Why conversational commerce may be our best chance to re-imagine the web
From WeChat to Messenger and Slack, our future is rapidly being colonized by social networks that effectively mean to function as operating systems. If this is our future, we will need the web more than ever, but it — and they — will have to evolve.
It’s easy to downplay the furore over Twitter’s proposed changes to its Timeline, but I believe it was caused by a much larger, somewhat existential malaise, that coincides well with the second big piece of social news this year (so far) — the downfall of Facebook’s Free Basics at the hand of Indian regulators.
“In India, given that a majority are yet to be connected to the internet, allowing service providers to define the nature of access would be equivalent of letting TSPs shape the users’ internet experience”
— TRAI ruling (PDF)
The neutrality battle is far from over, but in the end may also barely matter. From WeChat to Facebook and Slack our future is rapidly being colonized by social applications that effectively mean to function as operating systems. Should this come to pass, the impacts will not be trivial, and could fundamentally alter our relationship with the network through which we increasingly live our lives.
Will we be happy to live in a future where social platforms — and the ‘apps’ and content that increasingly live within them — become the primary lens through which we see the world? If not, how could we rethink this future to the benefit of all of us?
The rise of social mega-apps
Use of social platforms is positively booming.
With a user base of 1.59 billion monthly active users, Facebook now reaches 20% of the planet — and at least 80% of those users access their apps on mobile. Add to that the billion users on WhatsApp, and the company begins to feel invincible.
But these admittedly impressive numbers tend to deflect conversation from what may prove to be the real change in how we live and communicate using the internet: the rise of social “mega apps” that offer rich applications and services directly within what purports to be a social wrapper.
In China, almost everything a smartphone user could do with a web site or native app is now accessible inside messaging giant WeChat: study English, book a flight, play videogames, order a cab, send and receive money, watch ‘TV’, buy (physical and virtual) things, consult with a doctor, renew a passport, and so on.
You can, for all intents and purposes, live your entire life within WeChat. It takes a phone full of apps to replicate its entire functionality. It’s simple enough that anyone can use it, versatile enough that everyone has some use for it.
What’s more, once a community reaches critical mass, it hardly seems to matter whether a platform is already exhibiting mega-app behaviour. With a giant audience at their fingertips, people simply cobble together a facsimile of this experience on their own. A merchant in Kuwait for example, sells sheep through Instagram. No shopping cart, you just call him to inquire, negotiate and determine how best to arrange payment.
He’s somewhat in the minority however as more often than not these communications take place using chat.
From WhatsApp to Line, and even Instagram, messaging-driven commerce may not be perfect, but it’s good-enough — even for large brands.
“…they chat all day long on Instagram and WhatsApp with our customers…we literally have some customers who have ordered multiple times and never spoken to us…never placed the order online”
— Christina Hawley, Jumia Africa
Growth of social mega-apps is most pervasive in emerging economies, but their presence hasn’t gone unnoticed in the West. Buy buttons are popping up within Pins and Tweets, you can now call Uber through Facebook, and full-blown digital wallets will no doubt soon appear. It’s hard to tell how these western facsimiles will do however — and this may be down to intent.
“Philosophically, while Facebook and WhatsApp measure growth by the number of daily and monthly active users…WeChat cares more about how relevant it is in addressing the daily, even hourly needs of its users — its goal is to address every aspect of its users’ lives, including non-social ones.”
— Connie Chan, a16z
WeChat may be an app, but it’s behaving like an OS. Systematically building a platform for rich, versatile, and yet fundamentally human-powered interaction.
“WeChat is like a fully fledged cityscape where all the electrical and plumbing have been installed…developers can come in and build all kinds of unique and distinctive real estate that assists people as they go about their daily lives.”
— We chat about WeChat , MLS Group
Underpinning WeChat is the use of a simple Chat interface as universal enabler of conversations…with whom or what increasingly hardly matters: close friends, strangers, ad-hoc groups, your teacher, co-worker or client, a hospital booking bot, (or dreaming forwards a bit…) a bus stop, your smart baggage tag, a building directory, your insulin pump.
“You can send any kind of message (text, image, voice, etc), and they’ll reply, either in an automated fashion or by routing it to a human somewhere. The interface is exactly the same as for chatting with your friends.”
— Dan Grover
This super-versatile interface is often paired with a more traditional set of ‘pages’, complete with top and sub-level navigation to in effect provide the best of both worlds — a self-directed browsing environment that can easily pivot to highly-personalised and targeted conversations.
If WeChat is the most robust of these platforms, its nearest cousins may well be Japan’s Line, Korea’s KakaoTalk, and an unlikely outlier in the form of Waterloo Canada’s Kik. Far from prototypical, these mobile-only platforms collectively reach well over 1 billion monthly users around the world (600M on WeChat alone).
