Conversational marketing: where and how to deploy chatbots and AI

21st October 2019

Do you remember the old days before we all carried in our pockets a computer powerful enough to put a man on the Moon? Don’t worry, I’m not going to dwell on the good old days. Back then we didn’t have instant access to all our family and friends. We couldn’t check which movies were at the movie theatre and then summon all the navigational info required to get there. We are now used to have an answer to all our questions the moment we ask them. We are now always engaged in a conversation.

Conversational marketing is a strategy to engage conversationally with our customers across multiple channels, when, how and where they want to. It’s a strategy that crosses lines across business functions and informs marketing, sales and customer experience strategies.

Conversational marketing is focused on the client, meets them where they are, is aware of their context and answers them when they need it. According to research by Drift, 86% of people would rather use a chatbot than fill out a form

The advantage of robots over humans

Conversational marketing is not about using a specific tool. It’s a strategy that can be deployed across different technology to engage in one-to-one conversations with clients. From email and SMS to live chat on websites, Facebook Messenger and Alexa.

But when engaging in large scale conversations and chats, bots and AI have distinct advantages over live chat with a human representative: scalability, availability, speed and context.

  1. A chatbot or AI solution can handle multiple conversations at once and can be scaled according to need instantly. Instead of recruiting and training new support staff to handle an increase in live chat demand, you can just increase the resources or use a third party service that can handle the load.
  2. A chatbot can be ready to answer a customer query 24 hours a day, 7 days a week. It’s always ready just when the client needs it.
  3. A chatbot doesn’t need to review documentation before providing an answer. If it understands the query, then the answer comes instantly, fulfilling the expectation of immediacy from the client.
  4. The chatbot, if it’s model with machine learning capabilities, can understand the context of the query and provide options that advance the needs of the user. So if a user is asking for upgrade options on a hotel reservation, the chatbot can offer the upgrade directly and process it, amending the original reservation.

As of now, live-chat with a human is better for adding empathy and emotion to the interaction and for handling very complex queries.

Etsy uses a chatbot in Twitter to handle support requests over direct messages.

Rule-based and AI chatbots

Not all chatbots are made equal. Roughly, there are two main types of chatbots:

  1. Rule-based chatbots. This is the most common type of chatbot and the easiest to implement. You have likely interacted with one of these chatbots at some point. In this type of chatbots, all the possible conversation paths are mapped and written in advance, and the user chooses from different options to advance the conversation. They live on websites, but also on social media sites like Facebook or Twitter. There are more than 300.000 chatbots on Facebook Messenger alone.
  2. AI chatbots. These chatbots don’t rely on a mapped conversation with pre-defined options and can accept complex queries written in conversational language, which is more human and allows for increased personalization. They use large datasets of human interaction and previous conversation to learn and understand the user’s requests. There are two ways to include AI in a chatbot.
    1. The first is through natural language processing. The bot processes the query to understand the user’s context and intent.
    2. The second is through machine learning. In this case, the bot can learn from conversations to anticipate the user’s needs instead of reacting to them.

What’s the business impact of using a chatbot to manage customer interaction?

L’Oreal used a chatbot to guide users through gift recommendations.

The biggest impact of deploying a conversational marketing approach to interact with existing and future customers is the increase in customer engagement. Instead of relying on forms, you can offer them a more personal experience that continues to move them down through your conversion funnel. L’Oreal achieved this through personalized product recommendations and re-engagement strategies.

Chatbots are also really useful at qualifying leads and acquire customer information. You can build your bot to ask the same questions you’d ask at a lead generation form. But in this case, the conversation can take different paths according to the answers, providing next steps that are aligned with the needs of the user. For some, it’d be suggesting the download of an ebook or registering for a webinar. For others, it’d be scheduling a call with a sales rep. Aprimo was able to cut their sales cycle from 8 months to just 45 days after implementing a conversational marketing strategy

The improvements in customer engagement and data collection, together with the availability of bots mean that you can shorten your sales cycle, generate more leads and increase your sales pipeline. A company implementing chatbots for this purpose saw a 45% increase in conversion rate and a 10% increase in pipeline velocity.

The overall experience feels more human. I know how ironic that sounds, considering that we’re talking about chatbots. But the flow of the interaction is more closely aligned with how we carry on our interactions with other people daily.

Best practices for a conversational marketing strategy

Chatbots are great. But they can’t handle 100% of the queries sent by your users. That’s why it’s important to have a fallback to human interaction when there’s an error or the bot can’t handle the interaction anymore. At those points, the chatbot can offer the user the option to switch the conversation to a human.

Always be transparent with your users. They need to be aware that a chatbot is at the other end of the conversation. That will help them set their expectations for the interaction. But that doesn’t mean you can’t make the bot a bit more human by giving it a name, a face and a personality.

Your user has an objective in mind when interacting with your chatbot. It’s not there to have a cute conversation. So keep your messages short and aligned to your brand voice and tone.

Showing empathy to your users’ issues is key to set a good tone for the conversation from the start. Make your users feel valued and heard.Personalize the conversation with the information you already have. Don’t ask which subscription plan they have if you can directly offer an upsell package. If you can read the context of the conversation and get ahead of your users’ issues, you’ll increase their perception of you as aligned with their needs.

And always measure and optimize. You already know that, because you’re awesome, but it’s important to put metrics in place to measure the impact of the conversations for your business. Learning from feedback and past conversations will help you improve the experience and bring delight to your users.

L’Oreal increased engagement and conversions through personalized recommendations via a chatbot.

Steps to deploy a conversational strategy in your business

Identify the opportunities for conversational marketing

The first question you need to ask is where a conversational marketing strategy can have a bigger impact. You may want to improve the experience of your users when they need help, or you may need to accelerate the customer journey down your funnel.

Understand your users and their needs

If you want to engage your users in a conversation, you need to understand what language they use, what are they trying to achieve and how you can meet their expectations.

Devices, platforms and channels

Based on the understanding of your audience and your goals, you can start thinking about devices, platforms and channels. Are you trying to improve your customer support? Do your users usually reach out from a mobile phone? Do you have already an audience in social media? Maybe a Twitter bot, like the one Etsy implemented, would do the trick. Do you want to increase conversions on your site? Maybe an AI chatbot can increase your funnel velocity by guiding people to the content they need.

Content ideation for chatbots

Develop your scripts by running roleplay conversational sessions in an ideation workshop, as if you were having a direct conversation with your users. This will also help you ground your script in plain English, and capture unexpected conversational paths.

You should also use this step in the process to define the tone and voice of your brand when engaging in a conversation. If you’re looking for inspiration, Intuit, Shopify or IBM have useful design guidelines in their design systems.

Analytics and optimisation

You should monitor key metrics and analyze the conversations users have with your chatbots to discover optimization opportunities. There are four key metrics you should monitor:

  • Chatbot users
  • Session length or number of conversational steps
  • User ratings
  • Number of fallback responses

You should use these metrics should in combination to avoid misinterpreting data. For example, a short session length could mean a high success rate in solving user queries, or that users get disappointed and drop the conversation too soon.

Start engaging your users in conversations now

In a world driven by conversations and communication, it makes sense for businesses to adapt and provide users with an experience in a format they are already comfortable with. The results of focusing on the client’s needs and providing a more human experience with your business will be an increase in customer engagement and satisfaction, increase the knowledge of your users and optimize your sales funnel.

What are you waiting for? Start the conversation now.