Communicating with prospects & customers - a novel concept?
About Common Sense - About John & Ed
You've met Ed here on the site.
For many years John and Ed have enjoyed talking about thorny issues around B2B marketing & sales, and recently we've started to record those conversations/debates. We're calling it "Common Sense" based on our shared love of history, and the perspective we hope we bring.
You can follow our musings on Common Sense at:
There's lots of buzz about Conversational Marketing - does it merit the hype?
Communication is critical to sharing ideas and selling deals. Demand gen, lead gen, sales and service all rely on communication that's effective - and increasingly what's convenient. So in addition to phone, email and text, now folks are focused on chat and bots. Are they legitimate game changers?
This episode includes:
- Why are we focused just on marketing? Why not sales & service too?
- Are bots as powerful as claimed? Aren't they just marketing automation? Or are they a simple form of AI?
- Why users will determine what's best
- There's little downside to companies adopting the technology
- Why some companies treat it so differently than other communication methods which they embrace
We dive into each of those and more as part of this conversation about whether conversational marketing really merits the hype.
Transcript Follows:
Ed Marsh: Hi, I'm Ed. Welcome to Common Sense.
John McTigue: I'm John. How are you doing, Ed?
Ed: I'm good, John. How about you?
John: I'm doing great.
Ed: I'm excited about our new format. We're starting this in a new way. Tell us about it.
John: Yeah, so we thought we'd change things up a little bit and make things a little more informal, have more of a conversation, more of a point, counterpoint on some of the hot topics that are out there in sales, marketing and customer service and the blending of those three.
We thought we'd start today with just talking about conversational marketing, which is one of the hot buzz words out there. Tell us what you think, what is conversational marketing?
Ed: Actually first, I just want to talk quickly about the point, counterpoint template. You say we're going to have a conversation. I hope we can have an argument. I mean, I think the beauty of this is too often everything in business gets painted as this binary sort of thing. You got to do this, or you can't do that, and so I think by taking a binary approach to these conversations, what we're going to do is kind of argue to a middle point and illustrate the fact that this stuff is much more nuanced than people often talk about, so I'm really excite about it.
We'll start by having a conversation about conversational marketing. Maybe we even call it conversational business. I mean, conversational marketing is this hip term that people have started talking about, but business has been conversational since people met at the bazaar and began doing transactions using pebbles as currency. It's not just marketing. It's conversational marketing, conversational sales, conversational recruiting, conversational service. Even potentially conversational management within the company. It's really a question of dialogue and real-time synchronist dialogue and using tools to facilitate that and make it more efficient.
What kinds of tools do people use for it? Chat bots are one that people talk about. There's chat tools like Drift and Intercom. There's Facebook Messenger, as long as not everyone hashtag deletes Facebook. We'll see if Facebook Messenger survives. There's Slack, which is kind of a conversational platform. It's not really an exchange as much as it is kind of a campfire. And there's more. There's a lot of tools.
Is it always the answer? No. When is it appropriate? Well, sometimes. Does it work? Well, short answer, yes it works and it has more potential to work and that's why people are investing money in it, but if you look at Chat bots, I mean clearly there's times that it fails.
It's interesting, there's a study done by Drift and Salesforce that as of 2018, I'm going to read some of these statistics off, 15% of American adults say they've used a chat bot to interact with a company in the prior 12 months. Getting detailed answered or explanations is 35% of how respondents might use a chat bot. 34% of respondents say they would use a chat bot to find a human customer service assistant. Speed and availability are where chat bots are perceived to provide the most value to customers. 64% of respondents said 24-hour service is a benefit of chat bots. The second most mentioned benefit is getting an instant response, to speed and convenience kind of bubbling up through these. That's mentioned by 50% of the participants. 43% of adult Americans say they prefer to deal with a real-life assistant rather than a chat bot, and that's no wonder, right?
Alright, that's why it could work, can work, will work. Tell me why it will never work, why it's a bad idea.
John: Well I'll start by saying that it's not new. It's really, conversation in marketing or conversational marketing, if you said that to a salesperson, they would laugh in your face because they've been using conversation, that's their primary tool in sales forever, since the beginning of time. It's really not something new in that sense, and even things like chat bots or chat apps, they're not new. They've been around for a long time. Websites that are interactive to some extent, social media, people have been using Twitter for customer service for as long as Twitter's been around. These things aren't really new, and to get all excited about it is really just sort of laughable I think to a lot of people.
