As a lover of forensic science, I know how the smallest clue can help solve the biggest puzzle. And that’s why I’m excited about how data has become such an integral part of the advertising equation. Data and science are inseparable. The definition of science is turning knowledge into the form of testable predictions. But where do you start?
Ask a Question
It begins by understanding what are you trying to learn and how will it be measured. The more specific you can be, the better. The answer to which customer segments are most profitable will provide better actionable opportunities than how well are we doing?
What data are you collecting? Many companies collect lots of data, but is it the right data? Smart data helps solve business challenges. To get the best insights you have to start by gathering valuable information. Look for trends and correlations. Marrying both quantitative and qualitative data can also produce beneficial results.
Construct a Hypothesis
Creating dashboards and reports is very different from hypothesizing, analyzing and optimizing your data. Think about this step as how can the past build a better future.
Test your Hypothesis by Experimenting
Develop an educated guess, or a prediction that can be tested. Collect as many observations as possible about the challenge you are trying to examine. Get into the market quickly. Consider testing various customer segments and offers, and then optimize accordingly.
Analyze Your Data and Draw a Conclusion
We’ve all heard the statement “garbage in, garbage out.” That couldn’t be more true than in the data world. How clean is your data? Prepare for this step by collecting complete data that provides the most valuable insight into your customers' behavior.
Communicate Your Results
Does your company have a data-driven culture, or is the data tightly held by a few people? For data to truly impact your success it must be embraced, respected and shared. Whatever you do, don’t manipulate the facts to fit an agenda.
Here's an example of how the process works in real life:
We were challenged to increase sales for an e-commerce client whose products fall within the restaurant and packaged goods categories. Going into the analysis, we hypothesized that developing data-driven messaging by the various customer segmentations would generate positive results. The data was analyzed to determine purchase frequency, as well as what was being purchased. Our recommended customer segments were based on frequency of purchase: those who purchased only once, those who purchased once a year, and those who purchased multiple times a year. The constant in this test was the product being offered – which was the client’s biggest seller. We varied the actual discount by segment, giving those who purchased the least a more aggressive incentive. All segments responded favorably and revenue more than doubled vs. their most successful past promotions. This “test and learn” exercise helped the client realize how combining data with marketing could build his business.
As in science, data analysis should become the foundation for all business decisions. Think of it as real-time perspective of your customers’ behaviors. Strong analysis can smartly influence product, promotion, price, place and people.