<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=2006193252832260&amp;ev=PageView&amp;noscript=1">

4 Min Read

Why Marketers Need to Understand Predictive vs. Prescriptive in the Age of AI

Featured Image

Wondering how to get started with AI? Take our on-demand Piloting AI for Marketers Series.

Learn More

Marketers need to prepare themselves for artificial intelligence. These technologies make machines smart, and they're going to change marketing forever. While the size of that change is uncertain, most commentators agree it'll be big.

McKinsey says tech companies spent between $20 to $30 billion on AI in 2016 across industries. There has been a 4.6X rise in deals to AI startups from 2012 to 2016, according to CB Insights. Six marketing companies with more than $500 billion in funding made CB Insights' "most promising" AI list.

There's a reason companies are bullish on AI: AI tools make firms more productive. AI could boost rates of profitability by 38% on average says Accenture

We've seen this firsthand, having interviewed dozens of AI-powered companies through the Marketing AI Institute. And at our agency, PR 20/20, we've tested a series of AI tools and used natural language generation, a type of AI, to scale content creation and reduce Google Analytics reporting time by 80%.

This is only the beginning though. To understand the current and future potential of AI, let’s consider two types of solutions: predictive and prescriptive systems.

Predictive vs. Prescriptive Analytics: What Are They? What’s the Difference?

You don't need to learn every in and out of predictive and prescriptive systems. But you should understand the basic differences between them. Because those differences could affect your job.

Right now, many tools with the label "AI" only automate common marketing tasks. They might create and schedule social shares, write reports, or create data-driven content. AI is a broad term that includes a lot of different technologies at different levels of maturity. Many systems may use simple forms of AI, but they don't learn without human intervention.

That’s where predictive and prescriptive artificial intelligence systems come in.

A predictive AI system looks at your data and helps you predict a future outcome. One example, says data scientist Michael Lu in InformationWeek, is sentiment analysis.

Sentiment analysis tells you if text conveys positive or negative emotions. A marketer might use it to analyze how consumers perceive a brand or product. The system rates text with a score (good, bad, or inbetween). This score is "predicting the data we don't have" says Lu.

This predictive system tells me what sentiment a text might have. I take this information and use it to improve my marketing. How I do that is up to me, reliant on my past experience and human creativity.

I could look at messaging examples used by other brands to improve sentiment scores. Or, I could research best practices used by other companies to achieve that goal. I could even ask an expert for advice or use my own expertise about what's worked in the past. My next actions depend on the collection of experiences and resources I have at my disposal, and how I use them.

But it doesn't have to be this way.

There is a second type of system, a prescriptive AI system, that will recommend next actions.

A prescriptive system analyzes my available data and the outcomes of particular actions. Then, it recommends next actions. A prescriptive system doesn't know 100% what will work. But it can tell you which actions are most likely to work based on the information at hand.

A predictive sentiment tool will tell me what sentiment score a bit of text might have based on past data. A prescriptive sentiment tool will tell me what to do to improve that sentiment score.

Why Does This Matter to Marketers?

So, why does this matter to marketers?

Because predictive and prescriptive systems can dramatically enhance marketers’ knowledge and capabilities.

What marketers do to produce value for firms breaks down into a few main categories:

  • Assessment — Marketers analyze data, past performance, and best practices to learn what works. They communicate this to stakeholders and colleagues.
  • Recommendation — They use human creativity to recommend new actions that may be successful. These recommendations rely on data from the assessment phase. But they also include healthy doses of intuition, guesswork, and bias.
  • Implementation — Marketers create assets and execute campaigns. They may do this with or without the help of machine systems like automation software.

Right now, predictive and prescriptive systems exist that do all three. One simple example is Phrasee, an AI-powered email tool.

Phrasee assesses email subject line performance based on open rates. It recommends new subject lines that will be successful based on this data. And it implements subject lines, automatically writing them.

The tool beats human performance 98% of the time.

Before you succumb to doom and gloom, let's take a look at a few important contextual facts.

1. Right now, AI tools that assess, recommend, and implement are rare.

Phrasee is a rare example of a tool that does all three. And it does all three in a very limited area of expertise: email subject lines.

Marketers don't risk being fully automated overnight. But they must be aware of AI's predictive and prescriptive capabilities.

2. Humans are necessary to the process.

We were quite broad when outlining marketers' duties (assessment, recommendation, and implementation). That's because you need to realize how fast AI is encroaching on high-level human tasks.

But human marketers are not going away. AI will automate some roles. Yet in many other cases, AI will augment and enhance marketing jobs.

You may find yourself coordinating AI systems, rather than creating outputs yourself. You might take insights from different AIs, and use them in creative ways. Or, you may lean on AI recommendations to fuel high-level human strategic work.

You may even find yourself in higher demand than before. Firms will covet marketers who can coordinate, leverage, and improve AI systems.

3. Competing providers, investments, and capabilities make for a fragmented market—and opportunities.

There are many companies building AI solutions with varying degrees of capabilities. Many are in early funding stages and still working on functional products. This means finding and implementing a tool isn't always easy.

Forward-thinking marketers will lead the charge to AI adoption, not suffer from it. In the process, AI champions will become indispensable to drive and manage change.

We see similarities to digital transformation here. Marketers led the charge to train and transform organizations to leverage digital technology. Many organizations benefitted from the internet thanks to forward-thinking marketers. This trend culminated in CMOs being predicted to spend more on tech than CIOs.

AI transformation will unfold in a similar fashion, though much faster.

Early digital transformation champions reaped huge performance rewards and made their careers. Marketers who are AI champions have the same opportunity ahead of them. They can leverage both predictive and prescriptive systems to become more valuable in the age of AI.

Photo Credit: Kyle Popineau

Related Posts

How to Profit from Your Brand's Predictive Analytics Goldmine

Mike Kaput | July 21, 2021

In an exclusive interview, an AI expert shares what brands need to know to get massive value out of predictive analytics using AI.

How Data Enrichment Improves Predictive Modeling

Laurie Hood | March 30, 2021

Before you go all in on artificial intelligence, it’s important to understand the foundations of successful predictive models. Here's how.

AI for Analytics: What You Need to Know

Mike Kaput | September 20, 2021

AI is transforming predictive analytics by enabling predictive analytics marketing. Here's how to take advantage of this in your business.