Let's talk about when they're most useful, because I think that's the thing that we probably need to nail down the most, and how to use these tools because really, the point of all of this is to create a two-way conversation. What's new about this is that it's digital exclusively, this what they call conversational marketing.
It's being used in some new ways. It's being used as a lead generation tool. That isn't really something that chat features in the past were used for. They were used mostly for customer service, and now we're starting to use them in sales as well, so a quick response to a question about a product, for example, and having a salesperson there for a quick conversation. That's a little bit new. In the past they would have picked up the phone and would have had a quick conversation, which is just as effective, but now it can be automated. It can be online exclusively, and it's getting to be something that everybody expects from your website. Those are some new things about it.
Then obviously the technology side of it using automation and AI to even remove of the other person from the other end potentially. That's new too. What's old is new again I guess is the bottom line here. I'm a little bit excited about it. I'm not nearly as excited as a lot of people are just because of that and just because it has a lot of potential for abuse and I think we'll get into that next.
That's my next point, really is it can be intrusive. Anything you do online can be intrusive, and it can degrade the visitor experience if it's overdone. I think the things I see the most that I really kind of hate are these homepages on websites where somebody asks you a question when you first land of the page, "Why are you here? What are you looking for?" My response of course is, "Leave me alone. I'm here to browse. I'm using a browser so I'm browsing." Those things are interruptive.
Then just some other things, just trying to find out why the visitor is there, it's too early for that. Let them explore. Let them find out what they're looking for. Maybe deploy these things somewhere else on the website where they're more, their intent is more clear. They're obviously looking for something and they're exploring your product page or your pricing page or something like that.
What do you think about the intrusiveness of this so-called conversational marketing?
Ed: Well so I think that what's new about it is that there's people talking about how to solve for the buyer now. I mean, if you go back to when chat was originally used before, it was because companies didn't want to staff their customer service department so what they did was they put a chat thing there so that one person would handle five at the same time and let people feel like they were somehow being taken care of without actually focusing on taking care of them, and it's a very different application now where we're talking about trying to help people get to where they're trying to get to sooner, and your browsing example is a great one. I think a website is no different than a store. You walk in and you want to have just the right amount of attention to optimize that experience. I mean, it could be somebody to put their arm around your shoulder, walk you across the store, show you exactly what you need and help you pick it out, or it could be somebody to just say, "Hey, you know what? As long as you're here, I'll keep the lights on and if you got any questions let me know."
The trick is how can you adapt chat so that it responds properly to the person without being intrusive? That means you got to experiment a lot to get it right. We've all been victimized by crummy bots. I mean, my wife was showing me stuff the other day, a customer service interaction she had with an online retailer that clearly was a bot. She said, "Well this doesn't really answer my question." I said, "Of course it doesn't. They're trying to answer 50 questions with one canned response." That interrupts, that certainly degrades the customer experience.
If you can focus on expediting what people are really trying to do, looking for an answer, connecting with the right person, booking a meeting or a demo, registering for a webinar, some sort of ... It's not necessarily buying a product or service. There's these incremental steps along the way and they just want to get those taken care of at the time when they're ready easily.
Placing stuff strategically on the right pages where those conversions are going to happen, language and questions that are written in a way that are natural and friendly, but without being, well I'll say in millennial speak if you will, I mean not everything is awesome with an exclamation point, and so you have to adapt it for your appropriate audience. If that's the audience, then write that way. If that's not, then adapt it.
Everything should be solved I believe. In order to make it work, everything needs to be solved for the mannequin or the stuffed animal that I suggest the companies put in the seat at their conference table that represents the customer. Every conversation is always brought back to the reality of how does this impact positively or negatively the customer, rather than just a selfish thing of how do we solve for our own interests?
You're right, it can really be a pain in the neck. But, if they're done right, I think they can actually reinforce a great experience.
John: Well let me tell you another example of what I hate, is when you get into a quick conversation with one of these things and it asks you, "What can I do for you?" Or something like that. That's an appropriate question. I state my request, "I want to find out how I can cancel my account." Of course it goes crazy and it says, "Well, in case this conversation's dropped, please put your email in the chat box."
Ed: Right.
John: I'm going, "Well why is that? So you can send me more offers and more spam and that kind of thing?"
