Uber's Chris Messina on conversational commerce

Recently he’s been writing about a new trend at the intersection of brands, bots, apps and AI, where messaging is no longer just a communication tool, but a transactional one as well. Chris has labeled the idea Conversational Commerce. You can check out his recent post on the topic here and track the trend yourself with #ConvComm.

I had the pleasure of hosting Chris at Intercom’s San Francisco offices to chat about Conversational Commerce, his new role at Uber, the legacy of the hashtag, and more.

If you like what you hear, check out more episodes. Subscribe on iTunes, Stitcher, or SoundCloud or grab the RSS feed.

What follows is a lightly edited transcript of our conversation. Short on time? Here are seven quick takeaways from our chat:

1. Consumers familiarity and comfort with messaging creates a new market for offering services and experiences. That medium is the foundation of Conversational Commerce.
2. Messaging integration frees developers to deliver services where users already are, rather than an app store. When you remove the store model and focus more on virality and social channels, conversations will lead people to discovery.
3. Rather than designing for web, mobile or desktop, or bot or chat, developers will build experiences around the context in which people want to interact with a brand or service.
4. How the rules of privacy will apply to bots and AI – and whether they’ll be held to the same standard as humans – is still a massive unknown.
5. Brands must begin thinking, what is the resonance they’re creating, and how will it come through in the bots and agents they deploy?
6. Old behaviors (e.g. IRC culture) hit the mainstream when a product market fit is finally realized. For Chris, the hashtag was his moment of recognizing that mobile was going to be a massive, important medium.
7. What’s the next product design frontier? Hardware like the Amazon Echo that brings the internet into people’s everyday lives.

Adam Risman: Today we’re joined by Chris Messina, Developer Experience Lead at Uber. Chris, welcome to the show. For the sake of our listeners, could you quickly introduce yourself and just give us a feel for your career trajectory? Where you’ve been and how you got to where you are today?

Chris Messina: I’m one of those odd-shaped people that whenever I go to interview at a big company, they don’t quite know where to put me. This has happened several times. I studied communication design actually, and arrived in the Valley about 10 years ago, starting out with the Mozilla project.

I helped co-organize the original launch of Firefox and then got involved with a bunch of community projects like BarCamp and starting co-working here in San Francisco. I had my own consulting agency and then I’ve also worked at Google, both as a Developer Advocate and a UX Designer. Then I took some time off. I worked for some digital art startups, and now I’m at Uber.

Adam: At Google, you were doing a lot with developer tools, right?

Chris: Yeah. I created Google Developers, which gave Google a mouthpiece to speak to developers in a more uniform way.

A new era in messaging

Adam: For the past year or so, you’ve been tracking a really interesting trend that involves the convergence of messaging, bots, AI, and a slew of other interesting technologies, which you branded Conversational Commerce. What’s the story there?

Chris: In January of 2015, just over a year ago, I started to notice this shift. Path Talk was actually one of the biggest ones that really gave me an insight into where this was going. This feeling that you could talk to brands the way that you can talk to people. I wrote about it then and wrote about it again this January.

Adam: Why will this finally be the year the concept comes of age?

Chris: One of the things that’s interesting about being in Silicon Valley, being in San Francisco, is that we kind of live in the future here. We’re two to three years, easily, in the future. There’s two interesting things happening now.

First, there’s this app fatigue. There’s constantly new stuff coming out. I’ve become a lot more involved in Product Hunt in the last year and seen the challenge new apps have of actually getting discovered. People install apps, and they never go back to them. Most people who have smartphones don’t even install apps as it is; they just use the stuff that comes stock. That creates a real problem in terms of growth and opportunities for companies and services that want to meet people where they are with the devices they have.

On the other hand, you’re like, “Well, what is an easier, simpler way to deliver services to people at a lower cost that meets them where they are?” Most people on smartphones are comfortable with the idea of messaging their friends. Given that there’s this huge amount of activity there, you’re not going to call an app, per say, but you might interact with it through a messaging context that’s more comfortable or familiar to you. That creates, or at least has the preconditions to become, a really interesting market for offering services and experiences.

