Chatbots were seen as one of the biggest trends and most exciting digital marketing tools in 2017. While enthusiasm was tempered a bit this year, they are still extremely promising and can be a very effective part of your marketing strategy.
Chatbot adoption has been slower than expected, partly due to limitations of early bots, and the hurdles of understanding how they work and how to set them up. However, they are still projected to be very prevalent by 2022, and account for $8 billion in cost savings according to Juniper Research.
It's time to give them another look.
We think chatbots can be very beneficial for many companies, and our recent broadcast covers how to evaluate whether they're right for your business, and how to get started.
The technology has come a long way, and is still evolving and constantly improving. There are currently some very good options available that businesses should consider.
- Improve customer experience
- Save companies time
- Qualify leads
- Increase engagement
- Increase conversions
- Aid in customer retention and repeat sales
You can read the transcript of our broadcast below. Prefer to watch the video to see some of the chatbots we mention in action? You'll find that below too.
Chris: Hello everyone, and welcome to WebStrategies Live, the video broadcast for the modern marketer and business owner where we share with you how to generate leads and sales online. I'm Chris Leone.
Phil: I'm Phil Woods.
Chris: And we are broadcasting live from WebStrategies' headquarters just outside Richmond, Virginia.
Today is an episode that we've wanted to do for quite some time. It's on a topic that's starting to gain a lot of momentum and has for the last year or two. But there are a lot of tools that are out there and that is now making this accessible to businesses of all sizes, and that topic is chatbots. So, today, we're going to talk about chatbots, why they're important to you, what they are, and how you can increase conversions on your website using them.
So, the first thing that we want to talk about is, what are chatbots? If you're not familiar with chatbots, they're essentially a software application that simulates conversation with a human user.It's very much like live chat which has been around for maybe 10 years or so except the person on the other end isn't actually a person. It's a computer that's doing the talking back to you. And there are some reasons that this can just be hugely advantageous to businesses which we're going to get into, but it's really exploded recently because of the technology and the processing capability to run these types of applications.
So, that's in a nutshell, what chatbots are and we're going to go through some demos, so you can see what they actually look like. Now, there are two primary places that you're going to see chatbots. The most common place to see a chatbot is on a website. So, maybe in the bottom right of a website, you'll see a little chat icon, you may think that that's actually a live chat with someone on the other end but it very well could be a chatbot instead.
The second place that's also common to see chatbots now are within messaging apps, like Facebook Messenger. So, a lot of these brands have a chat in their Facebook Messenger, when you go to the page, you'll see it, and if you start talking to it, it'll converse back and forth with you. So, those are the two primary places that you may have seen chatbots and likely have interacted with chatbots in the past. Now, a couple stats to throw out there, just to give you an idea of how pervasive chatbots are and are becoming. Chatbots are going to power 85% of all customer service interactions by the year 2020, that's according to a Gartner research paper.
Also by 2020, over 80% of businesses are expected to have some sort of chatbot automation implemented. So, four out of five businesses within just a couple years are going to have some form of chatbot. And I would say maybe in the last 12 months, we've seen a huge influx of chatbots not only just on sites that are using it for marketing purposes, but also for customer experience.
Phil: Absolutely. And we're certainly getting a lot more clients now asking about it. So, it's making its rounds in the marketing circles, people are talking about it. Success stories are being thrown out there and people are really starting to get interested in it.
Chris: So, the four big things that chatbots can do that might peak your interest. First and foremost, chatbots can help generate leads on your website. And this is a show about generating leads and sales online. So, most of what we talk about today is going to be focused on the lead generation aspect of chatbots.
Another very popular way to use chatbots is for customer service. So, if you're dealing with a lot of customers who are coming at you at all hours of the day, this is a way that you can scale that.
And then if you're just trying to increase engagement on your site, maybe you have a lot of people coming to the site and you're trying to get them to spend a little bit more time, maybe view a little bit more content, chatbots can increase engagement because it's another way to deliver content instead of making someone click around to different pages on the site. Maybe you can deliver content through the chatbot.
And then finally, qualifying leads. That's more of a subset of lead generation, but something that we've seen a lot of success with is having a chatbot ask questions of your users and then only try and schedule time with those who answer a certain number of questions the right way.
