Cross Channel Design

My goal here is to understand trends in interactive media that challenge both the User Experience designer and the business strategist when it comes to combining different channels of communication between a business and its customers. As I go about it, I will provide examples of specific services I worked on and how such challenges were approached. We need to figure out what questions to ask in order to define the right user experience strategy (more than comparing solutions).

Let's set the stage 

By "cross channel" design I refer to a type of interactive design that involves various modes of interaction (primarily voice and graphical interactions) to access the same information system. For instance a user may check her balance either through an interactive voice response system (IVR) (i.e. by calling her bank) or through the bank web site. Voice and web are two separate channels typically used by companies to interact with their customers. They have been the corner stone of automated self-service applications. A face-to-face conversation in a bank branch is also another form of channel but it involves a human at the other end, not a "robot" . Live channels involve two or more people interacting in real time either in person, by phone, or via an online chat. Typically a customer and a customer service representative. Such channels are the most expensive to deploy and increasingly companies use automated self service to cover the most repetitive and predictable interactions at a fraction of the cost of a live agent. Asynchronous channels imply a delay in the dialog: a customer may send an email to tech support or receive an SMS indicating that a transaction was completed.

Multimodal Interaction

A digression on multimodal user interfaces appears to be very a propos at this point. Multimodal interaction has been discussed in high tech circle at least ever since research took off on virtual reality and the so called data glove or wired glove some 20 years ago or more. The MIT Media Lab certainly pioneered some of the research with prototypes such as the famous Put That There (1979). The bottom line, however, is a man-machine interaction that combines most of the ways a human interacts with the world: seeing, speaking, hearing, touching (although in commercial products, touch is used mostly to simulate clicking; touch to sense texture or temperature has not really reached the masses). Similarly, smell is not really part of the equation yet except for some prototypes. Multimodal UI can also involve a gesture: the early handheld devices already offered gestures to specify an action to be performed (such as delete).

Trends

Why did I mention here multimodal interaction? Because it guides us to understand a trend where slowly traditional channels are morphing into integrated modes of navigation. Both designers and business strategists need to see the connection to figure out what steps to take, what to think about, and be ahead of their competitors. 

Each "Channels of communication" specializes in one mode of interaction, usually voice UI or graphical UI. It becomes then obvious that emergence of multimodal interaction will affect channel integration. It is not clear yet exactly how: when you have your smart phone in hand would you go to a web site to check your account balance or would you call your banK? Asking the question is perhaps more important than answering it (as used to say my 12th grade math teacher). It sets the terms of the issue to solve.

Currently the consumer industry handles channels mostly very separately: the customer gets a toll free number to call an IVR and a URL to access a web site. Even agencies in charge of designing and developing those channels act as two very different and independent entities. Inside the company different managers may oversee those 2 channels. Airline companies typically handle independently their automated voice service and their web site. Even branding may not be fully consistent.

But the communication devices we are using - such as the phone - are re-inventing themselves: the iPhone is starting to offer multimodal interaction, smart phones are integrating music, video and web. This puts pressure on channels as I defined them earlier to offer a better integration.

Design Challenges

At the very least channels have to use a consistent language: if one channel uses the term "payee", then the other should use it too as opposed to using a term like "recipient".

The steps and processes used to achieve a goal should also be consistent. This goes as far as re-thinking top level menus offered on the site home page and the phone service.

Another example could be using the same security infrastructure and UI guidelines for user authentication. In 2007 I worked for RSA Security designing a voice authentication system. Its main goal was to allow financial institutions to offer transactions over the phone that require higher levels of security (typically enabling money transfers). However at the time I was not able to incorporate in the design the challenge questions that customers were typically used to on the web site (such as "what is the name of your first pet?"). Such question could be triggered if you are not using your regular phone.

The limitation here was technical: the infrastructure could not be shared, resulting in a customer experience that was not as good as one would naturally expect. Instead we had to come up with another way to double check the caller's identity on the phone, hence creating an undesirable inconsistency.

That being said, both channels are of course intrinsically different. If Voice Verification is used on the phone it changes the customer experience but it remains consistent with the voice medium itself and as a result feels natural: entering a password on the phone (action that is more relevant to the web UI) is not very natural.

Some actions are indeed performed a lot faster online: if you need to answer a series of yes/no questions you can quickly do it by submitting one form using radio buttons. On the phone you will have to answer the questions one after the other. But then some users tend to use more the phone and other more the web depending on their lifestyle (in my experience additional user research is needed here).

When you step back you have to ask yourself t what's the life and behavior of your user throughout the day. When would voice be better? when would web be the medium of choice? Too often companies act based on peer pressure: we need to offer voice because our competitors do (or because touch tone is outdated), we need to offer voice authentication because it works and it is now available. 


Incidentally RSA Security dropped its Voice Authentication Product after the sales teams failed to bring any customers 6 to 10 months into the project. No budget was invested in doing direct research to understand how users behave in relation to their money, the communication devices they use, their age and lifestyle. I found a number of studies that show usage of phone versus web in banking or who uses cell phones and how. There was no definitive proof that what we were trying to build would affect revenue and customer service in a significant way (one that would of course justify the investment). Our main argument was new government regulation that forced US financial institutions to provide higher security to their customers to prevent fraud and identity theft. There had been a trend in fraudulent phone calls that were similar to phishing on the web. But this clearly was not enough to convince businesses although it has been resurrected in collaboration with Intervoice.

Future posts will follow up on this issue as I intend to continue exploring it.