Chatbots have become increasingly popular among both businesses and consumers, with roughly 56% of customers claiming they would prefer to chat with a bot than call support.
However, a chatbot can do much more than simple support queries, and there is still a major gap between high-performing chatbots and mediocre chatbots.
Some studies predict that bots with mediocre engagement only see a 35-40% response rate while quality bots regularly see 80-90% response rates.
Therefore, if you notice that most of your chats end without a conversion, or you’re simply looking for ways to improve engagement, here are a few proven ways to dramatically improve your chatbot’s success.
As the internet enables companies to become more global, most websites now host a variety of visitors from different countries.
For example, Neil Patel stated that one of the best ways he’s been able to grow his traffic is by translating his content. In fact, while most of his traffic was previously from the United States, American visitors now make up only about 25% of the six million monthly visits he receives.
Therefore, it’s essential to create a multilingual chatbot.
For example, while this appears to be a wonderful chatbot for ordering food, it’s not very helpful to English-speaking visitors.
Unfortunately, most chatbots only exist in one language, as translating them can be rather advanced.
First, some words are the same in multiple languages. For example, “no” is the same in English and Spanish, so if someone responds to a question with “no,” the bot won’t be able to detect if they should serve the English or Spanish version.
In addition, translators often don’t account for linguistic nuances, so some of your conversations likely won’t make sense if you simply translate the conversation.
To solve this, you can use a Natural Language Understanding (NLU) tool like Dialogflow to create dialogue in English and then adapt it to various other languages.
Microsoft also has an NLU solution, LUIS, which can accurately translate up to 11 languages with various linguistic nuances.
One of the best ways to leverage chatbots to increase conversions is retargeting. Similar to retargeting ads, a chatbot can recognize returning visitors and even detect their interests.
For example, if you have a visitor click on a pricing page for agency services, you can retarget them with a discount or offer a consultation sign-up.
Here’s an example of what the dialog flow would look like:
So how do you create a retargeting chatbot?
Retargeting chatbots cookie visitors that interact with your chatbot so that you can see what pages they view and how often they view them.
Therefore, you can schedule a follow-up message to be sent a few hours after the visitors first interaction with your brand. The set up process depends largely on the chatbot platform you use, but the end result is more or less the same.
Triggers for certain pages
The chatbot on your website’s homepage is likely somewhat generic and probably sends visitors to relevant pages or connects them with a live agent.
While this is an effective way to separate masses of traffic (such as on your homepage), it isn’t a great experience for more segmented pages (such as a specific service page or your pricing page).
Therefore, you can use triggers to create different conversations for different pages.
Here’s an example of a somewhat generic catch-all bot for a homepage, and it works well for sorting a wide range of customers.
However, the pricing page has a more specific bot that addresses pricing questions.
Similarly, they have a specific bot for pages that target agencies.
While the process of creating triggers is slightly different for each chat platform, it usually involves tagging the desired pages and then setting up a trigger to show that sequence when a visitor lands on that page.
One of the best ways to create a better experience (and ultimately improve engagement) is by solving a user’s key pain points within the chatbot itself.
For example, if a user booked a flight ticket and wants to know the flight status, the chatbot can provide a visual flight status.
This way, the visitor never has to leave the chat to accomplish what they need to find.
While showing a visual status card is a nice element to enhance the user experience, other rich elements, such as restaurant reservations and scheduling calendars, can lead to a direct increase in conversions.
To do this, you can integrate apps that you currently use into your chatbot. For example, here’s an example of a Calendly integration.
Enabling visitors to schedule a consultation directly in your chatbot is ideal for improving conversions.
Therefore, re-evaluate your current chatbot.
Are there any pages that you’re currently sending chatbot users towards to make the final conversion?
From there, consider how you can integrate your apps with your chatbot to enable visitors to convert within your chatbot.
As with virtually any marketing channel, the key to maximum optimization is A/B testing each campaign.
This also applies to chatbots.
For example, you may want to consider testing your offer, the copy language, images, and other elements of your sequence.
Specifically, here are a few things you can test:
- How far into the conversation you ask for an email
- How early you sell them on your product or service
- How many different options should be clickable on the homepage bot
- How much information you can ask from a person at each stage (company size, revenue estimate, etc.)
- Various calls-to-action (CTA) such as Click for Pricing vs. Learn More About Us
- How much information you provide (a sentence or two versus a paragraph)
- Different sentiments (using emojis or keeping it professional)
While A/B testing campaigns isn’t a new tactic, it’s often underleveraged and opens the door for a new level of engagement. Therefore, consider testing at least one element of your chat sequence weekly.
To set up an A/B test, you can do so in the dashboard of virtually any chatbot platform like ManyChat or Chatfuel.
For example, a platform like Chatfuel comes with an A/B testing plugin where you can decide how many users see each variant and which message they see.
Unfortunately, outside of the individual platforms you use, there aren’t many A/B testing tools that currently exist for chatbots. Therefore, if you want more advanced tracking, you’ll have to create your own tool and set up custom tracking and data visualization with graphs or tables.
All of these tactics to maximize conversions are only useful if you are tracking them appropriately.
First, here are a few basic things you should be tracking and why you should track them:
If the chatbot receives little engagement, the problem may be anything from the page it’s placed on (as not all pages need a chatbot) to the opening line of your bot.
Therefore, while it’s important to observe how many visitors engage with your chatbot, it doesn’t necessarily mean that your chatbot is low quality.
Your retention rate is how many visitors return to use your chatbot after one interaction. Ideally, you want a high retention rate as this means visitors were pleased with it enough on the first interaction that they were willing to use it again.
This is the number of visitors that come to your website and use your chatbot but don’t continue engaging with it. Ultimately, if visitors don’t continue to engage with your bot, it probably isn’t very effective and needs to be updated.
Goal Completion Rate
Finally, if you have a CTA in your bot (such as downloading a white paper or submitting an email address), you can track the percentage of visitors that perform this action. If your visitors continuously drop off before completing the goal, it’s a strong indicator that your chatbot or offer isn’t effective.
While these quantitative goals are important, you should also track qualitative feedback like user sentiment. For example, do people thank the bot at the end of the session? If you include a survey at the end of the bot, what feedback do users provide?
While you can track user sentiment manually, it’s much more effective to do so with a tool like Amazon Comprehend. This tool will track conversation data and give you a summary of user sentiment for a full overview of your chatbot’s performance.
Despite the rising popularity of chatbots, most businesses have only scratched the surface of their full capabilities.
As automation technology becomes increasingly advanced, these six tactics will remain staples in chatbot creation and development. Next time you build a bot, consider incorporating at least one of them into your sequence and let us know how it performs!
Guest author: Bojana is a content strategist with a finger on the pulse of the world of Messenger chatbots. She heads up the content strategy for Chatfuel, the leading no-code chatbot platform for Facebook, Instagram, and Messenger, and has overseen its growth from zero to 100,000 plus monthly visits. Follow her on Twitter @mss_bojana.