a16z Podcast - a16z Podcast: Bots and Beyond
Episode Date: April 19, 2016So... about those bots. Bots bots bots. Bots! In this episode of the botcast, a16z partners Benedict Evans and Connie Chan -- along with Chris Messina, longtime advocate of the open web and more -- pu...ll apart various threads related to the topic of bots, mobile, and beyond: the (evolution?) from web to apps to messaging to bots; chat as an interface; “conversational commerce”; and so on. They also discuss why framing messaging through the lens of WeChat both reveals useful things (what works/ might not work) and not-so-useful things (such as seeking the “WeChat of the West”). More importantly, how can we keep bots, and what they represent, in perspective -- beyond the fad? Especially when it comes to considering innovation on the 'web' vs. 'mobile' (remember the mobile browser!) and when it comes to removing friction (vs. adding limited interaction). What contexts (like customer service) are most useful for thinking about bots? And how can we even know, given it’s early days yet and we haven’t moved much beyond the command-line interface... Finally, as computing moves outside the classic work-centric paradigm to room- and urban-scale, how do we think about integrating online and offline, ubiquitous and “diffuse” computing through bots? The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation. This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments and certain publicly traded cryptocurrencies/ digital assets for which the issuer has not provided permission for a16z to disclose publicly) is available at https://a16z.com/investments/. Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.
Transcript
Discussion (0)
The content here is for informational purposes only, should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund. For more details, please see A16Z.com slash disclosures.
Hi, everyone, welcome to the A16s podcast. Today we're talking about guest bots. And to cover that topic in all its wonderful glory we have with us Benedict Evans, our in-house analyst and expert on all.
all things mobile and beyond. We have Connie Chan, our partner who covers all things China,
including messaging and payments and various threads related to how these phenomena are playing
out internationally, particularly in China. And we also have, as a guest, Chris Messina,
who is previously at Google. He's currently at Uber and developer relations. He's widely credited
with proposing the first use of hashtag on Twitter. So you can blame him for that whole phenomena
or bank him for it. And more saliently for this conversation, he's been a long time advocate of the
open web and was an early proponent of Oath. This conversation was recorded right after Facebook's
F8 event last week. So it focuses on the news that came out around Messenger and bots, a platform
for bots as a lens to have a broader conversation about many different threads around bots,
everything from chat as an interface, conversational commerce, just the topic of what comes next
for messaging and mobile, and especially as computing moves into our lives in more diffuse ways.
So let's get started. Today we want to talk about bots.
in the context of yesterday's Facebook news
and I think we should take a step back
and talk about what that news is
but also more broadly
like why are bots in the conversation
and also just this I mean
do you guys even agree with this notion
that there's like it was web
and now it's apps and now it's bots
and by the way we should probably start
by talking about what the hell of bot is in the first place
so yeah there's different ways
you can come at this question
so what Facebook has announced
is a platform
that lets a brand talk to you inside
Facebook Messenger
it sends a message, you reply, it can see the message and produce a response
and the messages from the brand are a bit more sophisticated than just a bit of text
in that they can be a card, they can have a couple of calls to action,
they can put three or four cards up that you can supply it back and forth,
which means that you don't have to know what you can ask in the way that you would otherwise.
You're not faced with just kind of a blank command line
the way you would with like an SMS service that you might have been using 15 or 20 years ago.
You get a little bit more of kind of an interactive model.
And so they've launched, there's a payment more plan in there, there's a few other bits and pieces.
Very oddly, there's no sort of social or viral mechanic in it, which we can talk about a bit later.
And this fits into like two or three big industry preoccupations.
Like there is one preoccupation, which is AI and natural language processing and the sense that you can sort of,
so you can talk to the bot and it understands what you said and send something back.
there's a second preoccupation which is finding like another runtime after the web and mobile
apps preferably on mobile in which you can create experiences preferably one that comes with
sign of some kind of user acquisition or engagement model that's better than the app store
and better than page rank because it's a sort of elemental thing that everyone worries about
with apps is well how do we get people to install our app and so the sense if it's well as if it's
just a chatbot inside messenger or we chat or something then you don't have to
to get people to go to the App Store.
And then there's a sort of a third question in here as well, I think,
which is sort of Facebook sitting on other people's platform.
And the smartphone itself is kind of is a platform for the internet in a way that Windows was not.
You targeted the browser, you didn't target Windows,
but on the smartphone, you target the smartphone operating system.
And it's not neutral.
Apple and Google change stuff, deliberately or not.
And that affects how you can get to your users.
And so Facebook and to some extent Amazon are sort of sitting on top of this and thinking, well, okay, what do we do here?
How can we live in this world?
How can we get to the point that we can carve out some space for itself?
So I think those are the kind of the three kind of mega trends.
There's like there's the AI question.
There's the specifically like the bot and the conversational commerce and the how do you build a new runtime and do you do that in a messaging app.
And then there's Facebook going, oh, crap, we're running on top of Android and iOS.
What do we do about that?
Yeah, let's start off talking about the first two.
So what do we think about the Facebook product itself?
It felt like it's almost like we've been talking about
when is Facebook going to copy WeChat for so long?
It's like it's finally happens.
It's starting. It's starting to look like WeChat official accounts did back in 2013.
Yay.
Connie clarify what those look like for our audience.
So when WeChat first launched official accounts,
a lot of them were more text-based as well with very simple, very simple graphics.
They didn't have the menus at the bottom just.
yet. And a lot of official accounts weren't focused on the WebView concept and the app within
an app. But it quickly evolved into that because the tech space chatbots just didn't work out
and very few of them got much traction at all. Why didn't they work out?
