How Using Humans And Chatbots Together Generated 182% More Qualified Leads

This post is part of Made @ HubSpot, an internal series of thought leaders where we learn from experiments done by our own HubSpotters.

The prospects for companies have never been as high as they are today. according to Google61% of people now expect brands to offer personalized experiences.

At the same time, the number of people who want to reach a company Messaging continues to grow. This means that even well-trained teams can find it difficult to stay in the chat.

At HubSpot, we felt the brunt of this pain in early 2018. We have a dedicated sales team that takes care of chatting with people on our website. At that point, we used most of HubSpot.com to connect with potential customers in a way that felt natural.

Unfortunately, the sales management noticed that we had lost valuable potential prospects who tried to reach us via live chat. We just didn’t have a guaranteed way to ensure that a member of our team met their needs – which meant that we lost a lot of time on potential sales.

When my marketing team at HubSpot contacted the sales team to discuss the issue, we found that live chat alone no longer met our requirements. To communicate appropriately with each prospect and create the most personalized experience our website visitors yearn for, we had to scale the productivity of our sales team with a chatbot.

Here’s why we chose to create a chatbot, how we designed the experience, and why it could be an outstanding solution for your own business in 2019 – and beyond.

Why Live Chat didn’t work for us

When the live chat sales team turned to us, we knew we had to make the team more efficient.

Live chat in itself was not a problem – but unfortunately we did not tailor the discussions to the visitor and his problems. We knew it was right to sift through each visitor’s requests as this was helpful solve for the customer and answer the questions of all potential customers.

However, as they all fell into one bucket, it became increasingly difficult for the sales team to keep up. And for the majority of the people who gave us no way to get back in touch, they were gone forever.

One time over a third of the people who chatted with us never heard something back.

In addition, hundreds of the chats we received each month came from users who only needed product support. This has consumed part of the valuable breadth of our sales team and made it more difficult for them to connect with website visitors who actually had to speak to a sales representative.

Ultimately, this bad experience created unnecessary friction for our website visitors and salespeople.

Why we decided to build a chatbot

After a few internal discussions, we set about creating a chatbot that would appeal to visitors, search them, and get them to the right place earlier. This would be a win-win situation for both our website visitors and the sales team.

To find out what this chatbot should do, we first looked at the live chat transcripts. They are an invaluable resource for conversation marketing because you can hear what the prospect is saying in their own words. There are technical ways to classify chats (like natural language processing), but a qualitative approach is fine to get started.

We then divided the intentions of the live chat into three areas: sales, support and an “other”. To start the chatbot, we asked users to choose the topic that best suited their intent.

After talking to our support team, we found that it could be difficult to get technical help for a product on the website. There are rich and thorough resources that are better suited to get such answers. Instead of leaving someone in the chatbot, we used what we know about them in CRM to point them in the right direction. The chatbot looked at the contact record and then provided a contextual website depending on the products used.

Although our potential customers could not connect with someone immediately, we have set them up for better long-term success.

Next, we realized that a large percentage of the people who interacted with live chat actually wanted to speak to our sales team. However, that did not mean that trying to connect them to a human immediately was efficient for both groups.

We interviewed and watched the sales team to understand their chat experience. We immediately realized that each sales representative needs to know three important facts about the user in order to tailor the conversation to their company – name, email address and website.

In an ideal world, the sales team said they wanted this context before chatting to a prospect – a problem we knew a chatbot could solve. We have programmed the chatbot to collect this information in a natural, contextual manner.

We also knew that a prospect would get frustrated if it had to repeat itself – indeed NewVoiceMedia found it most frustrating aspect of the customer experience. To combat this problem, we checked the CRM again. If you’ve ever filled out a HubSpot form, we’ve skipped the total questions.

By collecting this information in advance, the sales team was able to spend more time selling and less time searching for email addresses.

Our results

At the beginning of this project, our main goal was to distract support-related chats from our sales team. But even as someone who’s been building chatbots for almost four years, I admit – what we saw was shocking.

When compared to Live chat, 75% more people deal with the chatbot.

In addition, over 55% of people gave real answers to the basic qualifying questions and reached a person. While the levy may seem steep there, picking out people with little intention proved incredibly helpful for the sales team.

Ultimately, our chatbot’s success came in a number of ways: First, the chatbot anticipated the next step in the conversation and used quick responses to drive a visitor. Opening a chat and clicking a context button is less smooth than entering a text box. Now users no longer had to create the first message themselves.

It makes sense that multiple choice options lead to more engagement. Think of it this way – when you meet someone, the most difficult part of the conversation is often the beginning when you try to say something. But once the conversation starts, things get smoother.

The same applies to our potential customers who deal with a chatbot or a live chat.

In addition, 9% of the people chatting with the bot needed help – we were able to help distract visitors from people and provide the best place to get quick answers to their questions.

It is important to note that when designing conversations, the most difficult and most overlooked aspect is how to write like a human. As consumers, we rarely process in our minds that we want to “speak to sales”. Instead, we only know that we want to talk about “products that I don’t use yet”. Writing with the “tasks to be done“In mind is a great philosophy for conversation design.

Snack stands for your company

At the end of the experiment, I found a few food stalls that can help you and your company succeed.

First, chatbots are a great opportunity to meet your visitors where they are. If you want to start with a chatbot, check out the chat transcripts or ask your sales team to understand the types of questions that are usually asked. Divide these questions into a few critical categories.

It is also important that you and your team think about how you can best help people’s different buckets. The more you can personalize – possibly through your CRM – the better.

Ultimately, using a chatbot to give people the help they need is a great asset to your potential customers and your business. In our case, website visitors now had the best resources for their support issues and our sales team was able to capture more qualified leads.

The best part? Both sides have done this smoothly and efficiently.

Editor’s Note: This entry was originally written in February 2019 and has been updated for completeness.