Co-Founder/CEO of Vujà Dé Digital, on a mission for conscious capitalism and reinventing the media agency model.
getty
While marketers have been grappling and struggling with solutions to the loss of third-party cookies over the past couple of years, new and smarter ways of investing media dollars have been evolving. As the cookie deprecates, it will be more important than ever to track triggers to desired marketing behaviors.
How do you know if you’re reaching the right consumers or whether your media and marketing mix is working? Simply waiting for the cash register to ring isn’t enough, and media mix modeling (MMM) or multi-touch attribution (MTA) is going the way of the dinosaur. What’s replacing it? I believe it is big data media mix modeling.
The latest solution to smarter media buying and performance analytics can be found with big data media mix modeling. It’s becoming the more sophisticated way to monitor and predict future marketing performance and guide media investments for the highest ROI and ROAS.
The democratization of machine learning, the processing power of big data on virtual machines, and access to Google and Amazon big data analytics have changed the game of strategic media investments and campaign performance forecasting.
Where does big data media mix modeling have its roots? Direct response television (DRTV) was the prototype for BDMMM. This was the beginning of the unification of linear media investments with digital campaigns.
So, what makes big data media mix modeling so powerful compared to previous options for campaign analytics and media planning? Big data media mix modeling provides a unified field view of marketing. It shows your media investment and campaign performance bringing together all your marketing, economic, political, seasonal and even retail data together in one view.
The beauty of these elements coming together is that you get a true performance view of what happens over time with every possible variable. Big data media mix modeling allows marketers to see the direct sales correlation when adjusting channel investments across media types, seasons and investment levels. The insights can also get smarter over time with the incorporation of machine learning.
AI correlates spending levels to show you where you’re over-invested so that you can cut back or reinvest in higher-performing channels to achieve a higher ROI. For example, rather than using a gut instinct and a couple of data points to increase your spending on Instagram, it can tell you that TikTok overperformed with a direct correlation to other media elements.
Big data media mix modeling can tell you the recommended media mix and where to increase or decrease spending based on actual performance and the interplay of all variables tied to revenue and sales performance. Its aim is to offer less guesswork and more calculated real-world decision-making based on statistical facts and not conjectures.
What’s wrong with multi-touch attribution (MTA)?
Tagging issues have been the downfall of MTA. This tactic always breaks down with tagging issues, which become onerous across all the various media types. You need one tag for each media type, including brand and non-brand, and with the walled gardens of Google and Facebook: You’re never getting a complete 360 view of your campaign performance across media types and investment levels. Marketers are forced to make decisions with partial or incorrect data, which makes continuous improvement a crapshoot.
What about media mix modeling (MMM)?
While it’s not as onerous or expensive as big data media mix modeling, it’s not good for deep analysis. Big data media mix modeling reports data at computer speed in a shorter time frame, and in multiple increments with an extremely high confidence level. You can model it against your KPIs. Neither MMM nor MTA can boast that output.
So one could ask, why wouldn’t marketers flock to this smarter approach to media planning and buying? A few things can make it difficult to implement, even though it is by far the most sophisticated marketing analytics and media planning strategy and tool available to marketers today.
Cost, media investment required to calibrate the model, and skill set required to implement are three of the most common barriers to implementing and executing big data media mix modeling. You need access to a data scientist or data engineer, a few hundred thousand dollars for access to Google or Amazon cloud data management and analytics services, and a meaningful ad budget, over a minimum three-month window.
The only other potential challenge can be getting access to retail data if you’re a CPG brand. This can be overcome by paying to get the data to have a full view of performance. The reality is that BDMMM won’t be as effective if you’re missing a significant percentage of your retail sales data.
The keys to big data media mix modeling success lie in access to accurate and as close to complete data as possible. Partial data or data in disparate places that are not easily pulled together in a timely manner can lead to an inaccurate model.
The power and accuracy of solid big data media mix modeling will pay for itself over and over again. As the saying goes, no pain no gain.
Forbes Agency Council is an invitation-only community for executives in successful public relations, media strategy, creative and advertising agencies. Do I qualify?