What’s more, social and mobile use are often so intricately linked that ownership of one only drives the other.
Mobile-first = Social-first
When it comes to social app use, research shows a striking difference between consumers in emerging and developed economies.
Internet penetration in developed markets is high, ranging from ~80% in Japan and the U.S. to a whopping 87% in the UK. And yet in many of these countries, social app use is still a relatively low 60–80%. Look to India, Brazil, and Indonesia and you’ll find a very different picture: while a mere 15–20% of people use the internet, a staggering 85–90% of them use social apps.
The reasons for this rapid growth is hard to pin down. A recent IBM study (PDF) cites “the concentration of Generation Y and younger users in those regions and the cultural emphasis on maintaining regular contact with friends and family”. The growing diaspora of economic, social and conflict migration likely also play a part, as does the flexibility and financial savings of using data to broadcast messages one to many.
The biggest single factor may however prove to be the huge growth in availability of low-cost mobile devices. By 2018, Gartner expects the average selling price for a basic and utility phone to be $78 and $25, respectively. (A cursory look around the net suggest we’re already there).
Combine these factors and it’s not surprising that in many emerging markets, a person may enter a phone shop and simply ask for Facebook or WeChat. When a single application appears to enable an entire world of people, places, and things, one can be forgiven for forgetting (or even fully grasping) the presence of its underlying enabler.
“Indonesians surveyed by Galpaya told her that they didn’t use the internet. But in focus groups, they would talk enthusiastically about how much time they spent on Facebook. …“It seemed that in their minds, the Internet did not exist; only Facebook” he concluded.”
Shocking? But then haven’t we been here before? How many of us have known someone who persistently mistook the “Big Blue E” and everything it enabled, for the computer itself? Or believed the browser on a Mac or PC was the sole entry point to all the things that device could do?
From destinations to just-in-time conversations
It’s useful here to pause for a reality-check. Mega-apps notwithstanding, a large part of what we do on social networks is share bits of our lives and our culture.
“…the emerging global society is visual. All these photos and videos are our way of trying to see the world. We feel compelled to make images of it and share them with others as a key part of our efforts to understand the changing world around us and our place within it” — Nicholas Mirzoeff
While many of us simply chit-chat and exchange thoughts with friends, we also share a veritable ocean’s worth of socially, culturally and often temporally significant artefacts (social objects if you will). Objects that both spawn and fuel conversations. Object that (historically at least) originated, and were consumed almost entirely on the web.
The role of the web as connective tissue for social conversation is changing fast however; and for many users, may already be effectively be besides the point.
A (social) feed, then, was a mixture of personal posts and pitches to look at other things. It was an endless stream of captions contextualizing a Great Big Outside. But each year the feeds got richer. The captions expanded into previews. The previews expanded into full photos, videos and posts. The remaining links underneath came to resemble vestigial metadata. — The Awl
The final nail on the coffin may well be that within many social apps, much of the remaining web-hosted content currently opens in an embedded web view. Combine this with the “Facebook is the internet” meme and we come to another potentially shocking realisation: for many users, Facebook may not just be the internet, it’s probably now the web as well.
If it’s not already obvious, the increasingly siloed and pervasive use of social apps (especially those that creep into OS territory) only serves to reduce the value of most other apps.
If you’re an Android user, you may already feel the logic of this. Thanks to a platform messaging architecture that revolves around loading and unloading of intent-driven “fragments” (i.e. a task or view within an app) you’re already used to diving in and out of apps to get stuff done. Round-tripping through chains of short, targeted UIs (with often little though to what app actually spawned them) creates a sense of flow that makes traditional app-to-app multitasking feel kinda quaint.
Apple’s brilliantly designed 3D touch is just one more break in the “manually open a thing each time you want to do a thing” model we’ve been stuck with for some time. But their current implementation is only scratching the surface of what we could do. Initiating key actions without formally opening or deep-diving into an app is great — but what if you could initiate these actions from almost anywhere?
What if an app icon wasn’t merely a pointer, but a social application object. An object that could be shared, discovered and acted upon as part of a larger task?
How would all this look? The narrative is lately converging around some form of conversational-platform-cum-OS. A platform that enables conversations…that go on to prompt actions and experiences…that then prompt more conversations. This seems plausible, although Chat may not always be the best way to initiate, let alone complete an interaction.
It’s likely however that conversations will continue to play a big part — but not the way we’re currently used to. After all, it may not just be mobile devices we’ll be talking through, or people we’ll be talking to.
The social life of things
‘Smart’ things are slowly but surely creeping into our life in the form of connected light bulbs, wearable trackers and things we can program to sense, inform, assist and report.
We’re still testing the waters with these objects. We don’t yet know what their boundaries should be. When it will be useful (let alone socially comfortable and ethical) to track, measure, observe, and remotely control. The current generation of objects is also a product of the same world that gave us app stores and the ‘destination’ mentality.
