Going Forward in Reverse: Big Data? Meet the Pandemic...

Heart pounding, edge of your seat fun. "Gone in 60 Seconds" is a cult film so bad it transcends the genre into greatness. In one particular scene, our protagonist who is a legendary car thief ("booster") steals his ultimate prize and nemesis, a 1967 Shelby Mustang GT500 name code named Eleanor. In the ensuing chase by the entire LAPD, our hero, "Memphis" (played by Nicholas Cage) performs some amazing stunts, including driving forward at high speeds, backwards with the car in reverse, using nothing more than a rear-view mirror and a few over-the-shoulder glances to out-maneuver law enforcement who drive with the same aplomb with which Storm Troopers shoot, which is to say...they suck.

If you don't really understand one of the primary promises of the #DataEra, #attributionmodeling, let me simplify. The desired outcome is a model that tells advertisers which channels and what channel mix will yield optimal #businessoutcomes. It essentially answers the questions that has plagued media planners for ages: "Did it work?" "Will it work"? "How much should I spend?". The promise of attribution modeling is it flips advertising budgets from an expense into an investment; it allows agencies to turn budgets into ATMs and tell clients how much revenue they will garner for each dollar invested. Nirvana.

Here's the problem: models are built with one essential assumption: the way consumers will behave will be largely consistent with how they have behaved. Models are built by matching about two years of media spend to two years of consumer behavior; foot traffic, bookings, sales...revenue. Models of varying complexity will account for migrations between media channels, audience shifts be they demographic or geographic, competitive activity and new contenders among others. But the bottom line is still this basic assumption that people will largely behave for the next year how they have for the previous 2 years factoring in a few variables and coefficients.

Enter the #Covid-19 #Pandemic, which has obviously completely changed consumer behavior, and likely will for a long time if not forever. pre-vaccine, people will continue to largely shelter at home (at least the smartest ones will) and make minimal visits to retail establishments, curtail travel (which means fewer cars sold, less gas sold, fewer plane and cruise tickets sold). They will spend more time at home using video platforms (OTT, gaming and even linear TV and radio). They will spend more time online getting information and living vicariously through others via Google and Facebook. No matter how you slice it, we have seen seismic shifts in consumer behavior, essentially negating the value of two years of data. Well, not entirely; it's likely that the same media channels (except maybe OOH) will have the same ABILITY to elicit action, but messaging and frequency will have to change dramatically.

Much in the same way we all behave differently in airports and on airplanes after ONE day 19 years ago, it's not likely people will return to "normal" any time soon after a vaccine is discovered. People will realize it's just easier to get more things delivered; it's easier to buy a car online; it feels crazy to navigate a crowded mall during flu season.

Recognizing this, my consultancy, Pro.Vocation LLC, is working with clients to infuse the old with the old for something new. This means taking the really old, the original first party research for real time sentiment (sources, methods, confidential...give me a call), and integrating it with historical data; the goal is to use the stability of data as it relates to channel effectiveness and "persuasiveness" and re-calibrate frequency and messaging in the age of Covid to build predictive models to help organizations set expectations and revenue projections in way at least slightly more effective that a finger in the wind or best guess.

Friends, you're no Memphis. You can't escape capture forever by driving at high speed in reverse using the rear view mirror. We're all just average to daring drivers who need two hands on the wheel, a clear windshield and high visibility just to follow someone to a lake house when Google Maps hasn't mapped the rural routes.

If you're just looking back, you're living in a past that no longer and never will exist. You need to look forward. It's uncomfortable and itself feels backward, like media planning pre-internet. But that's where we are. The experts understand that and will invent ways to marry quant and qual to create the right kinds of models.

"Donny..."Lowrider.""