Are We Entering the Era of Millisecond Marketing?

Opinion: Without the right data, even the best AI engines sputter and stall

We’ve entered the era of millisecond marketing
Marilyn Nieves/iStock

Nowadays, it’s a scene you’ll pretty much see only in movies: Stock traders in brightly colored jackets frantically shouting and flashing symbols on exchange floors. But this was how stocks were bought and sold for most of Wall Street’s history.

Then, in the 1990s, electronic exchanges entered the picture. In short order, powerful PCs were making trades in milliseconds, leaving even the fastest floor traders in the dust.

These days, high-frequency trades are powered by increasingly complex algorithms that analyze multiple markets, and then transact vast numbers of orders at blinding speeds. The decision to buy and sell is largely out of human hands. High-frequency trading is not without its critics, but the lucrative payoff has spurred a tech arms race among the industry’s largest players.

What does all of this mean for the world of marketing? Quite a lot, it turns out.

Marketing’s technology problem

In contrast to the stock market, technology has come slowly and fitfully to marketing. Only in the past decade or so have digital tools to streamline how we promote and sell products become widespread.

Cloud-based customer-relationship-management platforms simplified the process of compiling and tracking prospects and leads. Advanced content-creation software made it easier to churn out designs and graphics. Marketing automation tools let businesses scale campaigns and reach bigger audiences.

But by and large, we were still in the “Don Draper” era of marketing. Marketing teams, helped by creative agencies, put their heads together to come up with catchy campaigns. Data and research might play a part but, just as in Mad Men, these efforts rested on a decidedly human process informed by experience, gut and intuition. Art would be generated by a design team. Copy would be scrutinized and finalized by editors. Ads would be placed, blog posts written and email campaigns launched.

Then, we’d wait—fingers crossed. Some campaigns crashed and burned. Others exceeded expectations. But even in an era of Google Analytics, it was hard to predict what would work and what wouldn’t.

In some ways, we hadn’t advanced all that far from the days of marketing pioneer John Wanamaker, who back in the late 1800s famously quipped, “Half the money I spend on advertising is wasted; the trouble is, I don’t know which half.”

Rise of the machines

But since 2011, the handful of tools in the marketing technology marketplace has ballooned to more than 3,800. And they’ve evolved from simply automating repetitive tasks to generating their own competitive insights. Marketing clouds—from the likes of Adobe, IBM, SAP, Oracle and Microsoft—bundle the best of these tools into integrated packages.

At the heart of these clouds are increasingly sophisticated artificial intelligence brains powered by, you guessed it, algorithms—from IBM’s Watson to Salesforce’s Einstein. They process vast amounts of raw data in order to better target customers, place and tweak ads, personalize campaigns and measure results—and they do it automatically.

Right in front of us, the cycle for creating, distributing and optimizing marketing content is being radically compressed. Instead of relying on human intuition, it’s fueled by in-depth consumer data gathered from social media and other sources. Ever more complex algorithms are calling the shots and removing the guesswork about what marketing materials will lead to what results—and they’re doing so nearly in real time.

Suddenly, we’re no longer old-fashioned stock traders sweating in the pit, gesturing frantically and hoping for the best. We’ve entered the era of millisecond marketing.

Millisecond marketing in action

The ad space offers some of the clearest illustration of the potential here. Not long ago, digital ads were largely the domain of specialized ad agencies, which had the design know-how to create ads and the industry connections needed to place them.