It’s the Big Data Era
When most of the business schools still teaching reach, frequency, and ratings, people have been shifting their attention from unmeasurable media to digital media. For TV buyers, they pay a lot to Nielsen for their survey data about TV ratings, which is not so precise even though Nielsen claims that they have a panel of 50,000 people representing entire American population. Statistically speaking, even though well-sampled data can represent the entire population, sampled data can never tell as much information as entire population data can do.
Big data has been a hot topic in marketing nowadays. To a marketer like me who has been hunting a job for a while, the most obvious fact is that there are more entry level analytic positions open to fresh graduates. It is hard to image that Google launched Google Analytics in 2005, the same year when the concept of big data was introduced and a year after Facebook was launched.
To me, the definition of big data is literally big, often real-time generated, and sometimes not in a unified format. The unprecedented amount of data provides a whole new approach to understanding people’s behaviors. Even though privacy is a big concern of many people, there’s really nowhere to hide on the Internet. Analytic tools use not only cookies but also so-called fingerprinting to identify unique users. These tracking features enable marketers to precisely target their target audience more easier than ever before.
Big data provide marketers with real-time decision support. It is possible now to drive some instant sales by showing ads to the people who are more likely to purchase. Big data also reveal what are the inner connections between behaviors, for example, a classic story about beers and diapers. Many retail chains have implemented these insights in their shelf management.
Be Aware of Those Limitations
Before we implement what big data tell us to do, we must know some of its limitations.
Firstly, the marketing moves that based on the result of data can sometimes not align with company’s brand strategy. Secondly, sometimes customers are annoyed because being fed too many same or similar ads. Thirdly, not all these suggested moves are reasonable because data along only tell about the correlations rather than true behaviors, for examples, people spend comparatively long time on interstitial ads not because they like these ads but because they spend time on looking for closing the ads. Lastly, data are not always reliable, for example, not only human beings create website visits but also bots.
It is also not wise to make future decisions solely based on big data. Since people all have experienced what’s the consequence of totally believing in weather forecast and in stock forecasting models, we must face the fact that data cannot tell much about long-term forecasting. It is not wise to totally believe forecasting models because previous data along can only tell the trend but not the unexpected incidents, like what will happen to United Airline after they have forcibly dumped a customer.
- Focus on objectives, both short-term and long-term.
- Don’t let big data hijack your decision making. Let human beings review the result first.
- Data can us tell nothing without good interpretations. Hire the people who are capable of doing the analysis.