The other thing to look at is how we’re actually starting to fall behind in some respects, because the Asian and Indian markets are ahead of us in terms of developing these conversational operating systems. We like to think of ourselves as creating the future and being in the place where all innovation is happening, but that’s not necessarily true, if it ever was true. That has given me pause to start thinking, “What does that mean, and what does it do to us in the way we think about what we can build if we’re no longer the front-runners?” It changes the whole perspective and paradigm.

Adam: It seems like a very important lens, particularly for developers, to begin looking through — if they’re not already.

Chris: Absolutely, on a number of fronts. Distribution and getting your app discovered is a big part of the puzzle today. Going to the app store and waiting for it to be approved is a big piece of the launch process. There’s reasons for that. There’s a quality that, especially in the case of its own App Store, that Apple wants to keep. But in the case of moving quickly, iterating fast, and again, being more adaptive and responsive and personalized to your customer, the messaging paradigm really frees developers to be able to deliver their services to people wherever they happen to be, as opposed to having to first install, download, and learn a new app paradigm every time they want to try something out.

The App Store’s Expiration Date

Adam: I think that idea of service discovery is really important here. In many ways the App Store was an accident to begin with. If that current model is not going to be around much longer, where are people going to find these bots? What’s that new marketplace like?

Chris: It’s funny. I wrote a post, I think in 2007, because the iPhone came out in 2007, about how Steve Jobs hated the App Store. I think it was fairly controversial at the time, and now the App Store makes a huge amount of money for Apple so that might have been a little premature to state, but the idea was that the extension model for Apple was really around the web and WebKit. Finally, you had a really performant browser on a mobile phone that didn’t just slim everything down to raw text. It’s interesting because we have this period of time where we brought the concept of a store into the computing context, which previously didn’t exist. There was a time when there was no Apple App Store and you just found apps through shareware and other weird distribution.

Now, we’re in an era of a glut of apps in this App Store that are really hard to be differentiated. There’s no natural virality, per say, in terms of discovery, except what happens through social networks. When you remove the App Store and you focus more on virality and social channels to spread apps, then it’s the conversations themselves that lead people to discover things.

You’re starting to see this in a number of contexts, whether it’s Slack, Intercom, or Facebook Messenger. People will say or mention a bot or service and you can immediately go and interact with it. Let’s say you’re on Slack. You can just connect the bot to your Slack instance, start having conversations with the bot, see if you like it, and then go from there. It’s a very fluid interaction, just like if you were to add a new contact in iMessage. You should be able to add and interact with bots in a very similar way.

Adam: So it’s almost self-serve meets word-of-mouth?

Chris: Word-of-mouth is certainly a big part of it. Also, the problem with installing software, of downloading this big binary that can take minutes, is the time to satisfy and intent. If a user is interested in playing a game or trying out the new food delivery service there’s a question as to how long it’s going to take from the moment the user conceives this desire to having that desire satisfied. In the case of a game, it could take tens of minutes, and maybe I have enough motivation to go back and actually play the game, but in the case of a food delivery service, I may have forgotten I even downloaded the app by the time it’s installed. Whereas in the case of a messaging bot or app, you add the contact, you start talking to it, and you can have your needs met, or not met, in a very short amount of time with very minimal attention. That economic shift will drive a lot more discovery and experimentation, as well as social sharing.

Adam: Does this shift that you’re describing mean the distinction between web and mobile is going to blur and ultimately disappear? Will people just want access to a service or functionality at that point in time, regardless of what device they’re on? To me, it seems more a discussion of large-screen versus small-screen rather than mobile or desktop.

Chris: You guys talk about the Job-to-be-Done framework, and I think that’s incredibly valuable and insightful. When you think about what it is that you’re actually trying to do, you have a suite of tools that may manifest in different screen sizes that are actually more productive in those different contexts. Just because we’re leaving, in some respects, the mouse and keyboard era of computing, it’s not gone forever. There are just other modes that enter our non-computing moments, and we expect or want to be able to have those capabilities or parts of those capabilities throughout our lives. We used to have desktop machines that were stationary. You leave work and then you leave your computer there. Then we had laptops, and you have your work on the go or on an airplane. Then we put it into the smaller form factor, and it fits in your pocket. Now you’re walking down the street and you can be distracted and walk into people while also getting your things done.