I think it's because in my personal opinion, I think a lot of commerce actually doesn't
need the conversation. And I don't think the dialogue, the back and forth chat is that additive.
I take customer service aside from that because that's an instance where I feel like I want someone listening, I want to pretend that there's someone on the other side, and things like travel agent or personal shopping something when you want personalization.
That aside, though, I think there's a lot of types of commerce where I know exactly what I'm looking for and the back and forth chat actually slows me down.
And I think that's what happened in China.
And actually in China, the app within an app model is much more focused on web views as opposed to bots.
So I actually think bots is an unfortunate name because it kind of insinuates there's this back and forth conversation happening where I think in many transactions, Messenger should be thought of more as a shortcut, a link to a web view that can give a much broader, richer experience.
There was a j-con Twitter sort of, what do we want?
Bots, when do we want it? Sorry, bot didn't understand the question.
And I think there's always that sense of, you know, it's kind of if all you have is a hammer,
everything looks like a nail. It's like everything needs to be a bot now. Well, there's some
interaction patterns where that's going to work really well. Like if you want, if you want,
if you've placed an order online and you want to change the order or you want to see if it's
available and you want to ask like an actual question, there's an actual question you want to
what do you have available is not a question. What are the headlines today is not a
question. What shirts do you sell is actually not a question. What you want is a page that
lists shows you all the headlines or a page that shows you all the shirts. When is my order going
to arrive. That is a question. Can I change it for something else? That's a question. And so to
Connie's point, it's like where you actually would be, it's almost like where you would be
emailing. Yeah. Then it works. Where you would be going to the website and looking for the support
email or the phone number or maybe opening a support chat on the website, then a bot is a much
better way of doing that. Right. And the unfortunate focus on the bot means it's in the chat bubble.
And that's very limiting from a design perspective, right? You lose the freedom of the full screen. You get a
much smaller real estate piece on the screen. And so that's why the transaction ends up taking
longer in many cases. Well, I think it's important to mention, though, that the chat bubble is
relevant in the sense of identity, because part of what Messenger announced is that Facebook has
your identity. They know who you are. And I think that's a key component to this conversation.
Like, how does that fit into the making that transaction so seamless? Because at the end of the day,
a user just wants things to be easier regardless of how we categorize it. Well, there's no reason the
web view can't keep the identity persistent as well, though.
But it doesn't. I think that actually, so the thing that's so interesting about this,
and I'm actually curious to know more about the Internet in China and why, you know, chat would
have been a better platform to launch these services there than, for example, here, where the
web has been very open, it hasn't been censored. And so as a result, you know, both brands,
any commerce companies, as well as advertisers and other folks could build on it in a very
sort of freeway. And as a result, you get a lot of really, you know, 10, 15, 20 years in,
you get a lot of crap. And so people are sort of looking for, I think, a simpler, more universal
experience that's also persistent across devices and does have some sense of who they are.
So the web...
It doesn't require being in the chat.
So I feel like one of the things that would be really interesting, I think, for us to talk about
is actually separating canvases and the way in which you can deliver experiences or
services in these different channels and also look at where the behavior is. So I actually agree with
you that chat isn't the right thing for all things. One of the critiques that I got, of course,
from the piece that I wrote about conversational commerce was, well, which is your critique,
which is, well, you know, it doesn't make sense to put everything in chat because it's
slow, it's stupid, bots are dumb, today at least. And I actually completely agree with that.
Like, my background is in design. So I would not be the first person to, you know, say, let's move
everything that's been done on the web, which is this very immersive environment that's very
interactive, very rich, full photos. It's great for browsing. It's great for seeing things. I spend
way too much money on Amazon because of the richness of that channel. When you go to the
Amazon homepage, it's, and you're logged in, it's a very personalized experience and everything
is on display. If I go into a chat context, that is not the case. I have to poke and I have to
prod, and the things that I can do in that channel aren't necessarily made available to me in a very
obvious way. Now that's changing. If you look at like telegram bots and so on, they advertise
what they can do. They sort of tell you, they're sort of like a help command that will give you
the list of things that they're available for. But visually, I can't just sort of casually browse and
see what's there. So that's actually one of the, I think, benefits and constraints about chat and
messaging as a canvas. But that also suggests some of the limitations that you will have to hand
off to a web view or to video or to voice or to other things. The question, I guess,
is what is how do you sort of express the intention of what you're trying to do and how quickly
can you execute that task there's the sort of several bits going on here one of them is like can you
use your can you use an app to create a better acquisition path and so you know to look at a
completely different model which is by do maps you go into buy do maps and you search for a service
and you can only search for stuff that's local so to speak so you search for restaurants you
wouldn't search for cinemas or you search for taxis or so on and you have the services in
integrated, and then you can go in and tap on them, and you'll get a native interface.
It won't be a chat interface at all, but you're using the map as a channel, as opposed to
just giving an app store and a search box.
And then the same thing in social, it feels like, well, if you're going to put it into social,
you should have some acquisition model in here.
And the weird thing to me about, you know, opening Facebook Messenger is there's just like,
I say, there's four bubbles that they promoted, and then there's a search box.
And I go into the sample, the kind of the content partner, so they've
got, and there's no social model at all there at all. So if I'm an e-commerce vendor and I've
decided I'm going to be on Facebook Messenger there, that doesn't change my acquisition
model at all. There's no, this isn't giving me new ways of getting users. I don't know that
that's, that's entirely true. I mean, for one thing, the thing that is different about searching
in Facebook, unlike Google, for example, is that it's all sort of either filtered through
your friends, so you can see which, if you search for a service and your friends have liked
the thing, you can see which of your friends have liked the thing. And you can also see
on mass, on aggregate, how many thousands or hundreds of thousands or millions of likes a page
has. And so that's a different page rank, if you will, that's more social.
