Date: May 11, 2021 By: Alicia Lin
Why raw data?
It seems so much easier to look at the Facebook Analytics or Google Analytics pre-prescribed charts and tables to view your marketing data on ad performance. However, there is great value in simple queries formed from owning your company's raw data.
Raw data is essentially the key to endless opportunities in terms of the potential insights gained from data collected. In this blog post, I will be going over a simple train of logic that provides great insight. This is just one out of infinite queries that could be performed.
In the table below, the query counts the number of events by the hour and the total value associated with an event and the utm_campaign that is related to the event. With Facebook's standard events, the primary events viewed would be AddToCart, Purchase, and ViewContent. However, once again, with the ability to make queries with your raw data, this can easily be adjusted for custom events.
To demonstrate the flexibility, we can easily add the content_id parameter to the table, and now, not only are we able to have an idea about how specific events and their campaigns relate to the number of events and total values per hour, but we have an idea of which specific products are associated with them.
Tables are great and all; however, with thousands - even millions - of rows of data, it's difficult to grasp the big picture idea. In the table below, the ViewContent event is filtered by one campaign. We can see how well this specific campaign performs in terms of having customers view products throughout the 24 hour day. This is not specific to one day, but rather the total from each day added together. An insight for example is that this campaign does not perform as well between 1am-3am.
Again, this might not be useful to your specific campaign needs. To reiterate, the benefit is that raw data is highly customizable. For example, in the chart below, we can see the hourly performance, but separated by the months for a specific campaign and event.
Not what you're interested in? We can also do a daily breakdown by the hour and compare the days. This looks overwhelming, but can provide some insights. For example, we can see that between 9pm-11pm on the 7th of this particular range of dates, the campaign performed the best.
Charts that can be created from queries can fit your needs - no matter how complicated. Take for example this last chart. Simple. This chart dispays the performance of a specific campaign and the ViewContent event over time. The data is smooth because the data source provides data points by the hour. Evidently, this campaign performed better on average around April 30th.
Event instances don't necessarily correlate with value though. However, raw data lets you create connections! This chart shows the total monetary value contributed by all the ViewContent events associated with a specific campaign. While there were more events, around April 30th from the chart above, we can see that around May 9th, this campaign performed better in terms of the value translated by ViewContent events.
The moral of the story? Raw data is a door to endless insights about your business.