The idea of trying to split the difference between web and desktop or mobile is becoming less relevant and it’s more a cloud of services or a cloud of brands. I was thinking about this the other day – our relationships with brands are going to change because the way brands enter our lives is going to be very different. If brands essentially become like super powers that you can turn on and turn off in your life at different moments based on the life state you’re in, that changes the dynamic. Say there’s an Amazon bot, and I’m having a conversation with a friend and we’re planning a birthday party. If I want to bring in the Amazon bot to suggest gift ideas – for example, we’re trying to come up with what our friend would want – that bot is very relevant for the moment in time. But then I want to turn off that super power. I think that’s interesting.

Going back to this question about discovery, it occurred to me that the way in which you can share contact information – I’m an iPhone user, so I can’t speak for Android – but the fact that you can send a contact from one person to another is essentially the act of installing a bot, if it runs over SMS. Drinkeasy, for example, is an SMS-based service to order booze. I’m like, “Hey, you should check out Drinkeasy.” You’re asking, “How do I get the app?”

I say, “There is no app. Let me send you the contact,” which has the name Drinkeasy and some short code. Just text it and you’ll be up and running. You’re like, “Oh my god, that’s it?”

You just add it to your address book and you’re done. That changes the whole dynamic. If you’re trying to think about whether it should be web or mobile or desktop or bot or chat or whatever, go back to that framework of what’s the job to be done, and then what are the ways or the moments or the context in which people want to interact with the brand or service. When is it most relevant? Really nail those and then maybe extend from there. You don’t want to over-promise and say, “Well, we’re on all platforms at all times” if you’re really not.

The New Developer Experience

Adam: This is fascinating as a consumer, but it must be equally if not more exciting on the developer side. It sounds like there is this fresh, clean palate to work with. What do you find most interesting in this space from a build perspective?

Chris: My job at Uber is a developer-advocate-evangelist role, but there’s also product and marketing. I wrote about this full-stack employee concept a little while ago, which was a self-description of stuff that I don’t do super well but have to know about. This idea of what a developer is has actually morphed and changed significantly but we haven’t quite rethought or reinvented the notion of the role. One of the things I’m hoping to do at Uber is to question that.

I don’t know if there is a new term that needs to be used, but I’m really interested in appealing to the developer-designer hybrid, the one that integrates the left and right brain, the logical and the artistic with a high EQ. In other words, a high social intelligence that has a lot of empathy for people and wants to create great services and great experiences that fit into people’s lives and meet people where they are. I think the Amazon Echo is a great example of this, where you can just speak to it. Granted, it’s version 1.0 in a lot of ways, but you can have a natural and conversational interaction with it.

To come back and try to answer your question, there’s probably two things that I think are super interesting. One is this question of how you design services that fit into a conversational paradigm. I know you guys at Intercom think a lot about this, given the way you’re setting up auto-responders or the ability to talk to your customers. How does that scale?

Adam: Right. How does it scale and how does it stay personal and conversational?

Chris: Exactly. As opposed to the really crappy bots that used to exist. They’re not even really bots from an AI perspective. You show up on a web page, this little pop-up shows with this attractive person, and they’re like, “How can I help you?” And you’re saying, “I’m going to turn around and walk out the door.” But Intercom opens this two-way channel, which is super interesting and it preconditions people to be able to express what they want and not think too hard about it. In this case, at least today, it’s to have humans interpret that request, and then figure out how to solve the problem.

What I’m interested in from Uber’s perspective is what are these developer-designers? Where are those folks? How do we talk to them? And what is the nature of the conversation around those skills so people are thinking differently about what they’re building? Then, from the conversational perspective, how do we design services that fit into that? How do we design APIs and platforms that can be broken up so that every part a conversation can be fed into these micro-services with answers or responses generated, then some aggregate come back together that’s the response to the user.

If I’m trying to book an Uber through Amazon Echo, which we just launched, you want to be able to do a bunch of other things that it can’t do yet. You want to be able to say, “Alexa, book me an Uber and also get me restaurant reservations,” all in one sentence. These things, right now, have to be done sequentially, because each of these is a different API call to a different service. But people shouldn’t have to think about those things. They shouldn’t think, “Oh, it’s OpenTable for reservations,” or it’s Caviar or Reserve.

