The State of Conversational Marketing in 2018

Two contradictory things happened to conversational marketing in 2018 at the same time. Real-time chat became something serious B2B companies budgeted for as a core part of the website, not a gimmick tucked in the corner. And a wave of “chatbots are dead” pieces started circulating, written by people burned by exactly the kind of bot that made 2018 chat feel serious in the first place. Both were true, and the difference between them mattered more than either headline on its own.

Chat widgets went mainstream

By 2018, a website chat widget in the corner of the screen had stopped looking unusual on B2B software sites. Drift, which had spent the previous two years pushing the term “conversational marketing” into the industry vocabulary, along with newer entrants and established players like Intercom expanding into the same space, made real-time chat a normal line item in a marketing stack rather than a curiosity. The pitch was straightforward: a static contact form makes a visitor wait for a follow-up email that might come tomorrow; a chat widget can qualify and route that same visitor while their interest is still fresh.

For companies selling to businesses with a real sales process, that pitch landed. Response speed had already been shown to matter enormously in B2B sales, and chat offered a way to compress the gap between a visitor’s question and a human answer from days down to seconds.

Messenger bots hit their ceiling

At the same time, the platform most associated with “chatbots” in the public imagination, Facebook Messenger, was showing real strain. Facebook had reported well over a hundred thousand active bots on the Messenger platform by the middle of the year, built by brands eager to be part of what looked like the next big channel after the 2016 bot platform launch. Many of those bots were rigid, scripted and easily confused by anything outside their narrow decision tree, and users noticed. Coverage through the back half of 2018 turned openly skeptical, with plenty of marketers openly admitting their bot experiments had underdelivered against the hype from a year or two earlier.

The lesson sitting underneath both stories was the same one: a chat interface only works when it’s useful to the person using it. A bot that answers a real question quickly builds trust. A bot that traps someone in a menu tree and never lets them reach a human erodes it, and it erodes it fast.

GDPR forced a rethink of messaging consent

Europe’s GDPR, which took effect in May 2018, added a real complication for anyone building conversational programs around collecting contact details through chat. Consent had to be explicit and specific, not implied by the act of starting a conversation, which meant marketers running chat and Messenger campaigns in or targeting the EU had to rebuild opt-in flows that many had treated casually up to that point.

What distinguished the programs that worked

Looking across the year, the conversational marketing programs that held up shared a few traits: they were scoped narrowly, answering a specific, well-defined set of questions rather than trying to replace a human for everything. They made the handoff to a real person fast and obvious rather than hiding it behind more menus. And they were placed on pages where visitor intent was already high, a pricing page, a demo request, rather than bolted onto every page of the site regardless of context.

Heading into 2019

By the end of 2018, the term “conversational marketing” had moved from a Drift marketing slogan into something most B2B marketing teams at least had an opinion about, which is its own kind of milestone. The open question heading into the next year wasn’t whether chat belonged on a website anymore. It was whether the industry could build bots disciplined enough to avoid repeating the Messenger backlash, this time with the benefit of a year’s worth of public lessons about what happens when a bot promises more than it can deliver.