Ed: Right, right, I mean that's contrived. If you were using a tool that you thought was going to drop conversations, you wouldn't use it. You'd find another one. You're not going to implement a chat solution that you think's going to drop and lose people, so it's clear to everybody involved in that exchange that's a contrived excuse for you to get an email, and I get it. I understand why they want it. There's probably a better way they could ask for it, but maybe there's another time to ask for it instead.
John: Yeah, if there is some obvious follow-up that's needed I can send you that by email.
Ed: Right.
John: I'll go okay, and then ask for the email address. Then that makes sense, but I think that this idea of using these bots for lead generation is a little bit tricky. You have to really be careful how you do that because ...
Ed: On the other hand, if you use it to make it easy for somebody to access information, so rather than, I mean kind of the classic example that's been used for the last year is instead of the five or seven-field form, how about if you have a bot where you just push one button to answer a year or no question, and then submit an email address. Certainly that reduces the burden on the visitor to get that information. In that way it is, I believe, an improvement for the user experience.
The other thing that you can do, I mean who says that you have to conduct the whole conversation that way? Why not have, for instance, "Do you want to talk to somebody or do you want to email?" If they click the Talk to Somebody button, then pop up the phone number that they can direct dial whoever's on the other end of that conversation right then right there for continuity. We go back to conversation, nothing is worse than, "Oh, well let me read the notes and see what you were talking to the last person about." I mean, if you could just continue the conversation with the same person, then that really, that solves for it.
Of course at the root of it, if companies are going to invest money in this they got to be able to demonstrate an ROI, and that means there's got to be some tracking, they got to understand data from the communications and from the transactions, they got to see what's working and what doesn't. That means that they also probably need to be able to integrate this, or they should integrate, it if they want to optimize it they should integrate it with CRM and marketing automation. Maybe some of the conversational tools seem to be moving and kind of overlapping in that direction. Drift is building a lot of those capabilities into its platform, kind of approaching it from another angle.
Conversation, I would argue conversation doesn't have to follow a single rail. If it starts in online chat, it doesn't have to continue and conclude in online that. If it starts in email, it doesn't have to continue and conclude that way. Why can't we allow people to jump from channel to channel as is convenient or appropriate or desirable based on the conversation? I mean, there's nothing worse in my experience than sending an email, asking a question, getting an answer back that then spawns three more questions and have no way to get those answers quickly. I mean, that's the difference between synchronist and asynchronist communication. If you can make it easy for somebody to just jump between tools, that's a perfect situation.
I think clearly companies need to have data on which bots are effective. Which ones are pissing people off and losing them? Which ones are engaging them and getting them quickly to the answers? They got to have that means that they got to have processes in place. The one that people think about often is how do you staff it and how do you route inquiries? If you implement it for the marketing department, what happens if somebody wants to buy something? Well geeze, that should be an easy problem for us to figure out how to solve, but somehow companies stumble over it. Or if they want to order spare parts or customer service, so you got to be clear. How are you going to hand it off? How are you going to answer FAQs? How are you going to move something to customer service? How are you going to pass the information from the conversation so you don't make people endure explaining everything three times? What you're going to do about leads that are generated, how the workflow's going to work. There's a lot of potential, but there's a lot of pitfalls and you raise a number of good points about some of those frustrating scenarios.
John: Yeah, I think the key there, as you said, is really developing a process so that you understand what that engagement looks like at every point during the conversation, afterwards, hand off to sales, does it evolve into a phone call?
Ed: Right.
John: All of those. You've got to have those things nailed down or it really can be a worse experience for the customer.
Ed: Absolutely, and that's a horrible outcome.
John: Yeah. Cool. Well I'm just going to push back a little bit on some of the buzz that's out there about artificial intelligence.
Ed: Alright, be a buzzkill.
John: Everybody talks about AI and machine learning and how this is the coolest thing out there because you don't really even need to staff up customer service and sales. In some capacity you can just have these bots answering questions for everybody. The problem with that is that first of all, most of the technology you see out there right now isn't even AI. It says it is, but what it really is is automation. You have a series of rules, and if someone asks a certain question or selects a certain item in a list, something else happens like another question comes up.
Ed: Right.
John: That's automation. There's nothing intelligent about that.
Ed: Right.