Yeah, but if I've gone in and said, right, I'm going to search for the gap, that's not
the, you've already solved a problem. The question is, how do I get people to know about my
service? And if you're not providing that. And I think the point is that webview commerce,
in many cases, not conversational. It's not necessarily triggered by a chat. There is no chat
functionality in bydo maps. And in 2014, you could already book a cleaner on demand through
bydo maps. You could already hire a driver on demand through bydo maps. You could already buy
a specific seat at a movie theater for a movie in bydo maps. And there's no chat functionality
or whatsoever. And if you think about WeChat, it's not fair to classify it as a chat application.
And it's not fair to draw all the conclusions that WeChat has achieved and say that all messaging
apps can achieve that because we chat is much more an operating system it is much more of a browser
and in a sense any app that has high frequency use case has login and ideally payments data
tied to it could create the same web view commerce ecosystem i don't think it actually i think it's
limiting to say that it's conversational based and that it's chat and messaging base i think that that's
probably fair i think that the challenge for us is like we're so early in this shift and
there is and has been this meme, which I have nothing to do with, and don't think that it's
that useful, of figuring out what the we chat of the West is going to be. And it's actually
not that interesting of a question. It's like, what is actually going to work and what are people
actually going to use? Right. And so I think given the numbers that Facebook announced yesterday
and the growth of Messenger once they... And by the numbers, you mean the fact that the use of
messenger is like so much higher than the peak of SMS, which I think was like 20 billion
messages. Yeah. So the number that they gave was essentially that combined WhatsApp and Facebook
Messenger have I think about
60 billion messages sent
I think monthly
daily
whereas SMS was
yeah yeah so it's basically like
three X well there's kind of a weird
strand in here which of course
half of that is WhatsApp which has no platform at all
at the moment on no platform story at all
and there's a sort of there's a strand in here
somewhere which is that Facebook had the
news feed that was the first Facebook product
and they port the news feed to mobile
and it turns out to work much better on mobile
or even better on mobile than it worked on the desktop
But there's no way to do a Facebook platform inside the newsfeed on mobile.
Or, at any rate, they didn't do a Facebook platform.
Well, that's not entirely true.
I mean, these are all sort of recompositions of different pieces of content from different sources, right?
But it was basically the same experience.
And then, I suppose, but what I was going to get at is that they went from that.
And on the desktop, Messenger was this sort of interesting little add-on that they built.
It was almost like a 20-1-20% project.
And then on mobile, they decide that Messenger, it almost feels like Messenger is going to be the mobile.
native product as opposed to the newsfeed, which is kind of the thing that's ported
or been transferred over from the desktop.
And so, to Connie's point, I'm sort of sitting and thinking, okay, why are you deciding
that if I am a retailer and I want to build a way that people can buy inside Facebook
and that people can share my products within Facebook, why is that happening in Messenger
and not in the newsfeed app?
And why can't I build an app inside the newsfeed app, so to speak, on Facebook?
Where my pages and likes are already stored.
Yeah, exactly.
Why does it have to be conversational?
Or is this going to go down a path that ends up where I am going to get web views inside Messenger?
And I can't stress how important the payments part is in making this whole ecosystem work.
WeChat flourishes because over a third of their users already have their banking credentials tied to their account.
And that's why so much commerce can happen with just one tap.
And so if you think about all the apps out there that already have your payments,
its credentials and a login and high frequency, they actually have a leg up, as opposed to
Facebook Messenger and pulling this off.
But actually, what's going to be very interesting is to see how, one of the big themes that
came out of Mark's keynote, Mr. Zuckerberg, I'm not sure, essentially was talking about
how they want to continue growing Facebook to like the rest of the world.
Like most of the world is not online.
You know, we are actually in the minority, those of us who have high speed bandwidth and so on.
The thing that's really interesting about Messenger is that it's a low bandwidth opportunity for actually reaching people in the developing world.
They've done a lot to really optimize that experience.
They're doing things with instant articles to cut down the loading costs of all these things.
And furthermore, one of the things that they launched that is just mind-blowing to me was the checkout plugins.
So I don't know if you saw that, but like, what they're enabling you to do is to essentially start, to your point, a rich, you know, shopping experience on the web and then tap a book.
button that actually takes you into the messenger context, whether you're on a mobile web view or
desktop web, and then complete the transaction in Messenger. So Messenger as an app and as a platform
becomes your credit card, sort of in a sense, that you're going into that, you know, comfortable,
familiar place to then continue the conversation and the relationship to say, I want to actually
check out and pay inside of the Messenger context. And they beat Apple pay to the web, which I think
is hugely important for the developing world and for people who live in environments that are not
at all like ours. So when I think about the failure of the web as the platform, it's because
it's so heavy. It's because, you know, ad blockers are in ascendancy because the web is so slow
for most of the world that we don't have sensitivity to that. And so when it comes to identity,
persisting across devices with payment information attached, the web is completely failed.