You have to think about the brand first, and then think about what you want to say to it, as opposed to, I have this desire or intention, help facilitate me getting the things that I’m looking for.

Adam:Right now it’s not, “I need a Mission Chinese reservation at 8:30 and I need a car to pick me up at 7:45 in order to get there on time.” It’s very disjointed.

Chris: Absolutely.

The Question of Transparency

Adam: As people begin getting more comfortable talking to bots, is it going to be very important they know when they’re talking to a bot as opposed to a human? Or will people become more comfortable with those lines being blurred?

Chris: I don’t know that I have the right answer yet. What I will say is that the American marketplace tends to prefer human-to-human interactions or interactions that feel more human. Whereas in other cultures, there may be a preference for dealing with a robotic or bot-like system that’s very procedural. It’s interesting. The thing that I haven’t quite figured out yet is whether or not the rules of privacy apply equally to bots as to humans. If I’m talking to a robot that’s in healthcare, and I’m asking it questions about some lesion that I’ve discovered on my body, I may feel less embarrassed than if I’m talking to a nurse practitioner, who’s like “Oh …”

It’s a weird thing, this anticipation of judgement that we presume in humans but we don’t presume in bots. There’s this willingness to engage with Google Search, and to say all of those things that you would never say to a person, just go see what comes up. But if you were actually talking to people, you’d completely self-censor. What if it was actually a human pretending to be a bot? Would they feel more at liberty to share more things?

Adam: Would there be less of an emotional attachment? Would you even consider that kind of thing?

Chris: Exactly. I think there’s going to be this uncanny valley that starts to shrink, in a sense, where you won’t necessarily be able to tell all the time. Depending on your context, your mental state and the attention that you have available, you may not even care. For example, when I call Apple it’s like talking to Siri. Eventually, I may end up talking to a human but for the most part this Siri agent does fine. It’s more about how well these systems respond to me and understand or interpret what I’m asking. If a bot or an AI system is beyond adequate and performs sufficiently well, I don’t know that people are going to care that much.

There may be very specific instances – I think this is where the brand question comes up – when the character, the quality, the personality of the agent, system or person that you’re talking to makes a big difference. You may prefer certain brands because of the way they talk to you. That may actually be a really interesting in the aural sense, a new type of branding that emerges. I think Pixar or Disney is well positioned for this because of all the characters they’ve created. That’s another thing from a service design perspective that people should be thinking about, what is the brand resonance they’re creating and how does it come through in the bots and agents they deploy?

Adam: Brands have always had voice and tone but now there’s a character involved, an additional layer of identity.

Chris: There’s consistency, too. It’s one thing to call up Virgin America and for them to play house music and for them to sound somewhat British, like the embodiment of Richard Branson. The thing that’s going to be new and really interesting is personalization.

With the move to the conversational paradigm for interfaces, we start to live in threads. I may call Virgin America today, have a conversation with them, and tell them that I’m vegetarian and that I like window seats. Two weeks later I call back and say, “I’d like to book a flight to wherever,” because I’m having a conversational exchange, or I’m messaging them on Facebook Messenger and say “Hey, can you book a flight for me?” They’re like, “Great. We’ve booked the flight, it’s vegetarian, it’s an aisle seat,” and it’s done. Being able to move between channels and persist my preferences is going to become incredibly important in differentiating.

Again, the privacy questions are super interesting and completely unanswered, but people are going to increasingly desire that, because that’s what we desire of our friends. Every time I talk to a friend of mine, I’m not telling them my life story over and over again; they actually have persistence of memory. Why don’t brands have the same persistence of memory so they can serve me better when I’m interacting with them?

Adam: That point of consistency seems like one of the biggest hurdles. Ultimately if the idea is a seamless transition from thing to thing and then you hit a wall, frustration re-emerges.

Chris: And how do organizations organize themselves around that? Today you call a bank and talk to four or five people, you’ve had to verify your last name and your social security number and your dog’s name and you’re like, “Seriously? This is one organization.” Or you have a perception that it’s one organization. Of course, maybe they’re funneling you to different companies in the Barbados.

Adam: Who knows how many bureaucratic levels are within a bank system.