Each thing currently believes it’s the sole object of our affection. It is often DRM-riddled, promotes stickiness, speaks only to partners, keeps its inner workings to itself, and requires a person’s undivided attention in the form of an app that must be downloaded often before even the smallest interaction.
This is all well and good for the odd hero device, but can’t possibly scale once these devices begin to seep into public spaces.
Thanks to new technologies such as Web Bluetooth and the Physical Web, the ability to discover and request quick, just-in-time, low-commitment interactions with objects and spaces will soon be available to anyone with a web browser…which is where IoT collides with the rest of this story.
Web pages are currently obscenely heavy, and while their location is easily shareable, their content may not always be. This is already a big problem and has spawned the less-than-optimal workarounds that are Facebook Instant Articles, Apple Newsstand and Google AMP.
So here is a question: How is a quick, contextually-initiated, one-off interaction with a physical thing (…a vending machine, parking meter, or virtual aisle-bot at Lowe’s) that different from an equally quick, contextually-initiated, one-off interaction in a digital context (…your utility company’s Slack bot, a WSJ reporter in a WhatsApp group chat, or a drive-through coffee shop’s “pay while you wait” app that lives within WeChat?)
We’re not completely sure that it is.
- Ask me to download an app before I can initiate the conversation and you lose me.
- Deliver me to some manner of home page instead of the exact content, interface or experience bundle I need, and you lose me.
- Deliver a slow, heavy, ad-laden, tracking-riddled UI and you lose me.
What’s more, you possibly lose me forever because, quite frankly, I’m now standing in your store/museum/airport/campus/hotel room, mid-interaction and you have the nerve to drag me through your entire org-chart’s worth of architecture and template baggage, or try to up-sell me?
Sorry. I’m out!
Platform-level? OS-level? Or World-level?
For better or worse, the “app as a destination” paradigm appears to be dying in favour of platforms that leverage a context (social, personal, historical, proximal, behavioural), implicit user action, or declared intent to deliver the most appropriate just-in-time micro-interface for whatever job needs to be done.
Google the term ‘conversational UI’ and you will find a growing collection of ideas, speculations and prototypes of next-generation light-weight pseudo-OSs based around some mixture of Chat, Cards and notifications. In most of these articles, the author eventually asks: Who will “win” at this new space?
Will it be Slack? Or Facebook Messenger? Or Kik? (I’ve yet to see anyone suggest WeChat…which is interesting.)
This feels like the completely wrong question given that what we may be talking about is a generation-defining evolution in how we interact with computers, and what may amount to the creation of a new key piece of global infrastructure.
It’s not so much the shape and metaphor of these new UIs that will matter — but who will control them.
“It’s admirable for (Zuckerberg) to say that Facebook is more social mission than business, and perhaps he even believes it. But Facebook is more like infrastructure now — like roads and bridges and plumbing, the kind of networks we all rely on.”
Exactly. So how’s that going to work?
We see three paths this future might take. Part-speculation, part-provocation, we have named each of these after the actors who will most shape them.
In this world, each brand or app (Facebook, WeChat, Slack, McDonalds, the City of Vancouver) becomes a platform. The ‘destination’ meme continues as each organization will likely want to own their platform.
Best case: We’ll be back to acquiring and managing hundreds of ‘apps’, each with its own API and slightly different way to do what is effectively the same thing (which, if these platforms grow in use, will likely now include wanting to own, choose, or partner with a payment platform). Same emperor, new clothes. If we’re lucky, brands maybe leveraging a common messaging framework, and this enables some (authoring-level) interoperability.
Worst case: There will indeed be a few ‘winners’, and they will get really, really big as the network effect of being ‘an enabler of everything’ works its magic. Through an embedded app/bot/service ecosystem, these platforms will enable millions of private and public services (a bit like China looks today viewed through a WeChat and Alibaba lens). Once these platforms become a de-facto layer of infrastructure, only drastic measures may save us from the vagaries their evolving business models.
Mathew Ingram makes an additional and crucially important point about the platform-level vision (entire article recommended).
“…even if it is possible [for one platform to ‘win’], I would argue that by the time it actually occurs, the entire market will have become so commoditized that there will be no point in owning it.
What value is there in General Motors, The Guardian or your local municipality in owning, managing, and defining what is effectively plumbing? Wouldn’t it be far better to leverage a much larger and established ecosystem?
(…which potentially pushes us back to a winners take all scenario)
Unless of course, we completely rethink where the plumbing (or infrastructure) layer should reside.
In the OS-level vision of the future, the operating system provides the service infrastructure (think Google NOW on steroids). How open and interoperable this infrastructure would be, remains a point for discussion (there are always many possible versions of the future). In the example that follows however, we’ve assumed a certain degree of openness.