Now, two ways that chatbots are generally set up, we're just going to skim over this just so you have a basic understanding of it because then we want to get into some examples of chatbots. So, probably the most popular way of using chatbots right now at least for small medium businesses who are working on a budget is using this decision tree format. Essentially what that means is, when you land on a site, it'll ask you a question and then maybe give you multiple choice responses. And then depending on how you answer that, it'll drop you down kind of this conversation tree, this one branch of predetermined questions and answers. So, there's really no machine learning natural language processing, artificial intelligence, or whatever baked into that. It's just more of if this, then that. But that can still be hugely successful and we're going to show you a couple examples of where that's been very successful.
The other more sophisticated way of using a chatbot which is where I think things will ultimately go, is within this world of natural language processing. Essentially what that means is, instead of giving pre-formatted questions and answers within a chat window, it leaves it open-ended and then you type in "I'm looking for X, Y and Z." And then it can pick out what you're saying, and then actually start to converse with you based on what you're saying. So, it's a computer's ability to infer both explicit and maybe implied intent within a message. That's natural language processing. Those tools are out there, they are more expensive and maybe the business model and how you buy those and how you integrate them into the business is a little bit different. But those are generally the two ways that chatbots are powered. So, decision tree format and then natural language processing.
Before we get into why use chatbots, Phil can jump over and show a couple of examples of a working chatbot.
Phil: This is a company, a client of ours that sells faux wood beams. And this chat icon is here, and depending on which page you're on, this icon might bounce, it might open up, but this icon is always there no matter how you interact with it. This is the first question. Welcome, how can we help you? And there's four choices. And then this is the decision tree formula that you talked about depending on where I go in here. I'm given another set of follow-up questions. And depending on where I go, there could be a contact form at the end or a quote form. There could be links to product information, links to FAQ, there's a variety of different destinations within this workflow.
Chris: Right. Do we have the Facebook example up as well?
It's going to look just like your normal Facebook Messenger. So, Food Network has a chatbot baked within their Facebook Messenger. Again, that's one of the several ways that you can deliver it. So, this is also set up like a decision tree format. Phil was interacting with this just before the show. And it's asking him a series of questions, like who's your favorite chef or what kind of food do you like? And then based off of that, it feeds you recipes.
Phil: Feeds you recipes. I see what you did there. You can also, in this particular one, it's a little bit more sophisticated because they have this synced up with a recipe database, you can type in. If you just wanted chicken enchiladas, you could type that in and it would spit you back a recipe.
Chris: So, a little bit a natural language processing. Again, there's...the technology that's built into these, it's either out of the box or sometimes you buy a tool and all the decision tree stuff will be baked into that and you set it up. Other times, there are technologies that you can build into this chatbot. In the case of Food Network, they might be using another software that reads into a database and does some fancy stuff that way. So, it works several different ways. We're not going to get into that because we're going to focus on what small and medium businesses can do for this which is usually not something of that level of sophistication.
So, let's talk about why use chatbots. Well, there's three primary things that a chatbot accomplishes especially for small and medium business, but even a large business as well.
First is scale. Scale, when I'm referring to scale, what I mean is chatbots allow you to pretty much interact with an unlimited number of people at the exact same time which is something that, a small business can't do or at least cost a lot of money to be able to do something like that. Chatbots are expected…this is a really interesting study that we saw recently and it was looking at both banks and healthcare websites. And what it was able to determine from that is that chatbots are able to save 50 to 70 cents per interaction, just because of the scale that it offers and it doesn't require you to hire as many people to manage that conversation. So, scale is one of the reasons to use chatbots.
Another reason to use chatbots is availability. 51% percent of consumers say that a business needs to be available 24 hours a day, 7 days a week. For a small and medium business, especially small business, that's pretty difficult to do, right? You have to either have somebody on call all the time or you have to use a call center and pay them to be available all the time. And if you're dealing with large volumes at certain times a day, that can become pretty expensive. So, it allows you to have availability at all hours of the day without having to hire extra hands.
And then the third reason to use chatbots is customer experience. Seven minutes is the average wait time on a customer service phone call. So, that can be knocked out immediately. And then, another study that we saw said that by the early 2020s, 75% to 90% of all bot interactions will be successful. In other words, three quarters, or 9 out of 10 people who engage a chatbot on the site are going to have their problem solved as a result of the bot. And again, because bots allow you to interact with people at infinite scale, that's a huge value proposition, right? Because otherwise, the way to replicate that is you have to have an unlimited number of people on call at any given time to answer somebody's question.
So, that's really the main reason that we're interested in chatbots and that we're excited about chatbots. It can certainly save you on having a call center or a certain size call center, but then also if you're using live chat right now, there's only one person to answer live chat or maybe a couple people, you could get rid of that and then you can have it all run through the bot or at least maybe run through the bot up to a certain point at which point a live human comes in to take up the conversation.
