No One *Needs* Your Bots
How to appeal to the masses.
Most people use tech products not because it’s cool to do so or because they are tech-savvy. They use them to get something done. To get some value. That value could be utilitarian, entertaining, or educational. If you are an avid reader of tech publications, tweet throughout day, or live on Product Hunt’s website or know what chatbots mean and why they exist, you do not qualify as “normal,” nor would be considered as part of a mainstream audience.
Discovery > Quality
In the minds of mainstream audiences, getting the right things to do what is to be done at the right time is much more valuable than how well those things do their job.
So, imagine you have a need for doing some specific task. And, you are finding it a bit difficult to find an appropriate tool that serves you well in enabling you to do what you wanted to do. This is what tech geeks refer to as an “discovery issue”. If you can’t discover something that suits your need or want, you wont be able to use it and the people who made that service won’t be able to make money off of your usage at the end of the day. Its as simple as that.
App Discovery
Fundamentally, the app discovery issue started creeping up after companies started going all mobile and app-first. When you have too many units of different varieties of a particular commodity in a warehouse, efficient discovery is bound to become a problem. This issue is only getting increasingly worse. Even in 2016. Fortunately, there has been some innovation in app discovery by third-party players like Kit, Product Hunt and Shelf that have added social feel to the conventional set-up.
So, now, if bots are the new apps, whats going to make mainstream audiences (again, you probably don’t count) to use them?
So, What Now?
This app discovery problem is one of the driving factors for the rise in conversational products. Tens of thousands of developers are building bots on the Facebook’s new Messenger platform already. When we add on top of that the bots being built on Slack’s platform and various standalone bots, we get a lot of bots. Already. The rate of bot development is only going to increase. So what we will be left with in the next 10 to 15 months is an incredible number of bots. Just like how we now have a ton of apps.
How Much Do We Need?
Studies show that an average smartphone user uses 5 apps regularly. So, basically, app developers can make apps all day, but an average Joe uses only a bunch of them on a consistent basis. This is probably so because a small collection of apps fit into the lifestyles of millions and serve all (or most important ) their needs and desires. A good chunk of these apps are social in nature. No surprise there. So, now, if bots are the new apps, whats going to make mainstream audiences (again, you probably don’t count) to use them? Whats going to make them install or search for and consistently use even five bots? Despite the app and smartphone explosion in the recent years, most people have had no intention of exploring the app stores to look for cool new apps regularly. So, whats going to encourage them to transition from apps to bots? Or, use bots in conjunction with their usual apps? What’s going to convince them of bots’ need?
What Techies Know That Probably Others Dont?
Yes, some of us know that bots are more efficient than apps for some purposes. Yes, some of you know that bot development is much faster and cost-effective than that of apps. But, does a 20 year old girl studying for her college finals on her pre-med track really care about all that? No, in most cases.
So, whats going to motivate her to try out this new thing called a bot after her exams are over? Whats going to help her understand why bots matter or why she should spend less time on Snapchat, Instagram, Google Inbox or Amazon’s website and consider spending some time playing around with some Messenger or Slack bots? Or, even worse, why would she download standalone “apps” that have bots built into them (Think Operator, Lark & Ozlo). The irony of this would confuse her even more.
She might test them out if she has an annoying tech-savvy friend or if she was curious enough to learn more about this new thing after stumbling upon some article online. But, this won’t happen with most people.
We usually are hesitant to try out new things if we fail to make some sense of it or see or feel some value that could possibly be attained. Things get even worse if I am totally in the dark about this crazy new thing. How am I supposed to even know about “bots”? Press cycles? Tweets? F8? Try again. Most people around you are not into all of that. At all.
So, we have this amazing new trend that is being touted as the next big platform and the next thing after apps. But, millions of people for whom we are making this do not have the slightest clue about it. They have no intention to step out of their digital comfort zones.
Even if people know about bots, how would they know what to use and what not to? How do they decide why they should pick bots over apps? In what situations should they not use bots? People have neither the intent nor the incentive to look into finding answers to these questions.
They have got everything that they think they need.
How do we show them the value proposition of this new thing to encourage them to take a moment to try these bots out? How do we make them feel the need?
How Will They Discover Bots?
What I think could be the best approach is organic discovery via both social graph and platform assistance.
So, the bot discovery story is probably going to be similar to the app discovery story it is not fundamentally rethought, redesigned and constantly factored in all the design, engineering and marketing processes of bots.
Currently, different platforms and developers have taken different approaches to help users discover their bots —
Conventional Approach
The messaging players could provide bot stores the way Apple & Google provide app stores. Slack already has its Slack Directory. And, recently, botlist came into existence. There’s BotArena and several others.
Such bot stores do not look or function differently than the conventional, incumbent app stores.
This discovery solution of taking the conventional model, which is now broken, and tweaking it a bit to have a new store for this new era of bots doesn't sound like the best solution. This is mainly because these stores are predicated on clear intent of the people to actually look for something. We know most people don’t do that. So, its safe to assume that they wont go to these stores, click around, explore categories, scroll up and down, and look for new bots to serve their needs that are probably already being served ( in some cases, inefficiently) by apps.
This just does not feel like the future. Its a good transitional step to get where we would one day.
Messaging platforms could have bot stores built right into their messaging apps. Kik’s store is one such example.
This solution doesn’t address the issue of lack of intent as well. This is no different than the traditional App stores if the messaging platforms are thought of as mobile operating systems.
The Facebook Way
Since Facebook is the biggest messaging platform and is in the driving seat the way Apple and Google were in the context of apps, its moves will be defining decisions for a lot of things that go on in the bot world. Facebook has, in some sense, had a three-pronged approach to bot discovery.
