Humans naturally crave personal connections. With much of our time spent online — especially over the past year — this can pose a challenge to advertisers. How can you interact with digital audiences at scale, while also ensuring that your messages are personal and meaningful?
Anytime, a Greek insurance brand, made investing in conversational technology one of the pillars of their marketing strategy even before the pandemic hit. And when the country went into lockdown last year, they ramped up the efforts.
They partnered with conversational marketing automation platform Spectrm to design a conversational assistant powered by artificial intelligence (AI). It engages consumers with personalised one-to-one conversations in display ads using AdLingo. Built in Area 120, Google's in-house incubator for experimental products, AdLingo integrates the conversational assistant into ads on desktop, mobile web and mobile apps via Google DV360. With this solution, Anytime can ‘talk’ with consumers about insurance quotes and, ultimately, provide them with a personalised offer. The chatbot is shown in a display ad format — without sending consumers to a brand’s website — so they interact with the bot wherever they are browsing online. That way, the interaction feels more natural and intuitive.
Tapping into conversational trends
When engaging with consumers online, generic and predetermined messages from a chatbot often leave much to be desired. That's why delivering a more meaningful customer experience requires personalisation, especially with non-human interactions. For Anytime that meant gathering insights into how their customers were interacting with different creatives and messaging.
Users are eager to engage with brands when they can provide natural and intuitive experiences that deliver value
The company first rolled out campaigns focusing on car insurance. Within the first 45 days of launching the chatbot, consumers sent more than 33K messages to the bot.
However, after looking at the conversations, the brand saw more interest coming from motorbike users who asked the bot questions such as "Where can I find motorbike insurance?" and "How much is motorbike insurance?"
Responding to this insight, they reworked creatives and adjusted the copy. In the updated version, the chatbot took users through a series of relevant questions that led to a personalised quote on motorbike insurance.
“That is one of the things that surprised us the most,” explains Ria Mylona, digital performance marketer at Anytime. “People were eager to interact with the bot more than it was required, by typing different opinions or questions. That gave us a unique window into their needs and allowed us to introduce new functionalities fast.”
While brands can get started with conversational marketing using simple decision-tree chatbots, these don't always offer users the personalized experience they're looking for. To really connect with consumers in a more human and conversational way, you need a type of AI that captures specific intent and can be trained to make more accurate predictions and match appropriate responses.
“Users are eager to engage with brands when they can provide natural and intuitive experiences that deliver value,” says Blake Hammer, business development lead at Google.
Understanding behaviour based on device usage
Not only could Anytime detect and pivot quickly to respond to interest in different products, but the solution also allowed them to determine which device — mobile or desktop — performed better in different scenarios. For example, they saw that conversation length lined up with device usage.
Since mobile conversations tended to be shorter — typically around one minute as opposed to three minutes on desktop — they needed to transition users to their website to receive a quote earlier.
“The users are tagged in Spectrm based on the ad entrypoint, so advertisers can easily track performance at a device level and gain insights on their preferences and behaviour” adds Benjamin Gibert, head of marketing at Spectrm.
Increasing customer engagements
By implementing the chatbot solution, Anytime increased engagements with their customers. Conversations averaged three-to-four minutes, exceeding the company's benchmark of 1 minute and 17 seconds. And the brand had more than 100K conversations with a 0.53% engagement rate.
“The results mean that users are curious about AI conversations as long as they are natural, responsive and don't require too much information to provide in advance,” concludes Mylona.
Anytime is going to continue to use its conversational assistant, even when customers spend less time at-home again in the future. They will use it to better understand their audience, including how their customers prefer to interact with them and which messages help drive the most conversions.