Chris: And how many layers of control for sharing information are in place to prevent certain types of abuses? That’s really user-hostile. Of course the premise is to protect people’s data and information, but the quality of the experience is sufficiently eroded. There’s going to be a real interesting opportunity for companies and brands to think about how they address the sharing and spread of information internally in a way that also respects people’s wishes and desires from a privacy perspective.

Adam: Right. As a consumer, just because of all those levels and firewalls, that doesn’t change my expectations. At the same time, a bank or a doctor’s office is an example of something that’s can be serious and not typically conversational. How does that tone adjust to this medium?

Chris: I don’t want to speak authoritatively about the healthcare business because I don’t know anything about it, except to say there is a lot of seriousness about it. If we become more willing to engage in an ongoing, continuous, partial attention way, that may make those moments a little less scary because you actually have an ongoing thread that you’ve established.

The conversations that I’ve had with One Medical, which is a fairly progressive healthcare provider here in the Bay Area, have actually been very positive, very warm and friendly. It would be amazing to scale that to all types of services. Just because healthcare happens to be, or is conventionally thought of as serious, it doesn’t need to be.

The government could actually be extremely friendly and cordial. It’s not for all sorts of reasons, but why does it have to be that way? I’ve been thinking a lot about the time we have and the way that we fill our time. When there are those moments that are just grading or ugly, why can’t we help to make the quality of those experiences somewhat better? This conversational thing might force that to happen in some ways because you’re going to choose and select for those experiences which are better, nicer, easier, more familiar, more caring, more comfortable, and more human.

Adam: I was joking with a colleague of mine the other day about what the new Turing test could look like. Maybe it’s emoji. Maybe that’s a small touch to lighten the mood for something that’s typically serious.

One problem with #ConvComm: you’re never fully sure if you’re talking to a human or a bot. ? pic.twitter.com/rynn54VcoS— Chris Messina ✁_____ (@chrismessina) January 26, 2016

Chris: It’s funny that you say that because that is literally what happened with Drinkeasy. It’s a service that you interact with over SMS. A lot of these services have been sent to me recently, so I’m trying them out. In this case, Drinkeasy recommended this $150 bottle of scotch. I thought, “Okay, is this really a human on the other side or is this like a bot?” Then the response was something along the lines of, “Could a bot make an emoji fart joke?” And of course, it had a horse farting with the emojis and I was like, “Okay, touche. You got me.”

It could have been a bot punking me; I have no idea.

The hashtag: From blog post to mainstream

Adam: Many of our listeners know you for creating the original proposal for the hashtag on your old Factory Joe blog. Did you ever worry that “Inventor of the hashtag” was going to follow you indefinitely and become your epitaph?

Chris: It’s so funny, for a long time I actually resisted taking ownership of the hashtag. It was probably 2011 when I went to Google and finally did an interview with the New York Times about the hashtag that I thought, “Okay, I have to be part of this story or the narrative is going to go sideways.” Furthermore, there were companies, Twitter among them, that wanted to own the hashtag and prevent other companies from using it through trademarks and things like that. It’s important to think back, even the stuff I’m talking about with Conversational Commerce comes from an era where there was a shift happening. My hashtag proposal was in 2007, the same year that the iPhone came out. Think about that, nine years ago. It took nine years to get the hashtag into mainstream adoption and usage, and it was free. There were no permissions; it was literally a public domain idea.

At a certain point, it seemed to me there was this dawning awareness money was going to be made from this behavior that had been taken from the land of IRC. I’m very clear this was inspired by things that were already happening and reappropriated it into the Twitterverse, but it needed to have someone there to explain where it came from and explain that part of internet culture and that part of the ideology, which lead to the growth, adoption, and spread of this idea and practice.

how do you feel about using # (pound) for groups. As in #barcamp [msg]?— Chris Messina ✁_____ (@chrismessina) August 23, 2007

It’s so interesting now to see the rise of Slack, which is basically IRC culture, built into a web and mobile product. It reminds me that a lot of the things us original geeks have been doing for a long time will eventually become popular behaviors once the product or the product market fit happens, or once the need and culture become so great that there’s no other alternative. The hashtag was my moment of recognizing that mobile was going to be a really big, important thing.