So it's interesting you talk about identity. So this is a great piece about Facebook in
Burma, talking about how, you know, because Burma went from basically zero to 15 million mobile
phones in about two years whatever the number it's something like that i forget the exact number
and they're only 18 hours drive away from chenzhen and so everyone has a smartphone and everyone
has expensive data and everyone has facebook on their smartphone and the way that you get facebook on your
smartphone is you go to the shop and ask for it and the guy takes a printout of 300 login and he picks
the next one on the list and he logs you in and if you accidentally get logged out or you lose your
Facebook app on your phone or you get a new phone you go back to the shop and he gives you
Facebook log in. So the idea that this is
about identity is kind of a
Western assumption that actually
those people in Burma are using Facebook like Twitter
with no connection to their identity at all. They're using it
to see funny pictures of goats and stories about
whatever it is, but not actually
about identity. But there's a small percentage of all internet
transactions that will always be about commerce.
And so you sort of have to have the context
for all the silly, stupid things that people
will do in their normal day-to-day lives
when they're not taking the platform seriously and they're
just using it for, you know,
shits and giggles, that then eventually becomes a platform that becomes more valuable over time.
I mean, the contrast that you might draw then is like with Snapchat or with KIC, for example.
KIC is used much more by teenagers and so on.
And so the question is, could they ever become a more serious platform for payments and transactions
as that teen user base grows and matures, or will they actually see a value in establishing a
Facebook identity that actually is more persistent and more useful?
I mean, they said yesterday that Facebook login is used across 70% of apps out there in the world.
So ultimately, people, at least this is Facebook's messaging, and I'm sure it's fairly accurate, given that I worked on Open ID, people hate usernames and passwords.
So they're looking for a simpler, easier way to get access.
So while I wouldn't dispute the point that you're making about these sort of new environments where Internet access or phones and devices are new, I think with a maturing market that starts to understand the power and utility of Internet access and Internet identity, Facebook is.
actually a pretty good platform to build on for the long term.
So one question I have is, you know, Chris, you were working up to this argument about why
conversational commerce. One question I have is I think, and I think all three of you guys
disagree about this, is where you guys fall in this discussion of web apps and now bots.
Well, I don't think, I mean, there's a pretty clear distinction one has to make here,
which is there is web and apps and then there is something else. And then one of the things that
may be is within a third party app on the device. It's clearly also,
Apple and Google's desire to create something here.
And so there's always a sort of context each year that Facebook says,
hey, this is the future of interaction on mobile.
And like six weeks later, Apple and Google say, sorry, kid, it's like it's our operating system.
We get to decide how that stuff works.
And I think actually what Facebook has done this year is quite sensible in that they're actually
not trying to do stuff that Apple and Google get to do.
They're doing it inside their app.
But then to Connie's point, it's not specific, it's certainly not specific to chat.
And it doesn't even have to be specific to social.
can be anyone where you've got some kind of mechanic for frequent use and some discovery model.
And the obvious alternative to say like, and the question is, like, have you got the frequent
use and the payment? Have you got some kind of interaction model where it makes sense to add
more stuff to it? And have you got some kind of discovery path? And so the reason I mentioned
Badiou Maps and Connie wrote a great blog post about Badium Maps is like there you're not
using social, you're using location as the kind of filter for how do you discover stuff. And the concern I
have looking at
now then Facebook
they could have done
it inside the news feed
where we wouldn't be talking
about bots at all
instead they chose to do it
inside Messenger
where we are talking about bots
it's like problem
the app store has two million things in it
solution create an app store
inside messenger
problem the messenger app store
has two million things inside it
like if you haven't actually done
something you have to do some other thing
all you've done is just move the problem
and say that's why I think
bidea maps is interesting
because you are actually solving a problem
because you're just constraining it
to local services, and you will not go into Bidu maps and search for stuff that isn't local.
I mean, you might do, but you wouldn't get a result.
Whereas the problem with the search box in Messenger is the same as the problem with the
search box in the App Store, or the problem with the Google Searchbox, which you can search
for absolutely anything, and it will give you some kind of answer.
And this is my point.
What I found really surprising about the Messenger implementation is there's no social there
at all.
There's no acquisition model here.
It's just a new channel.
I think the challenge for me, right, and it's so interesting because, like,
there's so many threads and so many, you know, edges to sharpen here because I agree with
many, like many of the things you said, but I also disagree with a bunch of other things.
So on the point of discovery, what I think will be interesting, and we really have no way of
knowing how this is going to happen. But my feeling in my sense is that there's a lot of stuff
that will happen probably in the next five to ten years in terms of, you know, you talk about
a map, and a really good sort of immersive map is just reality. And so if Facebook is
actually able to provide a way for smaller merchants in smaller environments, local towns,
places like that to actually help their customers get in touch with them, not by using the phone,
not by going through AT&T, not by going through Verizon or whatever, but simply going through
Messenger and through the Facebook page that they already have, that is actually a huge win for
them.
If you are a business trying to fear out which platform to build on, and all you hear are about apps
in the app store, and you need an app to go out there, right?
I've consulted with a bunch of nonprofits who are like, oh, we have to build an app,
like we've got dollars coming in saying, you must have an app, build an app.
I'm like, how are you going to get discovery?
No one's going to use your app.
right international players have the same problem all the time too
yeah in contrast you could build something
whether it's a bot or not whether it's just being in messenger
that allows you to actually be interactive and see what happens at a very low cost
and furthermore with a bunch of the things that facebook is rolling out
which we probably take for granted and think are sort of silly
there is social discovery so these can be the messenger codes
or they can be the new URLs that they've thrown out
and as well they're actually focusing a lot more on group dynamics
and i don't know a lot about the stuff that they're planning
but they're looking at groups in Messenger as being a big way of solving the distribution problem.
So the problem that Apple has that has no social network, the problem that Google has is as no social network.
