What is data analysis, exactly?
Data analysis is the act of reviewing information and deriving insights.
It encompasses many different aspects including cleaning, transforming, modeling, and inspecting data.
Today, we are focusing on data inspection.
Of course, we will make the assumption that you have already used Funnel, or a similar tool, to clean and transform your data so it is ready to inspect.
Why do marketers need to perform data analysis?
The world of advertising is increasingly dominated by digital. With that change, a marketer’s required competencies have shifted, and now require an understanding of data.
With marketers everywhere competing with their ad spend on similar platforms, the way to get ahead is by gaining a deeper understanding of what is driving performance. It is through analyzing marketing data that marketers can gain a holistic understanding and create new hypotheses to test. It is through testing and iterating that modern marketers improve performance.
Charts as a type of data analysis
Data can be communicated visually via tables or charts. These visualizations can be simple or complex depending on what the goal is.
One way to get started with data analysis is to create some simple charts that allow you to spot anomalies or patterns in your data. This is the first step toward understanding your data and being able to work confidently with it.
Once you are capable of creating simple charts, you can increase the complexity of your hypotheses and build more complexity into your charts.
Let's walk through the 3 chart types mentioned in the video:
Time series analysis
Time series analysis
A time-series analysis enables you to see the change in a metric over a specific period of time. This is a great way to start spotting trends or even seasonal changes.
With this analysis, you should analyze metrics on a daily, weekly, or monthly basis. Different metrics require different levels of granularity. This means daily reviews over a short time period might work well for one metric, but not necessarily another. Be sure to consider this when creating your hypothesis.
As you get to know your data better, you will begin thinking of new hypotheses about what impacts your key metrics most. One way of exploring these hypotheses is a correlation analysis.
Let’s say you hypothesize that, when you spend more on Facebook Ads, your conversions on Google increase. A simple correlation analysis could tell you if that is true. But be careful! As they (you know… the “math gods”) always say, “correlation doesn’t always mean causation.” In other words, just because conversions on Google increase at the same rate as the increased spend on Facebook, doesn’t mean they go up due to the increased facebook spend.
It may just be a coincidence, but it is still a good first step to exploring causation.
One of the most common questions posed to marketers: “Which campaign performed better?”
A comparison analysis helps to answer this age-old question by allowing you to compare two or more metrics or dimensions against each other. This is incredibly useful in a marketer's daily toolbox to make cross-channel comparisons and elucidate opportunities for deeper exploration. A great example is to look at the performance of two campaigns and ask, “Why did one campaign perform better than the other?”
For marketers, it pays to get familiar with your data and gain a holistic understanding of it. The best way to do that is to start creating hypotheses and trying to answer them by exploring your data. So get started today, and start performing cross-channel analysis.
Want to download this information with all the examples charts from the video?