The mistakes we’ve made when using chatbots, and lessons that helped us convert 5% more customers

Without any doubt, bots are the presence and the future of marketing and sales. They make every process way easier.

But it’s not easy to use them. When not implemented the right way, they know to be intrusive and the only thing they will achieve is actually turning your visitors down.

So, how do we do that?

How can we achieve the amazing results we read about in different case studies by using bots? How can we convert X more leads with them?This article is the story of how we failed to use bots, what we learned from it, and how do we convert over 7% of our content readers into hot leads and later into customers.

Let’s get started.

The problems we encountered

Userpilot is a user onboarding tool. Founded only a couple of years ago, we’re one of the pioneers of the Product-led growth (PLG) movement.

This being said, since we’re in a pretty much new market, a lot of our target users weren’t educated about the problems we’re solving.They weren’t aware of the troubles they had, they weren’t sure what exactly Userpilot does.

Keeping this in mind, no matter how great our content was, it just didn’t manage to convert readers into customers, or even worse, into leads.

So we needed to change something.We needed to change our entire approach to the content lead generation game.

So, we decided to use bots as our go-to way of generating customers through content.

We have used bots for many content types - playbooks, ebooks, statistics, articles, and other lead magnets. But in this particular article, we’re going to see how we implemented the bots on a series of different high-intent blog posts.

High-intent blog posts are essentially the articles that are searched for and read by the people who are almost immediately ready to buy Userpilot.

How do we know that?Because they’re targeting people who are currently using some of our competitors, but they’re not satisfied with their solutions.

In the nutshell, those articles are targeting these two keywords:

• Appcues alternatives
• Walkme alternatives

If you pay a little attention, you will also see that we did our homework, and we’re currently ranked first on Google for these keywords:

So we definitely had a bunch of people coming to our blog and reading the article.

But, although the article was pretty in-depth and actionable, they were not converting the way we wanted.

So, we decided to bump up our conversions by implementing bots.

This is the moment where the story begins.

Our first attempt at using bots, how we failed and what we learned from it

So, here we start.Long story short, our first attempt to capture more customers with chatbots went horrible.

We sat down, targeted everyone who was reading one of those two articles, wrote down the copy and created a flow.

As the end result, from more than 200 visitors, we engaged only 0.7% of them and got 0 emails and customers.

When it comes to the targeting, we wanted to engage with everyone who was currently reading one of those two articles, and who lived in the USA or Western Europe.

Mistake #1 - Not everyone who reads your content is ready to engage with you.

When it comes to the flow itself, we were asking if the user needs more clarification about the differences between Walkme/Appcues and Userpilot.

They could answer only with Yes, please! and No, I’m good for now.

Mistake #2 - Always offer a third (optional) option to your visitors.

If the visitors say Yes, then we ask them what’s the total number of monthly active users (MAU) they have, so we can qualify them better later on.

Mistake #3 - Don’t add friction to your chatbot workflow - Every unnecessary step is a waste of time and a chance to lose your customers.

Once the readers answer that question, we push them to schedule a demo, and there lies our third mistake.

Mistake #4 - Don’t be too offensive with your visitors - Someone who’s still isn’t qualified (like the readers in our case), definitely won’t schedule a demo with you. They need a bigger push and more education. When it comes to the final call-to-action (CTA), it should be something contextual to their current user journey.

In the nutshell, here’s how it looked like:

Reforged chatbot workflow that converts over 5% of the total content consumers

Every failure is a new chance for success. Without failures, there’s no room for big achievements.

Once we went through fire with the workflow from above, we learned those four lessons.

Now we were ready to go a step further, improve our chatbot workflow, and convert way more visitors into hot leads.

As the end result, we were able to engage with almost 8% of the total number of readers and make 64% of them to give us their email addresses.

At that point, we improved our targeting and we were able to narrow down our focus.Instead of just targeting people who were in the USA/West Europe and who were reading one of those two articles, now we also targeted people who spent more than 240 seconds (4 minutes) reading the article.

Why?

In the first workflow, we were targeting everyone, no matter how long he remains in the article.

But in the second one, our target audience was people who actually read the content piece itself. This means that they already went through the majority of the article, and they already have “opinions” regarding our content.

Achievement #1 - More in-depth targeting allowed us to engage with more relevant people and get attention.

Now we approach the workflow itself.

This time, we changed our game a little bit.

Here’s what we changed:

• Instead of just asking if they need help, we now also offered them a touchable lead magnet. Essentially, our initial message was: Heyo! Are you looking for Appcues alternatives? I have a sheet that could save you hours of research.

Achievement #2 - Offer tangible value straight at the beginning, it will intrigue your visitors.

• Besides Yes and No answers, we also offered them a third, custom option, allowing us to jump into the conversation immediately.

Achievement #3 - Not everyone has the same struggles. Someone needs a different kind of help. Offering “custom” answers allow you to engage in more conversations.

• If the customers answer with Yes, we will ask them for an email, so we can send them our comparison sheet.
• Once the customers give us their email addresses, we send them over a sheet in the inbox, and leave the chatbot window open, so we can spark further conversations.

Here’s how this flow looks like:

Overall success - from almost 500 people who read our content, we’re able to convert over 25 of them into hot leads, allowing our sales team to close them.

The bottom line

Chatbots are a great tool for converting your content consumers, but only if you implement them in the right way.

If the implementation is poor, your visitors won’t engage with you and you will lose a bunch of great opportunities.

So, at the end of the day, based on our personal experience, here’s what we learned:

• Always add value to your content, and try to educate your customers even more.
• Your main goal is to engage with your customers, think about the ways you can do that, no matter what
• Don’t be too offensive. Understand your visitors and the stage of the user journey they’re currently in. Offer the right CTAs for the best engagement