So there's no place for people to sort of congregate and share information with smaller communities of people where this stuff actually spreads.
And I'm thinking of international markets, not really in American markets, where the media is a big part of it.
And there's hype cycles and stuff like that where people are actually going to be sharing the names of bots or sharing the names of services and sort of sharing stuff in that way.
and that it's going to be a more natural way for people to discover these things and for these services to grow.
It's going to be a lot more organic and it's going to take time.
But I think it's a better answer and a better solution than go hire some consulting agency to build you some full bespoke app that people will download once and forget about.
But I think the whole problem goes back to the distribution point that Benedict is bringing out.
Because if you're focusing on onboarding, it's very unclear to me in that nonprofit example that building a bot is easier than building a web page and then creating a bot that's simply just a link to the,
web page. Both are hard. Right. But the point is one will work on all kinds of platforms and in browsers,
and one's only going to work in Facebook Messenger. And actually, that's, I think, something that's
going to hold the onboarding of bots even onto this new platform. Websites don't already have
this programming capability in-house, whereas if you focus much more on web views and use the bot
as more just a shortcut to go into the full-on web view, you can onboard accounts much, much
faster. Can we take a step back, though, and talk why people even care about bots in this
paradigm? Because clearly there is a need to have conversation in a way that isn't populating your
inbox with all this extra stuff. I mean, regardless of how it's, I mean, it's not that it's not
irrelevant, the design and the scoping of it. But why do we want this? Because I don't care
about the social sharing. Like, I want a frictionless interaction where when I book up a ticket and my,
I don't have to go back into my inbox to find the confirmation. If I want to change my fly,
that I can just very easily, like, just go right into and say, like, hey, this is my new time.
I think that's exactly.
I mean, there's different priorities conflicting here, because there is kind of from the user point of view, that customer service example of being able to message somebody on Facebook Messenger and there being a platform where they can do a bit more than just send you a text reply is fantastic.
Then the idea of kind of coming from the other end and saying, okay, instead of going to a vendor's website or.
installing their app, I'm going to chat with them.
That feels like a kind of a broken experience.
And then there's a third question which is like,
we want to have an engagement with this person.
Getting them to install an app is really hard and it's really binary.
Is there some other way we can create like a rich interaction experience
without forcing them to go through the app,
which is what the web gave you on the desktop.
It still, frankly, gives you your mobile,
which I thought you do have a web browser on your smartphone.
It's still there.
They actually probably use more of that
There's that S-C-CD comic
where one person's basically saying
Oh, what if you invent a phone
Where you don't have to download apps
And you can just
You can just access everything seamlessly
And then he said
He thought he was a genius
Until he realized
It was just web pages
Well this is the thing
But this is why the kind of the web view
Your kind of web view point is interesting
Because then it's like
It's another way to get people to a website
Is it a website?
is an app, I mean, that's a whole other conversation
and it's like, I don't really care what programming
language it is. It's just not, that's not something
that matters to a user. But it's a way
of delivering that, getting that rich experience
and this is kind of the, comes back to the kind of
primary point, the desire for a third
runtime. How can you give someone
the richest experience you would want
that doesn't have these weird constraints
of like you've got to get the app installed
that doesn't have the weird
constraint of it's a chat bot and it can
only do certain things? How can you
get people, give that people that
experience without having to get people an app installed.
Well, but I think that you're solving this from the wrong perspective, which is, on the flip
side of it, from the user's perspective, how do I get the most possible power and benefits
with learning as little as possible and doing as little work as I need to?
Exactly.
I want to think about what the user wants here.
And is that Google? Is that Facebook? Is that the app store?
I don't think users care, right? And so the question then is who gets out ahead?
Who builds a platform that is just rich enough and enticing enough and has
enough distribution, which is the entire
sort of marketing pitch that Facebook provided
yesterday is like we have all the users come to us
relative to anybody else
who's thinking concretely about
building relationships. So the thing that I actually
I think is really broken, I think is worth bringing
up is the way in which we
actually have articulated our brand
relationships is pretty awful.
I mean, if I were to like
look at conversion of
friends that I've made, like
what is my conversion like
figure, you know, 0.2%.
Every time I try to make a friend, I actually get a friend of them reject you before they actually become your friend.
Some of them don't even respond to saying hello, right?
They just walk by me.
I mean, this is email marketing essentially, right?
So this is the problem with these rich web views that you're talking about.
I mean, this is, again, going back to the eye blocker thing, like every time I try to read an article, you know, my friend, you know, Forbes or whatever has like 15 other friends coming in that are trying to sell me things before I can even talk to them.
So one of the nice things and one of the dynamics that's built into Messenger and chat apps that I think is so powerful that's returning.
some power and control to the person is just the simple block action that you can take,
that you can't take, I mean, I guess ad block is that thing.
It's sort of like saying, I don't want to hear from all these other brands and companies
and services on the web.
But any platform that enables those web views could block it too, right?
And WeChat, for example, there's all kinds of rules around how often an official account
can even message you from it.
I think actually I would describe the thing your characterization is exactly wrong, because
I can choose which websites to go to or not go to, but I can't.
can't choose who sends me Facebook messages or Twitter messages unless I block the more I turn on the privacy things.
And so, those are push channels, whereas the web is a pull channel.
That's actually not true in terms of Facebook Messenger.
Facebook Messenger explicitly said yesterday that the users have to first do some action to initiate that relationship.
It's a push channel, not a pull channel.
So it's not that I'm stopping something.
But that's also, that's a very clear sort of intent that opens up that channel.
