3 Ways Big Data Is Disrupting Marketing

Organizational inertia has always resulted in decision making processes lagging behind in the face of rapid change. The ability to track user behavior online has resulted in there being an enormous amount of data, and this new reality is often incompatible with decision making processes that belong to a world we no longer live in.

1. Predictability of Ad Spend and ROAS

Back when marketing did not have to be prefixed with offline or online, advertisers knew exactly how much advertisements cost, and had only a rough idea at best of what return could be expected. The most sophisticated advertisers could carry out marketing mix modeling to align spend with expected return using statistical analysis, but the feedback loop was long, spend was almost perfectly predictable, and return barely was. This meant that advertisers conducted such analyses only periodically at best, limited by availability of data; they decided a budget, with much control over cost, tried to gauge and attribute return to marketing, and adjusted the budget accordingly for the next period.

This relationship between the predictability of cost and return has practically reversed, or is on the verge of reversing, which has obvious implications on marketing decision making processes.

By design, the combination of online ad auctions, tracking of online user behavior, and the ability to attribute the former to the latter means that return is increasingly predictable while cost is less so, compared to the past.

2. Analysis Paralysis

With the limited amount of data available in the past, the cost of analysis paralysis to marketing organizations was little to negligible and perhaps even beneficial at times.

However, when it comes to the amount of data available today, the question is not how much water there is in the oceans; the question is, how much can you drink? The analytical techniques needed to work with this amount of data are very different from what was needed in marketing organizations in the past. The figure below indicates search volumes for the phrase "data science" between 2009 and 2015:

Searches for the phrase "data science" between 2009 and 2015

Rather than having marketing managers conduct statistical analyses, data scientists are needed to create systems that do that, constantly, in real-time.

This means neither the elimination of manual decision making nor a dystopic future in which human beings are replaced by machines; rather, with the number of metrics now exposed to marketers, it is all the more important to identify and understand the most relevant ones for your particular business.

The systems referred to above are still subject to garbage in garbage out.

3. How Much Of That Budget Can You Spend?

CFOs would perhaps not be too happy if I suggested that we get rid of budgets altogether, because they do serve an obvious purpose with regards to financial management within an organization. However, the way they are handled within marketing organizations is evolving.

While allocation of budgets and resources has always been an area where there is a need for stability, the number of channels has also increased significantly. Within online marketing alone, there is at least Google Search, Google Display, Bing Search, and Facebook, among others.

Combined with the points mentioned earlier, this means that marketing organizations need to be structured in a way that allows for a frictionless and data-driven allocation of budgets. This is certainly not to say that actual revenue is perfectly predictable or that budgets can be allocated perfectly, for there are pieces of the puzzle that are still missing (cross-channel and intra-channel attribution being two examples).

Marketing has not changed rapidly; it is changing rapidly, and agility and responsiveness are key for organizations. As Charles Darwin never actually said (but is of course credited as such): it is neither the strongest nor the most intelligent that survive, but rather those most responsive to change.