Go back to the original Slashdot thread about the iPod and you’re like, “Oh, I’m never going to need the iPod, it’s a stupid thing. What is this click wheel nonsense?” You see Steve Ballmer talk about the iPhone, how it’s a toy and a software-based keyboard, “No one’s ever going to want that; they’re not going to spend $600 for a phone.” Wow.

I have to maintain a very plastic idea about what reality is and how reality is going to reshape itself. The hashtag was my gradual acceptance of, “Okay, yes, I brought this thing into being, there was a collective moment where there were a lot of people thinking about this, but I wrote it down, started popularizing it, and helped to promote it.”

It also helps to put my life and the lives of other people who have contributed to internet culture in perspective. It’s a choice. If you want to give away your work, you can and it may end up actually having a material impact on the world. That was part of the story I wanted to be able to promote as the hashtag rose.

I don’t think it’ll be nine years for the conversational thing to happen. Obviously it’s been happening and people are building on it. So there’s a bit of Moore’s Law happening with these technological advances. They’re happening at a faster pace now.

Adam: And here you are tracking Conversational Commerce with a hashtag.

Chris: Yeah, I’ve been using #convcomm, which is funny because it could either go for Conversational Commerce or conversational commands. I’m really bad at branding. Stowe Boyd was the guy who actually called them hashtags. I called them Tag Channels. He won on that regard but the idea stood. I think in a similar way, it’s probably the conversational word that’s going to stick with the Conversational Commerce concept.

When I was talking about Conversational Commerce, I was trying to emphasize the shift from conversation messaging, which is for friends and for people that you know, to becoming more about brands and business and transacting in that context. In three to four years, or even by the end of this year, it’s not going to be weird, but prior to this past year it was weird to think about texting a brand and getting an actual reasonable response. That’s where this term, Conversational Commerce, springs from.

Uber: New Roads Ahead

Adam: It certainly seems like, with your new venture at Uber, this is going to be a key component of your job. What do you hope to achieve at Uber and what do you see as the biggest opportunities there?

Chris: That’s such a big question. I joined Uber in January, just over a month ago, because I feel like it’s at the beginning of this shift to bring the internet into reality. A lot of things are doing this, whether it’s Internet of Things or the Echo, which is a good example of hardware that’s bringing the internet into people’s everyday lives. The thing that’s so interesting to me about Uber, and about the platform specifically, is it’s changing some fundamental assumptions about the shape and design of life, particularly in cities. We have a conventional relationship to transportation as a way of getting from point A to point B. It’s perfunctory and it’s necessary and we don’t necessarily look forward to it.

Maybe there are some people that look forward to their commute, sitting in traffic and catching up on NPR or a podcast. But going back to that point about all the moments we have on a day-to-day basis that are just not great and could be optimized or made better, transportation is one of those things. It sucks up an enormous amount of people’s time and energy.

I was thinking about the time I spent commuting back and forth to Google for three and a half years. I commuted back and forth an hour and a half each way, and I did this, let’s say, 300 days out of the year. It’s a lot. To make this easy, let’s say I made $100 an hour. I ended up doing the math and I would have had another $315,000 worth of value over the course of three years. What could I do with $315,000? Well, that’s an interesting question.

Now, if you’re Google and you multiply that by the number of people who ride on a single Google shuttle, let’s say it’s about 40 people, that comes to around $12 million. Now imagine you multiply that by the number of shuttles Google operates. Let’s say it’s like 150 to 200. Now we’re talking like a couple billion dollars over the course of three years. Billions of dollars of valuable time spent in traffic.

If each one of those people was a driver and they were in traffic, think about all that cognitive excess capacity that could have been spent enriching yourself. Of course, Google does that because of the shuttle system and so on, but why can’t everyone have that? Why can’t all the people who don’t work at Google have access to the same time resource that’s currently being utilized to focus on tailights…taillights. That’s the thing that we know the most about, is sitting and looking at taillights. Uber is changing that paradigm and changing that relationship to transportation and finding more ways of enriching the moments that we have on a global scale. It’s going to be totally transformational.

Adam: As someone who’s had over an hour commute in the past, during which I’d become intimately familiar with the glare of taillights, that’s super exciting stuff. Chris, thanks so much for joining us today. This was a fun conversation.

Chris: Yeah, thanks so much for having me.