atomic_actions": []
         }
-        self.assertEqual(expected_reuslt, result)
+        self.assertEqual(expected_result, result)
 
     @mock.patch("rally.benchmark.runners.base.rutils")
     @mock.patch("rally.benchmark.runners.base.osclients")
@@ -93,14 +93,14 @@ class ScenarioHelpersTestCase(test.TestCase):
         args = (1, fakes.FakeScenario, "with_output", context, {})
         result = base._run_scenario_once(args)
 
-        expected_reuslt = {
+        expected_result = {
             "duration": fakes.FakeTimer().duration(),
             "idle_duration": 0,
             "error": [],
             "scenario_output": fakes.FakeScenario().with_output(),
             "atomic_actions": []
         }
-        self.assertEqual(expected_reuslt, result)
+        self.assertEqual(expected_result, result)
 
     @mock.patch("rally.benchmark.runners.base.rutils")
     @mock.patch("rally.benchmark.runners.base.osclients")
@@ -110,13 +110,13 @@ class ScenarioHelpersTestCase(test.TestCase):
         args = (1, fakes.FakeScenario, "something_went_wrong", context, {})
         result = base._run_scenario_once(args)
         expected_error = result.pop("error")
-        expected_reuslt = {
+        expected_result = {
             "duration": fakes.FakeTimer().duration(),
             "idle_duration": 0,
-            "scenario_output": {},
+            "scenario_output": {"errors": "", "data": {}},
             "atomic_actions": []
         }
-        self.assertEqual(expected_reuslt, result)
+        self.assertEqual(expected_result, result)
         self.assertEqual(expected_error[:2],
                          [str(Exception), "Something went wrong"])