Chris: So, Phil, let's go back to demoing the chatbot, so we could start to show a couple examples so people can get a better understanding of ways that we've seen this work really well.
Phil: So, in this example, Chris mentioned earlier four ways that chatbots are typically used, or used to generate leads, to qualify leads for customer service or to increase engagement. And I would say in this example, it can do any or all of those things. So, I'm going to walk through for one, this particular client wanted to make sure that they were giving their commercial clients more attention, better attention in some ways than the residential clients. The commercial ones were of higher value and you'll see here in this example, that's how they used their chatbot to qualify the commercial lead. So, one of the first questions, if I just say, hey, I'm looking for beam, so, I'm looking to find the right beam, it comes back and says, okay, just kind of tell us what type of project this is for. And it's very specific because there's residential or commercial project.
Chris: And again, we're trying to get better engagement on commercial leads specifically without turning the whole website into a commercial website.
Phil: Right. And the thing about it though, in this business, the commercial leads really need more hand-holding, you can't sell that product online. It's more complicated, it needs to be designed in many cases. So, what they really need is to capture these leads so they can follow up with them in the right way. But in the meantime, it says, "Great, we've helped supply beams to many commercial jobs, describe the project." I'm going to type in something about this project. Let's say it's, you know, a lobby in a hotel needs beams. Now, in this case, you think you're typing in a specific message, it's going to get you an answer. This particular tool is not set up to be that sophisticated. So, what comes back, it says, "Thanks for providing us that details. Can you provide us with your information so we can get in touch with you about the project? And this is just more like a simple lead capture form. But the science behind this shows that once people interact with the chatbot, they're much more likely to convert and give up their information than say if you just had a quote form on the page.
Trying to jump straight to that quote form, a lot of people are reluctant to take that step. But if you've already interacted with the chatbot, you've already typed back and forth a couple times, you're much more likely in this scenario to give up your info and that's part of the reason why chatbots are so effective.
Chris: Awesome. Let's go to the next example. So, we're going to look at the WebStrategies' website. So, we have several industries that we talk about throughout our website, one of which is credit unions. So, if you land on one of our credit union blogs, we have a chatbot set up just for that industry. So, that's another cool thing that you can do, it's not a single message on every page of your site. You can set up rules to say if somebody is landing on a page of the site that is about this or maybe has this word in the URL, show them this version of the chatbot versus if they land on another page, show them this version. So, we've customized this in several ways for that industry.
So, the first question, hey there, are you looking for ways to market a credit union? Let's say yes. Okay, so, a similar window's going to pop up. You've come to the right place. Do you work for a credit union? Now, we ask this because we get a lot of people in other agencies who are looking how to serve their clients who are credit unions come to our site. So, we want to flush that out pretty quickly because we're not going to talk to them. But if they say yes, I work for a credit union, great, we work with credit unions across the country. Now, this is a chance for you to kind of pitch yourself a little bit. Would you like us to share three trends about digital marketing as it relates to credit unions? And let's say, sure, I am interested in that.
Now, this is something that's really cool. Is that these chatbots can start to integrate with other tools. So, you know, we use HubSpot for our marketing automation and for our CRM, we have our calendars plugged into that as well, and in this case, Kristin is somebody who talks to our credit union prospects. So, Kristin's calendar is linked into HubSpot which is then linked into this chatbot. And so, if you click schedule a call right here, it actually reads her calendar and then puts calendar times up here and then you can select the time. And so, we've actually gotten meetings booked using this. It's totally hands-off, you know, maybe in the middle of the night or late at night, somebody comes around and then they book a meeting and we show up the next day. And we have a new meeting on our calendar. It's pretty cool, right?
And there's a lot of different technologies that are at play here, and again, this is primarily just a decision tree with some other technology baked into it and we've seen that be pretty successful for us. In fact, when we launched that, we saw that we got as many meetings booked through this chatbot as we did through a normal contact form on our website. So, pretty cool stuff.
All right. So, let's jump back to the best ways to get started. You know, I'd say the first thing here, this is very, very important. Make sure you're clear about what you're trying to solve before you introduce a new shiny tool like chatbots. Chatbots are definitely the shiny tool of 2018, of 2017, and artificial intelligence is becoming the next shiny thing too, right? But you have to have a very clear idea of what you're trying to solve for. So, in our case, we're a site that gets a lot of volume in tens of thousands of visitors a month to our website. And that's a lot to sort through and a lot of them are not qualified but some of them are. And we want to try to find a way of maximizing the conversion rate that we get on that qualified traffic without wasting our time on the unqualified traffic.
