Online action in the era of WhatsApp

Update: since we published this article in 2016 a lot has happened. But it is still not entirely clear how messaging apps will change our behaviour and while we have seen many tests and pilots, messaging apps are still not part of the standard toolkit for campaigning.

Reimagining online action for mobile devices

Computing and interfaces have come a long way. Our generation grew up with a keyboard, a mouse and a graphical user interface to navigate. The way we have executed online actions such as petitions and email to target actions in the past 15 years is built around the assumption that websites and forms are easiest way for our supporters to participate.

But as devices are changing and smartphones have outnumbered other computer-like devices, a new era of interfaces is on the way to dominate our supporters' habits. Messaging apps like WhatsApp, Facebook Messenger and Telegram are taking over the world.

These applications combine what UX Designers call "Text based interfaces" with graphical elements. The main way we interact with these apps is typing (in most cases using our thumbs). At the moment the primary usage of such applications is to send messages to friends and family. But we're using these applications so much that we're getting used to a new means of interacting with our devices and the internet, which has massive implications.

Artificial intelligence and conversational interfaces

With the rise of semi-intelligent computer programs we can have conversations with our devices through typing. We ask "What is the weather like" and the device can "understand", check the latest weather report and respond "the weather is nice, but do bring a coat as it might rain later".

Interfaces are being reinvented as conversations. These conversations will be dominated by the platforms people are used to already, like WhatsApp. Just like browsers have dominated how we interact with the internet in the past ~20 years, messaging apps may one day become gatekeepers.

Working in campaigning, the key question we have to ask is "what does that mean for online activism?". How can we make it easy for our supporters (yes, even those who grow up with a smart phone in their hands) to participate in our campaigns?

When thinking about such fundamental shift in how people interact with us it helps to reimagine certain elements of our work and to build prototypes. I'm not suggesting messaging apps is the one trend that will dictate our work future - but there is no harm in building prototypes to explore what might happen and how we can use new tech to create real world impact.

Imagine "email your MP" without leaving the WhatsApp interface

A typical example of an online action in the UK is an "Email your MP" action. The user interaction looks like this:

• Supporters receive an email
• Supporter clicks on the call to action link
• Supporter their browser opens, they land on a website
• The supporter looks at the landing page, scans the information and may read some copy
• The website has a form on it asking for the postcode
• When the supporter enters their postcode the page brings back their elected representative
• Through this form supporters can send an email to their MP

How could this look like in a messaging interface? Well, let's start by saying it's more complicated. Why? Because it's a conversation. Every single interaction may have numerous different possible responses which means the whole process is more of a tree-like structure because there is not a single simple way to the desired goal.

For this scenario let's assume that we will take the fastest path through the action.

• Supporter receives a message asking him/her to take action (in a messaging app)
• The supporter can say yes, no, might say “I’m busy, will do it later” or will ask for more information --> for the purpose of this scenario the supporter will say “yes”
• the bot responds by asking for the name and the email address (so the MP can respond)
• the supporter provides both and the bot validates the information
• the bot asks for the postcode so the constituency can be identified
• the supporter provides the postcode but can use an interactive widget to find his/her postcode
• the bot brings back information about the target and presents an interactive widget with the message that will be sent to the MP
• The supporter can edit the message, double-check if all information is correct
• Then the supporter can hit send
• the bot confirms that the message has been sent and presents the option to share this with friends
• the bot may also respond with updates about the campaign at a later stage

As we're going through this scenario you might have noticed that providing information that usually sits on a website or landing page will also have to be reworked as a conversation. Content like the detailed background information, social proof and an explanation around the theory of change may be part of the initial "call to action" message to supporters - but a lot of it will be too lengthy for this medium.

We'll have to train our bots to understand who the organisation is, answer basic questions about it and figure out a good mix of text based and visual interfaces to explore this information.

Let's make it happen

We live in a world where ideas also need funding. If you think this might be an interesting thing to explore, prototype and run the first action with your organisation, feel free to reach out.

New things happen when brave organisations are willing to invest in new, innovative solutions that help our sector explore what the future of campaigning might look like.

This functionality is on the roadmap for our tools but without additional funding we won't be able to work on it for at least another year. If you think this is for you please do get in touch.

Risks, hypothesis and thoughts

There is a lot to learn about these new interfaces and how people will interact with them. What's outlined in this blogpost is really a thought-experiment, not much more. We've only done very basic research, we've not conducted user tests, we've not had the time to run pilots and we have only been able to play with the different technical components required to build a system like this.

There is a long way ahead to an affordable "ready to go" solution.  Obviously the speed of global messaging bot adoption as well as the resources invested in projects like this will be the biggest influence on when the non-profit sector will have reached the messaging era.

A few good questions that need answering as we're building the first version:

• Do people want to receive whatsapp messages form organisations? What can we learn from the first few bots out there, like the CNN news apps.
• People don’t trust robots - how long will it take to overcome this? How can we make people feel comfortable messaging to bots?
• Would people prefer to just take the action in a browser like they’ve done it in the past 15 years?
• What style of language will people prefer when talking to bots?
• How long should the perfect message be?
• When should we use rich interactive elements as part of conversations and when do people prefer free text conversations?
• How long will it take for artificial intelligence to be smart enough to keep the conversation going without constantly asking for more info and without pre-selected “yes” “no” type questions?
• how can we provide a fall-back when the bot fails so our digital staff can take over conversations and respond directly?

If you can think of other good questions and hypothesis please tweet at @moreonion or email florian@more-onion.com