By Maggie Petrella, Senior Digital Data Analyst
Today, nearly everything online is trackable, which means the world of digital data is almost endless. For organizations committed to sharing their message through digital media and marketing, this vault of information is invaluable – and sometimes overwhelming. Knowing how to track, understand, and use the multitiude of facts and figures is critical to success. At Hatcher, we know how to do just that, and we provide that expertise to keep our clients informed on how to engage their target audiences effectively. In other words, when the digital data speaks, we listen.
Here’s a primer on our various analytical approaches.
Data analysis falls into four types: descriptive, diagnostic, predictive, and prescriptive.
When we think about the uses of data analysis in marketing, typically the first thing that comes to mind is the descriptive approach: reporting on the performance of past and ongoing campaigns. A descriptive analysis tells the story of what happened — the count of shares on a LinkedIn article, the number of views of a video ad, the cost per click on a Facebook post. Descriptive is the most basic form of analysis, but more complex analyses cannot happen without having this foundation to build on. That’s why the majority of analytics done is descriptive.
The second type of analysis is diagnostic. A diagnostic analysis separates confounding information and explains the results. If descriptive analysis describes what happened, the diagnostic analysis conveys why. An example of a diagnostic analysis would the impact of programmatic ad placement on click-through rates.
Descriptive and diagnostic analyses are both backward-looking, or historical: They describe and explain events that happened. This is the foundation of data analytics and what we use when we write reports that describe and quantify activities for a campaign. The remaining two types of analysis, predictive and prescriptive, look toward the future.
Predictive analysis uses historical data to estimate what will happen by building on the foundational understanding established through descriptive and diagnostic analyses. Predictive analysis is done by developing straightforward regression models (a technique that investigates the relationship between a dependent and an independent variable) based on historical trends and patterns, or by using more advanced methods like machine learning. One example of predictive analysis is a forecast that identifies the number of clicks expected from a future campaign. Predictive analyses are particularly useful when setting key performance indicators and metrics of success for a campaign. At the end of the campaign, comparing actual results with expected results can offer valuable insights.
Now that we know what happened (descriptive), why it happened (diagnostic), and what we can expect to happen next (predictive), the last and perhaps most the most compelling analysis will answer the question, “Now what?”
A prescriptive analysis uses the understanding developed by the prior three analyses as a foundation for suggesting (or prescribing) courses of action. It might recommend certain design features to implement or specific audiences to target to maximize views, clicks, shares, or any other metric.
At Hatcher, we combine all four types of analysis to draw valuable insights from the expanse of digital data. This approach helps us understand what is going on and why, what to expect, and how we can affect that future for our clients. With these analytical tools, we can leverage data to effectively communicate your message to those who need to hear it, and reach a breakthough in achieving your goals.