I mean, I think the control thing here is actually there's a fundamentally important point here,
which is, you know, there's a great chart from BuzzFeed
and information from BuzzFeed on where people watch content on BuzzFeed.
And, you know, it's one quarter is direct,
and then it's YouTube and Facebook video and Facebook images and Snapchat.
And you look at those things,
and you see different content formats and different acquisition models.
And now we've got another content format
and another acquisition model where the dynamics are a bit different.
And it's not really a media format in the way that those are,
but it's another format.
And when you look at all of these,
you see this sort of splintering of,
how you get users, what content you create,
what the interaction model would look like
if there is interaction models.
And all of them are controlled by people.
And I think, you know,
you can look at Snapchat video
is exactly the same as Instant Stories
or Google AMP in a funny kind of way,
even though one is that the bandwidth consumption
is very different.
But they're about somebody with a closed platform
creating new content formats
and telling people,
if you want to be in this format,
in this platform, with this distribution,
you have to create new types of content,
think about users in new ways
in order to get access to those
at least when it comes to Snapchat
and I think this is relevant
given what I'm trying to sort of anticipate
is the next wave of people using the internet
and what their expectations and norms will be
and if you grow up
using messaging as your first sort of application
of the internet then you sort of expect
a kind of dynamic interactivity
that you don't get from content
that is created that is static right
this is I think the key point
if you look at Snapchat
in the, one of the, at least my intuitive sense for why Snapchat is compelling, especially
for young people and for friends, is that it's very hard to, to fake it. It's much more authentic
content and you're going for it because you can't sort of compose, you know, this immaculate
blog post and put it out there and it's very, you know, nice and pretty. It's more sort of
rough and raw. And so messaging sort of also does that. And it's, and it's, if you contrast that
with a lot of the marketing and advertising, and again, you go to like a lot of these websites,
Like, one of the reasons why AMP and Instant Articles is interesting is because it sort of cuts all that crap off.
And it says, look, you're here for the content.
You're here for the raw stuff.
You're here for the good stuff.
And that's what you're going to stick around for.
And the more that companies have abused customers and sort of taken advantage of their attention, I think people are moving away from that.
And so that's sort of like the shift that I see.
And Messenger is such a narrow channel.
And it's so restrictive.
It's actually not rich.
And that's actually a feature.
And so that actually opens a question.
Which is the point.
It's moving back in the direction of the consumer
to actually give them back some of their attention
because they're overwhelmed by all this crap that we've created.
The control is moved from the publisher to the platform.
It's not about the user's control,
it's about the publisher's control.
It's the experience.
And it's like, I guess what I'm saying is that
even with all the ads on Facebook and so on,
Messenger is such a narrow channel
where there is no sponsorship and there is no ads.
Totally yet.
Successively, though, you see generations of people using social media
abandon the previous generation once they move in a monon.
mode because they start to abuse the users.
I mean, the places in which Twitter is putting ads, it's just, it's almost sort of
uncomfortable.
You sort of go to like, you know, what used to be a single tweet page, and now you've got
like a sponsored tweet below it, and you've got all this stuff distracting you all
the time, you know?
So where does that balance of, you know, I guess, relationship between these brands and
the consumers and the people using these platforms sort of fit in?
At the same time, commerce is 100% reliant on marketing, right?
Commerce is built on marketing.
And so if you completely strip down the way that you can market a new product to just text in a one image, it's very limiting for these commerce companies.
Well, I feel like we're also talking about this in isolation because it's not like these other elements, the web and everything else go away.
It's just like this is now adding a rich, exactly.
It's sort of adding a richer way of now interacting with an existing brand.
Like when I book my initial plane ticket, I'm probably going to go to a travel website.
And then when I have the rest of those interactions about updating my travel and doing all the reservations, I'll do all that within a bot.
So I think there's another strand here to talk about maybe, which is that Wii chat has achieved the point that its success is circular, that you have to be on WeChat because everyone's on WeChat, everyone's on Wechat, everyone's on Wechat, all brands are on it.
And Facebook clearly isn't in that position at all, certainly not on Messenger.
And it's an open question as to whether it will be able to get that.
but that's why the onboarding matters
and that's why web visas is actually a much faster way of onboarding
it is but I said the thing I was going to say is
and there's an interesting kind of question here and this comes back to
I mean as again I mean I just a degree radically with the idea
that this is shifting power to the user which shifting power to the platform
the thing that it reminds me of is I mean
so there's sort of two metaphors here
one is that what Zuckerberg is extremely good at his surfing user behavior
and that if anything is unbalancing the surfboard
he pushes them off and so this in turn reminds
reminds me of the line in the godfather where Vito Corleone who tells Saloso, I'm sorry, I can't tell my judges, my tame judges and policemen to protect you because they will protect a gambling racket or prostitution racket. They will not protect a heroin smuggling racket. And if they did, they would be in trouble and they would get fired. And I think this is kind of the point that, you know, it's not in Facebook's gift to say we will be a good partner to you because it's actually not their decision. It's the user's decision. And so whenever they've done, you know, the things that they've killed, they
because the users didn't like it.
But, you know, it's so funny, because you just said that, like, all the power
exists on the platform, and yet Facebook is beholden to user behavior.
So what we're talking about...
It's not directly a decision by the user.
Facebook can decide to put this stuff there.
Facebook can also decide to remove it.
Facebook can decide what you're allowed to do and what you're not allowed to do.
But I think you raise a really interesting point, and I think that there was an article
this morning that I haven't read yet, but I saw the title, which is, you know, provocative
enough, which is like...