Firstly, it has a Snapchat-style discovery section for its partner bots.
This move by Facebook feels like a temporary hack to bring bots to the attention of its users. It doesn’t seem like this will be a permanent solution. Once we have thousands of functional bots, Facebook will have to make a lot of design changes if it goes ahead with this solution. I was surprised that so may people I know either have no clue about these bots or have noticed this section but not felt compelled to click on them out of curiosity. Its like how Snapchat power users do not use the Discover section even though its right there in painfully obvious fashion. This is basically an in-your-face hack. This didn’t work, despite being right there, probably because people fail to grasp or foresee the value of the new micro apps.
The second discovery approach within the Messaging app is good old-fashioned searching.
This is essentially a messaging-native equivalent of searching in the conventional app stores. Except, it’s much worse. Searching requires you to know exactly what you are looking for. You got to type in the name of the bot. Yes, there is no “downloading” and “waiting time”. You just search the name, pick the bot, and can start interacting in a new thread. But, there is no categorization and room for serendipitous discovery. So, its like a worse version of the App stores. So, our pre-med friend from earlier not only needs to have the intent to use a bot but also know the name of it.
The third, my favorite, and the most efficient, scalable approach so far by Facebook has been the contextual approach. Its a messaging-first approach. No need to look for a bot in a bot store or install a bot natively by searching in the search box. Ride-sharing is a great example of this. We can request a ride from a car service without ever needing to download an extra app or leave a conversation.
Now, is this convenient, faster, and native? Yes.
Does it facilitate easy discovery? Not so sure about that.
Would our pre-med friend happen to just stumble upon this when she is planning to go out for a movie with her friends? Probably not.
She is much more likely to use the Uber app the way she is likely to use the Amazon or any retailer’s website or app instead of looking for the bot. This approach is definitely the best one yet. It is 10x better than getting people to download apps for using the bots. It is 5x better than looking for Poncho’s or CNN’s bot via the search section within FB Messenger.
I think we could do better. It is early. Way too early. But, we need to start thinking now so that bots do not meet the same fate that apps did in the context of discovery.
What I think could be the best approach is organic discovery via both social graph and platform assistance.
Our pre-med friend might have a friend that is tech savvy : ME
Lets say we are chatting on FB Messenger or Telegram or Slack. What if I could summon the bots into my conversation with her at the right moments? So, lets say, I summon in the CNN bot while talking about Mr Trump with my friend to update us on his latest racist remarks. Or, I could summon Poncho to inform us about the weather in Princeton so that we could plan whether we should go shopping. This makes this new bot thing cool for my non-techie friend in a second. This educates her. This helps her see the value even without using it per se. This unleashes the power of bots on-demand. This could help create social viral loops for marketing these bots. No intent is needed on my friend’s part to experience the beauty of these bots. Intent is created within the messaging context by someone who she is talking with and hopefully trusts.
Our techie friends simply make the bots appear in our interactions.
But, what if she does not have a tech geek like me to chat with? This is where platform assistance comes into play. There is not much doubt in my mind that Slack and Facebook would become the earliest gatekeepers of this new bot world. They own the messaging context and have growing platforms, growing user-base and the tools in place. So, they have all the access they need to all the bots that are built on top of them.
So now, lets imagine again. My friend and I are chatting. Both of us, like millions of other “normal” people, have no clue what bots are. We have Facebook, Instagram, Snapchat, Uber, Evernote, Google Inbox and Spotify on our iPhones. We are making a plan to go see Captain America right tonight. What if Facebook identifies the context of our conversations and natively surfaces Uber bot to plan our travel, IMDb bot to show us the movie review and a Fandango bot to book tickets? Assuming messaging platforms implement this in a not-so-creepy fashion and there are not many privacy concerns, we could have contextual discovery wherein the inherent platforms pick up on our keywords from our conversations to introduce us to new bot services. Such approach could open up new bot ranking algorithm and monetization opportunities for Facebook, Telegram and Slack that are increasingly becoming the new Googles of the bot world.
Once we start using these bots and the inherent platforms have enough data on our bot usage, they can use that data to make personalized recommendations for new and better bots. Again, this offers room for monetization. This was something that a lot of us hoped the incumbent App stores would implement. But, sadly, they did not. And, hence, the discovery issue cumulated.
Platform-assisted discovery would help people see and feel the value of bots in real time under right context. Once the value is understood, there will eventually be a need for them. And, then, people will be well-equppied with the right intent and knowledge to explore and user new bots.
The Messaging Platforms Simply Make the Bots Appear in Our Interactions.
One last time, let’s imagine again. Imagine a world where Google expects us to know the exact title or url of the webpage for googling things in 2016. Imagine a situation where we would be expected to know before-hand about something we would like to learn about or discover. I am not even sure how that would work out.
That. Would. Just. Suck.
We cannot expect people to know the exact name of the bots for discovering and using them. Most don’t even know they exist yet. We should come up with scalable methods to facilitate discovery with least effort or intent on the users’ part. They are happy with their apps. The bot movement is more of a developer-driven movement as Matt Hartman mentioned. That mainly stems from the fact that people do not know that they might need these bots. They do not feel the need. Yet. It is the responsibility of the developers to create an environment wherein people could learn why, when and how they should pick bots over apps and websites for certain use cases.
We just gotta show them these bots in a relevant and useful manner. I think such organic discover based on a combination of platform assistance and social graph leverage would be a great way of doing so. By getting the discovery process in a relevant, messaging context, we would both help with exploration and on-boarding about how, when and why of these bots.
