Overview:
DataMilk customers can use Google Analytics (GA) to independently verify DataMilk’s data and results. While it is expected that GA may measure KPIs differently than DataMilk’s analytics for known reasons (What to expect if I compare GA to DataMilk) , it is also possible that GA is producing inaccurate results.
This article provides a list of various issues that DataMilk has identified across its portfolio of customers.
1. Outliers by value
GA may not be removing outliers by value, which can lead to issues in your reporting. DataMilk automatically classifies outliers (How does DataMilk identify and remove outliers?) by order value and behavior.
2. Checking how GA calculates Ecommerce Conversion Rate
Ecommerce Conversion Rate should not be higher than 100%. Still, DataMilk has observed this happening in GA when more than one conversion event is generated by the site per session. Sometimes this is caused if a customer returns to the confirmation page or reloads the page, it’s possible that the same event could trigger again, and a duplicate conversion could be counted for the same order. DataMilk automatically removes duplicate conversion events for the same order.
See below to see how to identify this kind of issue in GA:
a.) Multiple purchase transactions per session or purchase transactionId
GA may be counting multiple transactions per session, which is uncommon user behavior, and/or counting multiple transactions per Tracking ID which can cause duplicate conversions. Transaction IDs are unique identifiers for each purchase transaction. DataMilk only attributes Purchase Transaction IDs once for a session and also discards previously seen pixels and suggests setting up similar filtering in GA as a solution.
An example of multiple transactions per session in GA: A single Session has counted 8 transactions which gives a skewed result of 800% conversion rate.
An example of multiple transactions per trackingId in GA: A single Tracking ID incorrectly has 5 transactions and 3 sessions which gives a skewed result of 166.67% conversion rate
3. Checking traffic sources and their conversion rates
It is important to be aware of traffic sources in GA to identify any prolific buyers who are visitors who perform an unusually high number of product searches or make an unusually high number of purchases. It is important to filter out these non-typical behaviors from GA reports as they falsely skew KPIs for the majority. The majority are the shoppers that generate the most revenue for the site and hence its important to get a clear picture of those shoppers in order to make the most effective decisions without being skewed by non-typical behaviors.
One source of such behaviors can be store personnel, who are using the site to help shoppers who come into the store.
Here is an example of prolific buyers from a GA report showing sources and their conversion rates. ‘example.sourceA’ shows an unusually high number of transactions and conversion rate.
4. $0 transactions being counted
GA may be counting orders with zero dollar value as transactions that may impact KPIs such as conversion rate and/or average order value. DataMilk recommends manually filtering these out in GA Integration as they are typically incorrect. DataMilk automatically cleans data for this type of issue and excludes it from its metrics.
Here is an example of a GA report that shows multiple $0 transactions being counted towards conversion rate.