Over the past decade, B2B marketing has undergone a massive transformation. According to Forrester Research, the entire B2B sector represents more than $1 trillion in digital commerce every year, more than double the size of the B2C economy. But what’s next? Is artificial intelligence the rocket fuel to take us through the next decade of marketing technology?
AI is not about inventing a new task or a new way of being intelligent, but simply mimicking human intelligence. It’s about doing the same old human tasks at super-human scale. Given that it’s a machine doing the work, AI can be done at infinite scale since it can read and process billions of data points with perfect memory.
The abundance and availability of data around marketing is why marketing processes are such a sweet spot for AI. Marketing is also a process that still has very low yields (0.03 percent, from an initial inquiry to closed business) and thus provides ample opportunities for ROI. We’ve already seen everyone jumping into the fray, including Salesforce with Einstein and Microsoft with its new Dynamics AI platform. In addition, Sundar Pinchai, the CEO of Google, recently announced that AI will be the central component in all of their products.
AI is different from just sorting data or predicting leads based on a limited CRM database. It’s a lot closer to Amazon Echo than to a spreadsheet. The next frontier for AI will be on the frontline of the brand communicating directly with the buyers rather than some back-office process. But to do so, it must be able to understand and communicate in the language of buyers and gauge deep insights about them. Given that most of the world’s knowledge is expressed in natural language, it has to be able to understand and communicate in human language and not in scores or numbers.
Hyper-personalized conversations at scale
There are many areas in marketing where current and future AI can be applied, including lead ranking, buyer identification, data cleansing, dynamic account assignment, opportunity forecasting, and the next sales action. But the most interesting and valuable use for AI is the ability for marketers to have a one-on-one personalized conversation with buyers who know their pain points, goals, and ambitions. The value of hyper-personalization comes from its ability to eliminate one of the scourges of marketing: worthless spam. It allows the brand to scale a personalized conversation to millions of buyers as if there were a personal concierge attending to them.
What exactly do I mean by that? Today, strategic account or field marketing managers act as such a concierge; they have in-depth knowledge of the accounts, business landscape, and industry and know how to align their conversation to the buyer’s business priorities. Until recently, such conversations happened only with an exclusive realm of highly paid people. But now AI can allow each of a company’s 10 million website visitors to have a unique conversation with a brand.
We already know this tactic works when conducted by humans. In fact, according to McKinsey, personalization can deliver five to eight times the ROI on marketing spend.
Types of hyper-personalization
Every industry has slightly different methods and channels in which they communicate to their buyers. To make hyper-personalization work and to avoid sounding disjointed, AI has to be applied consistently to all the ways in which a brand communicates to their existing and new customers. Three ways marketers can apply hyper-personalization across the buyer’s journey are:
- Dynamic ad copy: Today, we don’t think of advertising as part of a conversation, because it’s stuck on a billboard or a website and doesn’t apply to 99 percent of the people viewing the ad. What if the ad copy could actually change for every buyer and account? We know that would be more effective. What if we knew that the ad impression was going to a female CMO of an auto parts manufacturer in Detroit, with a new partnership with Mercedes, targeting Tesla as an account, with budget for a new digital marketing system? AI could develop personalized ad copy tailored to this CMO — “Marissa, download a case study of how Mercedes is using our marketing engine to transform their digital experience.”
- 1-to-1 emails: This is the most exciting opportunity for hyper-personalization at scale, since emails still remain the primary communication for deals. While generic messages don’t really work, researching buyer interest to create personalized messaging does. If a human does this, it is inefficient and rarely effective. With AI, you can understand buyers’ interests at scale and craft highly personalized emails to them.
- Unique website experience: The same advertising conversation has to continue to a brand’s website. Today, about 50 percent of the people bounce, and 97 percent of the people don’t really find what they’re looking for. Imagine a Netflix-style personalization engine where richer content is weaved from individual pieces and net new web content is generated with suggestions for what a prospect would like based on past viewing experiences. What if the same CMO of the car parts manufacturer clicked through the personalized ad and we knew that she likes her content in video format? Would you want her to be on a generic website or just show her the video case studies you have for the auto industry?
How can marketing organizations make this happen today?
In some not-too-distant future, each hyper-personalized conversation across ads, the web, and email will be automatically generated from scratch. But that’s not possible today. Natural language generation is still a complex developing field. Thankfully we can hack around this problem by weaving in pieces of existing content.
There’s nothing wrong with that, and in reality, that’s what your smart sales and marketing people are doing when they’re asked to do something custom. Companies like Netflix and Amazon have become masters of product and content recommendations over the past decade, and there is absolutely no reason why other businesses can’t apply the same techniques based on data on their own websites or third-party identifiers.
By combining a company’s internal clickstream data and simple machine learning, effective personalized experiences can be delivered today so we are not sending our users down a structured maze of industry or product categories.