Stay marketing-savvy and tech-savvy. Get the latest in martech by subscribing to MarTech Today.
Why B2B needs artificial intelligence
It’s not a coincidence that artificial intelligence is rapidly colonizing marketing tools designed for business sales. It’s a necessity.
Artificial intelligence (AI) is more than a stylish trend. It goes beyond rules, providing the ability to understand content or language, find patterns that can be applied to the future, digest all kinds of information and make reasoned decisions.
One by one, B2B vendors are rolling out their AI chops — targeting platform Demandbase, CRM and marketing platform Salesforce, account engagement platform YesPath, conversational platform Conversica, and B2B predictive marketer CaliberMind, among a growing list of others.
To get some insight into what this means for businesses selling to businesses, we talked with Raviv Turner, CEO and co-founder of CaliberMind. (Turner will be co-presenting “A Scientific Look at B2B Buying in the Age of AI” at our MarTech Conference next month in San Francisco.)
At the top level, he said, AI helps to solve key challenges that are particular to B2B.
First of all, selling to a business is complicated.
Turner noted that previous research by Gartner found there is an average of 5.4 people on corporate buying teams, a figure that is now estimated to be 6.8 people. Unlike consumer selling, in which buying decisions are usually made by one or two individuals, business buying decisions require an extraordinarily broad and informed consensus.
“AI helps you determine the buying team,” Turner pointed out, through platforms that map organizational charts, sift clues to determine the team members for given purchases, analyze and segment their profiles and predict what kinds of offers might appeal to them.
More than intent signals
Then, there’s the B2B condition that sales cycles are long — and getting longer.
Turner noted that research firm Sirius Decisions recently found that sales cycles are 25 percent longer than they were just a couple of years ago, with a deal in the $50-100,000 range taking an average of seven to eight months from start to finish.
This means that a seller had best match marketing and sales to the customer journey, providing product information when information is needed, competitive comparisons when required, offers when that stage is reached and revised offers when appropriate.
Such tracking of a customer journey, he argued, requires more than, say, intent signals such as supplied by a Bombora, which knows when people from Company X have started looking at pages and downloading white papers about a product like servers.
Turner argued that intent signals are not fully predictive, since they only tell that someone — possibly identified, possibly not — from company X is in the market for servers over a given period of time.
That’s only one step in the corporate buyer’s journey, Turner said. It indicates interest, but not exactly what stage the process is in. It also doesn’t account for what he described as the three different models of B2B purchase decision-making: centralized as in a headquarters-centric firm, decentralized by business units, or federated, which includes both.
And decentralized models take even longer than average sales cycles, in part because the decision hierarchy is often less well-established.
A branch office may have to get buy-in from a variety of users just to purchase a new office printer, for instance, which can mean more decision-makers than average and more time. And an average B2B customer journey over a lengthy sales cycle takes as many as 17 inbound and outbound touch points, according to Turner.
Conducting the orchestra
Unfortunately, AI can’t change decision structures, the number of team members or the sales cycle.
But it can keep track of data across silos, so all the clues, profile attributes, intent signals, site visits, email exchanges, white paper downloads, major moves by that account, coordination between marketing and sales, and dozens if not hundreds of other factors can be collected and analyzed into a predictive score.
Data silos are a bigger problem in B2B selling than B2C, Turner said, because there are more decision-makers involved and because B2B is not transactional. That is, it’s not a consumer plopping down a credit card to buy a new coat.
It’s a bid request, a bid, an offer, a counteroffer, an authorization letter, a purchase order and sometimes an electronic payment. In other words, even the payment process in B2B is complex, with relevant data often living in different places.
Key to orchestrating all that diverse data is that AI is able to handle unstructured data, the non-format where most business communication lives — emails, natural language processing and analysis of phone calls, social posts and more.
But there’s so much of it. Making sense of big data is now becoming a requirement for up-to-date B2B. The abundance of AI layers in tools and platforms is making this kind of data wrangling feasible for a wider range of companies, because it means you don’t have to hire a crew of data scientists to get on board.
And then there’s return on investment. As the cycles get longer, the teams larger, the decision processes more complex and the relevant data more abundant and various, AI-infused tools are needed to figure out what spending resulted in that sales lift.
Businesses sold to businesses long before anyone knew what AI was. But now the competitive edge is riding on rivers of data, and AI has become the vehicle for staying afloat.