So, in this case, this is more of a conversion play for us and a qualifying play for us. You know, customer service is a huge play for a lot of organizations as well. That's not for us but, you know, if you have a lot of existing customers coming to your site and they're asking very common questions and something that you can kind of pre-program, you can set up a chatbot to answer those questions or if people are just trying to set a quick appointment, if you're like a plumber and I need a service call tomorrow. You could do exactly what we did with that example that I showed you with the credit union, but instead, it's booking time on…you know, it's booking service appointments through there. So, there's also that potential customer service element to these as well.
If you have an engagement problem, if you have a lot of people coming to your site but you feel like they're not getting to the critical content that you want them to get to. You could try and deliver that same content through a chatbot because, again, ultimately, if you can try and add a little bit more value to that visit, you know, chances are they might come back to you a little bit later. So, that's another way to deliver content you already have but through a different experience that feels different, but is ultimately having the same effect that they otherwise would have had on the site.
Phil: Yeah. So, I think one of the other points around getting started and this is something we come up with or come up against all the time. And that is, you've got to shut down whatever stereotypes you have about chat or whatever common objections you're going to get internally with your organization. And those are like, oh, you know, who's going to run this? Who's going to manage it? Who's going to monitor it? You know, this is not live chat, this does not need to be manned. Though some of the tools will allow a user to be logged in, and if you see what looks like a good prospect interacting with the chat tool in a live sense, you can jump in. But it's not mandatory, it's not like the old-school live chat. So, that concern of who's going to do it, it should be a non-issue in this case.
Phil: The other is, you know, when someone says something like, oh, most people find that annoying. I know when I go to sites, I find that annoying. The truth is, if it's set up the right way and you're asking the right questions, most users actually find value out of that interaction. And a lot of people, you'll be surprised, will interact with it and you'll end up learning a lot from it. So, our feedback to people is just jump in, try it. In most cases, you can figure out a pretty straightforward decision tree. It's either help qualify your leads or at least steer your leads in the right direction towards the right product or the right service offering and then trying to capture that lead in some way. If you set out with that kind of narrow set of objectives, you'll end up learning a lot from trying some of these tools.
Chris: Yeah. So again, your chatbot could schedule a meeting, a phone call, your chatbot can send somebody to another place on the site, your chatbot can capture information and then you can do a call back later, your chatbot can simply deliver information directly through the system. Or it could maybe even send it to a certain sales rep. So really, it's unlimited, the applications of this, but that can be a little bit of a burden when you're trying to get up and running because there's so much that you could potentially do. But I would focus first on the exact problem that you're trying to solve for and then work backwards from that.
The final piece of advice that we have on getting started is just do it. Don't wait, don't overanalyze it, just get it on the site, get a feel for it, and then you'll go from there. That's what we do with pretty much everything, and we find that usually, our biggest revelations don't usually come from exactly what we expect to happen where we overanalyze like, well, we want it to happen exactly this way and we'll see what happens. It doesn't usually work that way but there's something like maybe off to the side that we learn. And then we say, hey, let's double down on this and then that's where we find our most success. So, just get out there and start doing it.
Finally, which tool is right for you? There are a lot of different tools out there and they come in a lot of different shapes and sizes. If you're focused on lead generation which is most people who watch this show are really trying to figure out lead generation for their website, there's a few different places that you could go to.
Drift is a chatbot service, ChatPath, similar to that, and then HubSpot is coming out with their own as well, it's in beta right now. We have access to it because we're partners. I think it's going to be going out to the public in beta as well in August. And we've tried all three of these. Right now, our preference is Drift, it just works the best and they're really developing the tool well in a short period of time. ChatPath is more affordable but maybe a little bit more limited. And then HubSpot, I think, probably has the biggest potential in the long run but they're behind the other two in terms of features and just maybe general clunkiness, but I expect within the next few months, most of that will be ironed out. So, Drift is our pick, if you're just focused on lead generation.
If you're trying to do customer service, you have a few options. You can still use Drift, ChatPath, HubSpot, you'll be able to use soon. And again, that's if you're installing these just directly on your site. If you're trying to integrate it into a Messenger app like Facebook, ManyChat and Chatfuel are two other services that will bake into Facebook as well. Those are really kind of the big ones that create a product for everybody.