Well, that's all anyone does anyway.
Exactly.
Right.
Who is time to read anymore?
You don't think I've read all those stories in my newsletters,
about, you know, CEOs being like the new statesmen or people or whatever.
And whether you believe that or not, I mean, what you're talking about is actually the governance of these platforms
and the balancing of the interests of all these different parties.
So, in a sense, Facebook is the largest global political party that exists or political, I don't even know, like system that exists.
It actually is balancing lots of different interests.
By the way, I mean, just to go.
back to like my whole sort of the genesis of the term conversational commerce was merely to point
out that in software that has been mostly dedicated to conversations between friends and family
members for however many years, we're now starting to see in the West commercial entities
show up, non-human entities show up. So it isn't to me that it's about conversational commerce
forever. It's about conversational software as being a new paradigm for releasing and launching
services, features and tools in a way that previously people didn't have that expectation.
And so that's the shift that I see in terms of this new runtime. And those conversations will
take many forms, whether they're in the form of chat and, you know, with chat bubbles and the
conventional kind of like back and forth, whether they'll be like Alexa or Cortana or Siri or other
voice. Voice as distinct from the tech spots. It's that interactivity that I think is interesting
and the dynamicness that's presented in that relative to the static content that's largely defined the web,
where there is a piece of content on the server, you request it, it comes down to the browser,
the browser interprets it, it shows it to the user, and that's basically what you do.
And then you have a form, and it goes back and forth and back and forth, and back and forth.
Now we're stripping away a lot of the decorations on that content,
and we're making it just pure conversation in a way that anybody who speaks any language can participate.
I guess what I'm saying, though, is one, if that's the definition,
definition of conversational commerce. That is not new because it's existed in Asia for many,
many years. Totally. And that's why it's new here. At least it's new for me.
Another really interesting thing that surprisingly, the media hasn't talked about is how the bots will interact with the offline world. And this is an area that I think WeChat really shines. Because in WeChat QR codes are not WeChat, sorry, in Asia QR codes are much more prevalent. The idea of adding an official account through the
the offline world is very, very common.
And it goes to that whole issue that Benedict mentioned,
which is how do you discover these bots?
How do they get pushed to you?
How do you find them?
And if all you can do is say, well, you put the QR code on a billboard,
then you haven't actually solved the problem at all.
That's no different from putting your domain name on the billboard.
I mean, you actually haven't solved a problem at all.
A tiny bit better.
But yeah, you know, saying follow our bot,
you know, spending a million dollars on media to say follow our bot
is no different at all from spending a million dollars on media
saying like us on Facebook.
or go to our webpage. You haven't actually solved a problem.
To me, this is also a case where we really don't know how things are going to...
I mean, this is your point, where how things are going to evolve,
because when I can imagine a bot that is as good as my relationship with Alexa,
in the sense that I can... It has a personality and something that not...
Like Tay, just something that I really just want to interact with as my go-to place,
we can't even conceive of that yet, because it's so raw and command line style still
that that could be a potential future.
I mean, where I would actually follow a bot, if Connie, like, Connie, if you said, like, hey, there's a store and it has this great bot, and this bot is actually a fashion assistant. And it's like on a bot, I might follow that bot if it's named Anna and click on it and create this personality of like these things that I'm following. It's not just this transaction at that point.
I think it's funny, too, that you use the verb like follow, right? I think that the thing that's different and the dynamic that actually is more exciting and interesting to me about this is that, again, it's actually sort of bringing,
these services and brands and bots or whatever it is into your world in a way that people
who are not super familiar or savvy with technology can understand. If they text their moms or
they share photos after weddings and stuff like that, they can do that with these, whether
their official accounts or bots and businesses, Facebook calls them, whatever it's going to be.
And that that pretext, those set of behaviors and the fact that we have fairly ubiquitous,
you know, mobile networking now means that people can text these things and message them whenever
they want to. When they're out of the bar at night and they're drunk and they're like, oh, I need
some new shoes like they can do that that impulse thing is there whereas trying to sort of you know
remember which website to go to and is it mobile optimized and I'm like I'm waiting for it and I've
got like poor reception like there's still realities on the ground that make using the web on mobile
devices and platforms just not that great but you can always just about always message someone
and that I think resilience is I guess one of the things that I feel makes the messaging context
robust. And so from here, because there is that behavior, now you can build on it.
But that resilience, what you just described, like being able to recall the name of the
website, that really is just a shortcut. And so that's why, again, like, I focus a lot less
on the back-and-forth chat dialogue because I don't think conversation is that additive
to most transactions. And I think the fact that it lives in Messenger fine, but it's really
just a shortcut to something else.
Maybe that's it, right? Like, I mean, the funny thing about the way that Threaded experiences
look is that it's actually sort of very similar to the way that RSS readers used to look.
and so you have sort of like your sources on the side
and you've got sort of content on the right.
So I would agree with you.
It's not necessarily just about the conversational nature,
but I think that the conversation is the invitation
that then invites you to actually have a back and forth,
which is about, I sort of started this point before
about if my conversion rate for friends was like 0.2%,
that'd be pretty terrible.
This question that so many marketers and brands and agencies
focus on just the conversion,
on just getting someone into the funnel,
and then they drop them off to whatever sort of other automated system exists,
is part of the way in which it's broken.
If you could imagine actually scaling a high quality, high class, high touch interaction
that feels like the friendships that you've had for years,
because I've been with Virgin America, I don't know, for 10 years or whatever,
since they first opened and, you know, God rest their soul,
like now they're with Alaska.
