Hey, just a few questions 1: We are noticing then...
# ask-questions
Hey, just a few questions 1: We are noticing then when we split by device we arent getting total reasons, do we know why this is? Is it because not all users have that data 2:How can I create a New vs Returning Dim to seg the data by? 3:Can someone please just breakdown in layman's terms what these 3 metrics are?
I am also struggling to understand how to calculate AOV, can anybody help ? 🙂
We are using Bigquery as a data source
1. what do you mean by 'total reasons' ? 2. You would have to create a metric for returning users. This is likely possible by looking for previous records in the data set. 3. Sure - Where are you seeing that data? It looks like just raw values for a metric - so how many were exposed (users) how many of that particular event happened (numerator) and the ratio of those two (numerator / users)
for AOV, you can just have a metric that selects purchase amount (total order value) with a sum aggregation, and then add a denominator metric on "purchased" - which would be a count metric for a purchase
1:sorry I meant total users 2:okay, so we wouldnt be able to split this out by dimension 3:okay awesome makes sense 4: sorry silly questions : how can I combine 2 metrics to form another metric ?
1. its hard to tell from here why the total users don't add up - if the randomization unit is not set, the user will not be included in the experiment, so that certainly could be it. 2. You can use this as a dimension, just requires a bit of work 4. When you add or edit a metric, you can chose a denominator metric
So when I device split we are getting : 8.8k users, but if we look at total its 15.4k users, what do you mean by randomization unit ? 🙂
Would it look something like this for CR of sessions for instance ? SELECT user_id, user_pseudo_id as anonymous_id, TIMESTAMP_MICROS(event_timestamp) as timestamp, 1 as value FROM
WHERE ( ( _TABLE_SUFFIX BETWEEN '{{date startDateISO "yyyyMMdd"}}' AND '{{date endDateISO "yyyyMMdd"}}' ) OR ( _TABLE_SUFFIX BETWEEN 'intraday_{{date startDateISO "yyyyMMdd"}}' AND 'intraday_{{date endDateISO "yyyyMMdd"}}' ) ) AND event_name = 'purchase'
Also another question Can we do pricing tests in Growthbook, we are on Shopify, so we want to AB tests new prices 🙂