But there are also now some more niched chatbots that are coming out there. I was doing some research prior to this show, and I found a very interesting company out there that's developing chat, specifically, I think for real estate companies. So, this is where it gets really interesting, right? Because all this natural language processing and being able to kind of semantically understand what someone means or intends in their messaging, that can be very unique for each individual industry especially if it's like a highly technical industry, right? There's a lot more data that you have to feed into it so it can understand what people mean when they throw out something that's maybe more kind of inside baseball about your industry. But one simple version that I saw that I thought was pretty cool was for a real estate company, and they've created this technology where you can come in and someone will says, "Hey, I'm looking for, you know, a four-bed, three-bath house in Richmond, Virginia. And my budget is, you know, $400,000." So, it's able to pick up every one of those and then, you know, start to categorize somebody because it understands what bath and bed and budget means in this context, right?
And then it can create a follow-up question that says, "Great, when are you looking to move?" And then say, "I'm looking to move within the next, you know, 60 days or, you know, whatever it is." And then based off of all that, it can kind of pick it apart and then, you know, send it off to the right agent based on those responses. So, that's the idea. The thing is that I thought was really interesting is that that technology, you can license. So, if you're a real estate agent and you have an office in town, you can go to them and you can license that technology and then you can build it into the...build it into your site. So, that's where I see limitless opportunities for this, right? Engineer…I mean, anything really. If somebody invests in that kind of artificial intelligence, natural language processing at the specific industry level, then you can license that technology and then have a much more sophisticated chatbot versus just using out of the box maybe decision tree type chatbots, okay?
So, there's a lot different ones out there. They will run you probably anywhere from maybe a $100 a month and up. The ones that we tend to like a little bit more might be $200, $300, or $400 a month, but we find that the capabilities within the tool and kind of the smoothness of the tool certainly make it worth it. And then, again, if you can just generate maybe like one more lead a week, or even a month, or even a year off of this thing, it can very quickly pay for itself.
I would encourage you to do a search for chatbots and then your industry name, and then see if somebody's out there developing something that's very unique to you, that you might be able to plug in to and then place on your website or even within your Messenger platform. The good thing is you don't have to be a developer to use this stuff. You know, they have very nice interfaces, you could jump right in if you're doing the decision tree thing. If you want to do something more sophisticated, you probably have to get developers involved as well.
Chris: I had a couple questions that were sent ahead of time. So, the first question...all right, so, this is somebody who pays attention to the news. Did Facebook shut down chat? And I think what they're referring to here is when the whole Cambridge Analytica thing came out, Facebook kind of put a hold on a lot of things and one of which was getting new chatbots approved within Facebook Messenger. So, I think that happened maybe back in March or whenever the whole Cambridge news broke. But they've lifted that, you can now submit new chatbots for approval. So, that's...as far as I know, that's totally wide open now. So, you can submit new bots.
Should we use Facebook bots or onsite bots, which do you prefer? You know, again, this kind of goes back to what your goal is. Again, this is a show that really focuses on leads and sales, so, I would probably say focus on site in that case Drift, ChatPath, or the new HubSpot tool that's coming out. I would point you in that direction. You know, for somebody to engage with the Messenger app, they've got to be on your Facebook page, they've got to maybe go from your website over to your Facebook page or something like that. It's some extra steps involved. The only caveat I would say to that is if most of your attention is happening on Facebook. So, in our case, most of our attention, a lot of our clients, most of their attention happens on their website. So, we're trying to, you know, monetize all that attention on their website. But if a lot of that attention is happening on your Facebook page, if you just have a killer Facebook presence and that's where most people are interacting with you, then I would go over to the Facebook app, right? And then use ManyChat or Chatfuel, or any of the other tools that are out there to build chatbot technology with into your Facebook Messenger. And what's really cool about that is when you land on…when someone lands on your page, a little Messenger box pops up in the bottom right and then you can interact with it right there. And it feels...it pretty much feels exactly how it would if somebody was on your website. But it's just happening on that Facebook page.
Okay. If you guys enjoyed this, please like it. Please leave a review on our Facebook page, we love those, we'd love to get the feedback and know that you guys are enjoying this. If you're watching this after broadcast and you have a question, leave it and then we get alerted to that, and then we can come follow up with you after the fact.
If you have any other questions about this episode, about the AEC framework, please leave us a comment below!
Agree, disagree, or just have something to add?
Leave a comment below.