Hey Growthbook Team :wave: Could anybody to help ...
# ask-questions
e
Hey Growthbook Team ๐Ÿ‘‹ Could anybody to help me understand why I don't see target and guardrails metrics data while running an A/B test? I'm wondering if it's the setup issue or the tool itself is designed to prevent peeking (if it's considered as peeking)? Context: in our company we're taking first steps in using Growthbook, so we've set up a small A/B test to check the whole flow (from setting up feature to a/b test evaluation and post-analysis). For this A/B test I'm trying to see my target metric and monitor guardrails.
f
just above this section on the right, there should be a list of metrics and guard rails,
we do hide results until a min number has been reached (which you can set in the metrics behaviour), but we donโ€™t hide metrics.
e
@fresh-football-47124
just above this section on the right, there should be a list of metrics and guard rails,
I've set up target and guardrail metrics for the a/b test. I've lowered sample size for this test (set up to 1), but still don't see results.
we do hide results until a min number has been reached (which you can set in the metrics behaviour), but we donโ€™t hide metrics.
Could it be something else? Like my metrics aren't connected to users or any other binding
f
So the guardrail metics are now showing up at all under the main results?
Iโ€™d be happy to jump on a call to help you figure it out
h
It looks a lot like the following is true, and the users in your metric source aren't being found in the experiment assignment source.
Copy code
Like my metrics aren't connected to users or any other binding
e
Hey ๐Ÿ‘‹ Thank you for your help! I've solved the problem yesterday and now see all the results after adjusting my queries.
f
great
e
I would also like to understand why should I add minimum sample size to the metric and not to the a/b test setup directly?
f
some metrics are noisier than others, so we let you adjust the threshold per metric.
The idea is that sometimes people people like to jump to conclusions based on very small sample sizes - and this lets you hide the results until you have more significant results
e
Amazing! Thank you very much ๐Ÿ™Œ bows
Hey @fresh-football-47124 sorry for such a long discussion, I have one more question and would be happy if you could help I'd like to understand what happens if I change experiment settings after launching the experiment? e.g I've set up a test, the results aren't shown yet and I'm altering target metrics behaviour. Would it affect my existing users or only the evaluation of the test?
f
adjusting metrics only effect the report- nothing in metrics will effect what the users are experiencing
๐Ÿ‘Œ 1
๐Ÿ‘ 1