It’s been said that “…the greatest fool may ask, more than the wisest man can answer.” Sometimes the simplest of questions can prove to be among the most difficult to resolve. Certainly, the question of getting the maximum value out of marketing dollars invested has been around (and often perplexing) business leaders since at least the early 20th century, when John Wanamaker coined his oft-quoted lament: “Half the money I spend on advertising is wasted; the trouble is I don't know which half.”
With the advent of digital marketing, machine-based learning, and AI-Everything ™, technology has revived the hopes of marketers in their ability to quantify the value they and their campaigns bring back to the business.
Data Scientists can now employ complex algorithms based on Bayesian and Markov-chain statistical principles using historical interaction data and turn it into predictive capabilities for use in forecasting models. All of this makes it a great time to be in business, and in charge of a marketing budget, but there’s one thing all marketing managers need to remember as they embark on their quest to measure and optimize spend – all of the magic relies on clean, abundant data.
Getting Under the Hood
AI, ML, and algorithmic attribution modeling all embody the sexiest parts of marketing attribution. These can be considered as the pretty exterior, and shiny engine compartment of a high-performing automobile. However, there are other components that are just as critical but are often overlooked. As budget owners seek to calculate the return on marketing dollars invested, attention should be placed on the fuel feeding the engine, the systems that maintain adequate visibility, and the parts of the machine that enable a nimble response to the curves that lay ahead. All three of these in this analogy have an important common denominator – data.
Fuel for the Engine
Just like a car benefits from higher octane gasoline, attribution modeling live and die by the amount and quality of the data fed into its calculations. Marketing managers should be cognizant of the axiom “Garbage In, Garbage Out”. Clean, standardized data streams are necessary to provide consistent signal by which to drive downstream decisioning mechanisms that in turn provide value for people-based marketing initiatives.
A Clear View
High performance and elevated speeds require good visibility. Clean, abundant sources of data enhance planners’ resolution and understanding of audience segments and personas, and lead to better visibility into marketing activity. It allows managers in the driver’s seat to better segregate important behavioral ques from mere noise, make more accurate forecasts, and more confident decisions regarding the next best actions to take at an individual audience level.
Online marketing has its share of curves, potholes and traffic. In terms of marketing technology performance, data once again proves to be where the rubber hits the road. Good data enables precision targeting and audience segmentation. For those times when a re-allocation of marketing spend is required, a strong data environment enables a faster cadence of evaluation, and improved reaction time via features such as automated alerts, anomaly detection, and rule-based marketing tactics.
Ready to Roll
Advanced algorithmic attribution modeling is perhaps the best answer to the question “What is my Real ROI?” However, it takes time, effort and preparation to achieve that goal. Business managers can begin to take steps toward that high level of performance, while realizing more immediate benefits, by laying the proper foundation for their data environments.
As Jill Avery, a Harvard Business School professor who specializes in marketing optimization puts it: “Algorithms are fabulous as long as they are based on good assumptions and good data.” Sophisticated algorithmic analysis can deliver fine grained details about the value of each touch in the customer journey toward conversion, but those algorithms depend on a complete set of data. Otherwise, the picture they wind up painting can have significant gaps, or worse yet, misinterpret which factors deserve credit for influencing revenue.