But like the thing that's interesting is that I actually message them,
this is so great, I sent them a DM on Twitter last night.
The last message that I had sent them was in 2009, and I never got a response.
It's just crickets.
Things have changed.
I got a response within two or three minutes, and I had a problem with my check-in,
so I sort of did this thing.
You're a perfect example.
But this person knows nothing about me.
I've had a relationship with Virgin for all this time.
And to me, that's sort of part of the broken piece of this, that aspirationally we can
move closer to getting to a kind of personalization and a resonance that brands would
love to have.
And I just don't see that that, I mean, it's possible.
And Amazon is a good example of this.
Like, they have done it.
They've sort of maybe not totally cracked them up, but they've certainly opened it on the web.
Most other brands and services can't achieve that level of personalization and customization and persistence.
And so I guess my question and my thought is that, well, if you strip away all the other trappings of pretty things and bells and whistles on the web to just being about a conversational channel, that actually forces you to sort of, you know, get into it and just be very raw and very direct.
And if you're going to be useful and valuable, that's where that's going to happen.
But, like, that example is, again, more customer service-centric, right?
So I think we all agree customer service is a fantastic way to use conversation in back-and-forth dialogue.
But it's really that when I'm going out to buy something, I already know.
So I think browse is, like, really bad.
That's one of the challenges that I think while, I mean, I've sort of always tried to be on the side of individual freedom and choice and all this stuff.
And the reality is these platforms remove so much of the cost and overhead of learning about and maintaining awareness of all the standards.
and all the sort of emerging things
that it actually enables people to conduct business
and transact, even if it's
lower margins because the platform is
taking a cut than if they were to build
their own really, really good implementations
of this stuff. I mean, again, if people can't even
get them to download, if brands
can't even get people to download and use their native
apps as it is. Yeah, the challenge
is that you have that nice, easy,
frictionless experience and you're on Facebook,
and then you wake up one morning and your traffic's
down by three quarters. I totally
hear that. You know, the internet, it's funny to say
the internet was like 20 years of stasis before mobile came along.
But kind of, you know, you have kind of a completely new kind of change in many more
different levels and much more continuous than you had on the desktop web, I think.
And this is just kind of one aspect of that, as we know, this is the saying we look for
another runtime.
It's a lot more complicated than that.
We're looking for kind of new interaction or new engagement or new acquisition at every
level of the stack in a way that we didn't really have with a desktop web.
Yeah, so I think like the fundamental shift that I see, and it's interesting to think about
the last 20, if not the last
60 years of computing
coming out of like the Cold War era
and building and designing
computers that were designed really around human
augmentation, you know, so you know this
coming out of park, and building
systems that were really about information retrieval
designed for experts.
The mobile device is the first
real sort of mass-produced device that
everyone seems to have and everyone seems to want
and that they incorporate into their lives
in these somewhat awkward ways.
We've persisted so many metaphors
from the office environment.
I mean, down to like the desktop,
which was re-labeled the home screen
or the home screen on iOS.
And we have this concept of apps
that represent different verbs.
And then you have to sort of,
from a cognitive perspective,
relate this little app icon
to the verb that you want to take
to then the services that you want to execute.
And you have to have accounts
and identity and payments
and all these things
have to be sort of set up
in each one of these little pockets
of functionality.
that that I think is leading to this sort of just or will lead to this just exhaustion.
So on the one hand, these personal devices I think are sort of leading to a place where people are looking for simplicity.
On the other hand, with ubiquitous networks and now sensors and other things like that sort of entering into the environment,
and the first kind of room computers sort of lighting up and giving us access to computing capabilities just sort of in the ether,
we're seeing diffuse computing kind of come upon us.
And so that relationship and that handoff
between these personal mobile devices
and computing in the real world
is going to become super interesting.
And it's at that kind of inflection point
that we start to see, you know,
what urban computing is going to look like
as opposed to, you know, office work-driven computing.
And I guess that to me is the most rich
and the most interesting.
And it's really bringing,
maybe it's not control back to the individual,
per se, because they're still members
of and beholden to the platform providers.
But it's forcing these brands to actually sort of cut back a lot of that complexity
and focus on delivering very simple, very straightforward services to people at scale.
And so I think that that's why, and because conversation is so universal and so accessible,
that's why I think it's super interesting and super worth thinking about.
And we are only at the very, very, very, very, very beginning stages of this shift.
I think I'd like to basically give caution to some developers, because there are,
is this kind of gold rush to build a bot right now, to really think whether in the bots that
they're creating, that chat is additive and whether chat is actually going to get rid of friction
and really evaluate your service. Are you a customer service-centric product? Is your product
something where personalization is very key to what you're doing? In those instances, sure, a bot, I think
allocating resources to go make a bot might make sense. But if you're not in one of those
categories, really think about whether the chat and the back-and-forth conversation is additive.
And if not, I mean, I don't mean to not build a bot, but think about how to design it, such that it
takes you to the web view much faster.
So you can show them more images, so you can do more richer, interactive experiences.
And spend more time thinking about how will this evolve in the offline world, right?
When Facebook first came out, all of a sudden saw a bunch of billboards that said,
follow us on Facebook.com slash XYZ.
Same thing with Twitter.
You would see the Twitter handle on the bottom of CNN or of all kinds of TV shows, right?
And so how will this new bot ecosystem live in the offline world, I think, is super interesting.
And so thinking about how this whole bot ecosystem will live in the offline world, I hope, will be much more than just follow me at XYZ username.
Okay.
Thank you guys.
Thanks, everyone.