John: I think they are heading in that direction, but the problem with machine learning is that the learning part, you have to have lots of instruction for a machine, an AI-powered machine to understand the language that someone uses, the wording, the inflection, all of those things in order to get meaning out of it and answer it in an intelligent way.
Ed: There's some great stuff happening on the telephone end of it, where they can extract information from tone and tempo of conversation and they can say, "This person's getting pissed off. Get them to the supervisor," and you lose that when you're talking about a type chat. You don't have those same signals, and so that's a great point about some of the kinds of nuance that we just haven't figured out yet.
John: Well the other thing about machine learning is it takes lots of data for that to work, so you have to have lots of instances, lots of different scenarios in order for it to really learn. How many visitors do you have to your website a day? Maybe 100? It's going to take a long time for a machine to really get it right if you have that kind of throughput. Now, I mean ...
Ed: What would be interesting though, I'd love to see people that take some of the data from telephone communications and figure out if you can overlay it on this. I mean, one article I read about telephone technology was that the single most definitive tip off that a call is going sideways is when the caller says, "That's ridiculous." Those two words, and people are programming their phone systems to pick that up and flag it and coach the operator on how to handle that.
There's probably some analogs that could be taken from this research around calls and overlay it on top of chat even as we are today, but I don't think anybody's doing it that I've heard of.
John: No, and it's going to be needed too for the reason you said. Yeah, but the human side is never going to go away. We're going to have to have humans monitoring these chats and instructing that the bot one way or another if things go wrong, so that's how they learn. I don't know.
Ed: Well and the Microsoft bot learned from humans. I mean unfortunately what the humans taught it was very unfortunate and very embarrassing for Microsoft, but that learning happens in two directions.
John: Yeah. Well that's my gripe about AI. Obviously there's a lot of buzz about this topic and conversational marketing in general and where it's going, and it has as you mentioned, it does have a lot of potential if it's done right. The question is how we're going to get there.
Ed: Right.
John: But then how is that going to really improve sales and marketing and how is that ... Is it really better than, for example, putting a person on the other end of a chat? Is it really going to replace that experience? When are you going to know when it's really working and better than the alternatives? I guess that's my key question, is where are we going with this thing?
Ed: I don't know what the answer is to that. I don't know how you, unless you do a massive A/B side-by-side test with some people only having human interaction and some people only having bot interaction, but where I think this conversation or dialogue or argument has gotten us to is that this point of this binary initial position, chat bots are going to solve all their problems or chat bots are rubbish, is itself absurd. It's a very nuanced kind of a conversation and the way companies implement it is critically important. If you approach it, "We're going to put in a bot and that's going to let us eliminate our switchboard or reduce the switchboard," it's going to fail.
Like everything that we've talked about in Common Sense, ultimately it's about solving for the buyer. Solving for the buyer. Now that doesn't mean that each specific step has to, but the outcome has to solve for the buyer. In that regard, some people, if I walk into a store, some days I want to be left alone. Some days I want somebody to show me where I'm going. Some days I want somebody that can actually answer the question without looking on the same rack or the same shelf or reading the same box that I already read myself. I mean, that's not assistance. Depending on my mood, depending on what I'm looking for, depending on what my goals are, I want different kinds of support and sometimes that could be a phone call, sometimes it could be a chat bot.
I think conversational marketing, sales, service becomes a tool to let companies solve for the buyer if they do it right and if their real intent is to drive more engagement and more sales. There's no doubt you're absolutely right, this technology and the tactics, despite ... I mean, the people at Drift might be offended or people at Intercom if they hear me say that it's in its infancy, but I think in general people agree. Where this is going to be in three years is completely different than where it is today.
John: For sure.
Ed: We don't know what's going to be the capability. On the other hand, any company, I would argue, that doesn't have this on their radar and isn't thinking about how to pilot it and how to test it and what situations to use it for, is missing an opportunity.
think we've argued to a reasonable place in the middle that it's something that companies ought to think about and there's pros and cons and a lot to be aware of. Is that a fair summation of where we've ended up?
John: Yeah, I think so. I mean, I've had good experiences with chat bots so when that happens I go, hmm, maybe there's something to this.
Ed: Right.
John: I do agree.
Ed: Cool. Well, that's it for this episode of Common Sense. Thanks everyone for joining us today. I'm Ed.
John: And I'm John. We'll see you next time.
