AI Fast Track: Study Shows How Marketers Are Adapting

• Updated on: 2017-10-06
• Read original article here

There is so much hype about artificial intelligence (AI). How does one distinguish between fact and fiction?

Amplero, a Seattle-based, artificial intelligence marketing technology provider, decided to find out. It commissioned Forrester to conduct a study on attitudes about the use of AI in marketing. Forrester polled 150 North American technology decision makers working at B2C companies.

Some of the findings are upbeat. For example, 78% will adopt or expand AI platforms in 2018.

In addition, 86% believe that AI can provide value to their businesses. And 92% are upping their technology spend in the next one to two years, with AI as part of the package.

As for email, 39% plan to allow AI to determine which content is served to customers, and 30% will allow AI to generate content, according to the team at Amplero.

“While email marketing innovators are no stranger to utilizing vendor algorithms and rules to automatically determine send times or run multivariate tests, core AIM platforms are enabling large brands such as Sprint and DoubleDown Interactive to draw on the entire customer data ecosystem to run thousands of simultaneous experiments to continuously determine the right experience or message for each customer,” says Olly Downs, CEO and chief scientist at Amplero.

Downs adds: “With the implementation of AI at the core of the marketing technology stack, enterprise marketers are finally able to move beyond rules-based systems and manual segmentation processes to deliver highly contextual, 1:1 customer experiences at scale that impact crucial business KPIs.” 

He continues: “As a high-performance, customer engagement channel for most large B2C organizations, email is proving to be an optimal use case for AI marketing technologies, often resulting in significant lift on not only short-term engagement metrics like opens and conversions — but also long-term revenue and retention KPIs.”

That said, there are limits to the general enthusiasm for AI. Over a third of the respondents doubt that “AI in marketing is currently real,” and almost eight out of ten believe AI requires human intervention to guide it.

What’s more, almost half wonder whether AI will lead to job losses, matching a survey conducted in the UK by The Drum and Sysomos.

Finally — and this says a lot — companies that have not implemented AI are more positive about its capabilities than those that have.

However, firms do have lofty goals when implementing AI. The top five priorities, all cited by 86% or 87%, are:

Their top KPIs? The big three are:

Meanwhile, lifetime value is cited as a KPI by 57%. 

Still, there are challenges, such as data. Of the respondents, 63% believe they have too much data to get actionable insights for their campaigns. Yet 80% plan to increase their data use within the next 12 months.

This is critical: A company needs “an easily accessible, consolidated view of the customer in one database, and if you don’t have that, AI and machine learning are not going to impact your marketing,” says Jeff Hassemer, chief strategy officer for Alterian.

He adds: “AI doesn’t know the difference between good data and bad data.”

And attitudes are not all positive. Here’s what the decision makers think:

Now let’s follow Forrester’s lead, and use some Marxist terminology — i.e., the haves and the have-nots. In this context, the haves are firms that have implemented AI, and the have-nots are those who have yet to do so.

It turns out the have-nots have a better view of AI than the haves.

For example: 64% of the have-nots believe it can help them identify their target audience. Only 37% think that.

Similarly, 63% of the have-nots are convinced that AI can help drive marketing personalization. Only 44% of the presumably sadder and wiser haves agree.

Here are a few more disparities:

Can you even define AI? Many techies see it in “a more science-fiction light, such as ‘intelligence exhibited by machines’ or ‘simulation of human intelligence by machines’; others simply refer to is as machine learning or predictive analytics,” the study states.

You probably have to nail that down before doing anything.