In 2004, Michael Lewis took note of how Oakland Athletics Coach Billy Beane adopted an analytical approach to baseball. In his book "Moneyball," Lewis wrote that "People…operate with beliefs and biases. To the extent you can eliminate both and replace them with data, you gain a clear advantage." Lewis was talking about how the Athletics put together a winning baseball team with one of the lowest budgets in the league, but the lesson, if applied correctly, works equally well for marketing.
Given what a Major League Baseball coach achieved with the help of big data, it's shocking that some marketers have still not yet embraced the concept of granular analysis to capture comprehensive customer marketing interactions in order to make more strategic, cost-effective media-buying decisions. Many in the industry still believe in aggregate data and base decisions on concepts like the advertising outlet that has the biggest audience or the highest click-through rate is the best bang for every company's ad buck. This is a profit-killing fallacy, and when marketers learn to embrace big data and true granular analysis over gut feelings, they'll stop falling for it. Armed with big data and the tools capable of pulling insights and recommendations out of it, marketers can be more strategic with their budgets and therefore generate more revenue.
With recent breakthroughs, we are awash in online and offline data that can deliver a better understanding of what really drives conversions. As consumers increasingly interact with advertisements across channels, devices, and time, marketers who use tools that can track, analyze, and capitalize on these interactions will be more successful and reap greater profits.
The common, expensive misconception about aggregate dataUndoubtedly, many marketers shy away from big data because the term itself is so daunting: big data. The words alone conjure visions of complexity and cost. Wouldn't it be cheaper and simpler to stick with aggregate data analysis? Simpler, perhaps. Cheaper? No; not if you factor in the massive expense of lost opportunity and costly actions based on incorrect information.
Aggregate data is essentially big data's opposite. Instead of using a platform that tracks and inserts every user interaction (clicks, views, conversions, and other actions) into a database and then mines this data for insights, big data doubters opt for a less computationally intensive approach that simply counts individual elements without putting them in accurate context. In fact, aggregate data provides little in terms of tactical optimization.
This occurred to me recently as I examined the television performance of a client using our platform. The company's media plan crossed a range of cable broadcast stations from the smallest to the largest. The largest stations provided the most revenue, but at the highest cost. The smallest stations couldn't deliver the volume of the larger ones, but they often produced the most profitable investments. Despite that distinction, the conversions delivered by these smaller stations numbered only in the single digits. An aggregate data approach would capture the sale, but not its source. And the marketer would only know that in aggregate, television either works or it doesn't. The marketer would lose any insight related to whether larger or smaller stations were more effective.
The "Billy Beane School" of granular data analysis
Unlike aggregate data analysis, big data can tell a marketing professional that one particular station is not only profitable but exceptionally profitable -- perhaps more so than the larger station whose aggregate performance pops off the page. The resultant recommendation that comes from this insight is to deploy more money into the smaller channel, subject to audience size and diminishing returns. When you multiply these insights across hundreds of smaller stations, the impact on the marketing return on investment can be significant. Each small channel adds more customers, more efficiently and with less work, since an algorithm in the platform extracts the recommendation from granular-level data.
Granular data analysis tells marketers what really matters. Traditional marketing theory holds that consumers need to be touched multiple times prior to a conversion. As consumers increasingly interact cross channel, cross device, and cross time, the number of touchpoints has only multiplied. For purchases where the average order value exceeds $200, the average number of touchpoints is greater than eight. Before you can determine the proper attribution to each of these marketing touchpoints, you have to make sure that you know about all of them.
Here's an example of how that knowledge might play out in favor of the marketer using a big data based platform. Let's say a major sporting league deployed a large television campaign to drive consumers to purchase a subscription product on its website.
- Forty percent of the response within one minute of airing the television spots occurred on mobile devices.
- Sixty percent of the response within one minute of the spot occurred on laptops and PCs.
- Ultimately, 1 percent of the total visitors who initially responded converted, albeit on average three days later and 95 percent on laptops and PCs.
For most of these conversions, a subsequent marketing event such as search or search/retargeting took place before the eventual purchase. The marketer who leverages big data knows that users first responded to the advertisement on a mobile device, laptop, or PC and then went to a laptop or PC on average three days later and searched for a keyword prior to purchasing. Proper analysis shows the marketer that users are not responding to search itself, but rather to the original TV spot that initially sparked their curiosity and decision to seek out more information. This understanding is critical to understanding which channel creates demand and what is the best use of the next marketing dollar. Investing more in search wouldn't produce more customers, but investing in television would.
Valuable analysis relies on big data
Without respondent-level data, marketers wouldn't be able to tie user behavior across multiple devices. They wouldn't understand, as in the above example, that TV might be the primary motivator, and they wouldn't know how to convert the customer. Unfortunately, there is still a mistaken belief among some in the industry that aggregate-level data can get the job done. This is wishful thinking. Quoting baseball statistician Pete Palmer, "Moneyball" author Lewis writes, "Managers tend to pick a strategy that is the least likely to fail, rather than to pick a strategy that is most efficient…The pain of looking bad is worse than the gain of making the best move." Marketers can have both the most efficient strategies and the best moves when they invest in platforms that are based on big data.
On Twitter? Follow iMedia Connection at @iMediaTweet.
"Machine Generated Data" image via Shuttersrtock.