[PS — We’ll use Quartz’ conversational news app as an actor in this story]
Architecture: In an OS-level world, Quartz would build a series of bots and related services that plug directly into the major operating systems of the day (mobile or otherwise). Thanks to some degree of OS-vendor cooperation, Quartz wouldn’t have to build a fundamentally different service for each OS.
Discovery and ‘subscription’: Users would access the Quartz service using some form of universal reference indicator (URI). You can see conceptual beginnings of this in a few current-day bot services:
- Users of present-day Telegram can add a bot by clicking a URI that ends in ‘bot’ (e.g. https://telegram.me/your_bot). At the moment, that URL is useless if you don’t have their app installed, but what if your OS could natively handle the action using a native rich-messaging interface?
- Clara and Amy are bots that leverage an OS’s native email support. No download. No formal subscription — just email them to invite them to a conversation (…only if and when you need then).
User interface and platform APIs: Rather than bundle its own UI, the Quartz app/bot/assorted services would leverage a collection of core OS-level components: rich media chat handlers, media containers, action-enabled notifications, VR-handoff prompts, a web browser (now effectively rendered an OS-level component). With much of their UI effectively handled, these apps would be far smaller and consist primarily of content and business logic. Some might still need to be installed (to more seamlessly tap into key device APIs), but others would simply store core aspects of their functionality, or stream the whole lot when needed.
Payments: Ideally, payment would be just another service that users could subscribe to (provided by third parties, a neutral or decentralised protocol, or the OS itself).
Expansion and customization: To further enhance the experience, the Quartz service could also intersperse ‘stock’ interactions with experiences; rich but lightweight task or context focussed app-bundles that would download at point of need. These apps could, for all intents and purposes, be widgets (iOS and Android both offer them) but it’s useful here to consider how apps that live within the WeChat pseudo-OS currently handle this scenario.
Chinese brands increasingly rely on the web to deliver ‘light apps’ (qing ying yong 轻应用): “one-off, zero-download, hyper-targeted mini-sites” that act as connective tissue between WeChat’s plumbing layers and the offline interactions that often lead in and out of the platform.
A user might for example scan a QR code in a shop, which would lead to a light app whose call to action would then lead to an interaction or transaction inside of WeChat (…which might in the end generate another QR code to trigger another offline interaction).
Once you accept the idea of the web as first-class citizen, it shifts the balance of power. The universally accessible, multi-device and multi-context web is there to enhance native infrastructure — not the other way around.
Kevin Kelley hopes Google AMP might be the start of…
“a containerized, embeddable article that travels to any site with brand, revenue, analytics, and links attached”
In the OS-level world we’ve described, this certainly seems plausible.
What could go wrong?
As we alluded to earlier, iOS, Android and whoever else may not wish to collaborate. They may instead each create their own slightly particular way to do all the things we’ve discussed. And if they’re too different, we’re back to building, marketing and discovering within a few giant silos (or several dozen silos if Amazon, WeChat, Facebook Messenger and anyone else who’s verging on OS behaviour decides to go it alone).
Despite all this, users could still fare quite well as they would by and large enjoy lightweight, ‘quick-subscribe at point of need’ applications that leverage data and context as a material, while benefitting from the performance, usability and device-to-device portability of an OS-level ecosystem. They would also be able to choose between a few OSs should one of them prove (socially, culturally, financially or ideologically) inadequate.
There’s a fine line between OS- and World-level and it has to do with intent. Our OS-level vision already pre-supposed some level of OS vendor cooperation. But what if we could go much further?
When we currently use the term “web” or “browser”, we mean a very specific thing. A thing that leverages open standards and whose values revolve around neutrality, openness, interoperability and user agency. But if apps are to become mere services within an OS that delivers the best UI for the moment — why should the web not be a part of this change? The web could simply become yet another service. Or a core component. Or better yet— a defining model for how all other services should behave.
A call for discussion….
“All models are wrong, but some are useful” — George Box
There may well be a few too many ideas in this article, and we’re the first to admit that some may not hold up to scrutiny.
It does feel however as if the seams are showing. The web, and apps as we know them have however already begun to change. Some of this change will no doubt be led by industry hype, but it would seem to our advantage to consider not just who will “win”, but what we could all gain or lose through such a dramatic change. And do so before someone else’s vision of the future decides it for us.
If you found these ideas interesting and want to chat further, we hope you’ll drop us a line. We’d love to start a conversation about which bits of what we’ve described make sense, which bits don’t, and how some version of it could work. For obvious reasons, we’d particularity like to chat with anyone who designs, builds, or works on a browser, an OS, a social platform, or some combination thereof.
Enjoy this article? Subscribe to yiibu’s newsletter for a short weekly selection of found tidbits and research about our evolving relationship with technology.