I’ve been having a short Twitter exchange with Paul Downey (@psd), someone who I hold in high intellectual and personal regard. I’ve made an assertion that has Paul snorting his coffee back up through his nose and into his keyboard: that the Web is not a communications medium. Justifying this claim can’t be done in 140 characters.

Now, there is a sleight of hand I’m pulling off here. You can build communications media on the Web, but my claim is that the Web itself is not one, and that has subtle but significant consequences.

(Much of this post is an edited version of a section from my free white paper on the future of communications in the cloud, Connect, Interact, Transact.)

To me, a “communications medium” is one which allows bi-directional messaging. It is something that doesn’t require both parties to be in the same “place” (physical or virtual) at the same time.

What the Web lacks is an ability to “introduce” you to the website, and to provide a means for you to share data about yourself including how the other side can choose to initiate contact back to you. All the technical components for doing this exist, but in the actual Web we use they are not deployed to any meaningful extent. This is a hole we tried to fix on the WAM project at Sprint a decade ago, and VRM is acting as a framework today to address similar issues. (It’s a separate essay/rant on why that Facebook cookie delivered to every website doesn’t cut it.)

This contrasts with the phone or SMS; if I contact you, then typically you will automatically be able to contact me in return. It’s very limited (roll on ENUM or its successor), but the gesture-response pattern of conversation can carry on over time and space easily.

“Push notification” doesn’t solve the problem. Communications media are far richer than just the ability to send a message and interrupt the user. For example, there is a need to prioritise, filter, pass security context, build governance systems, etc.

The outcome is that we need a “mandatory middleman” to provide a communications service (messaging, voice, directories, privacy, filtering, etc.) in order to enable any kind of business process that is not synchronous and initiated by the end user. Enterprises cannot solve these gaps unilaterally by building smarter enterprise web sites. They are dependent on communications services providers to act as middlemen and provide channels for messaging.

Why is this?

1. Web sites are like one-way mirrors. Whilst a web site allows the customer to reach out to the enterprise, the converse is not equally true. Only tools like SMS and telephony can ‘buzz, flash and ring’ to interrupt the user to participate in a time-sensitive business process.
2. The user is blind to messages at the enterprise web site. Most messages do not merit interruption of the user. Enterprises therefore need to send messages to a ‘place’ the user will ‘pass through’ and view in their typical day. These ‘places’ take on a wide range of forms, from a social media service to a smart tablet you hang on your fridge door. These ‘places’ are intrinsically provided by a communications service, and not a general-purpose enterprise web site. (This problem of being ‘out of sight’ is a common failing of online banking services, which offer a proprietary secure messaging capability, but leaves the user blind to the existence of a new message.)
3. The web remains deaf and dumb. Browsers continue to offer poor voice integration, and this is likely to remain so for several more years due to the enormous inertia of the legacy voice network technology.
4. The open nature of the web is a blessing for innovation but a curse for security. Combining the capabilities of multiple media, such as sending an email or SMS to notify the user of a message on the company website opens the door to security problems such as phishing.

These issues tell us that there will continue to be intermediary communications services that link enterprises to their customers, and that these cannot easily be disintermediated by enterprises. The value of each communications network to its users is proportional to the number of other people you can communicate with. This increasing return to scale implies there will be a small number of communications platforms acting as middlemen between enterprises and their customers, each with large numbers of users.

So whilst these communications networks may surface their user interface as a web site, they are not about the fundamental property of the Web which is hyperlinking. They are about message passing, subscription/distribution, analytics, prioritisation, etc.

The situation is rather more complex than this, as each medium is broken into its messaging infrastructure (e.g. PSTN, SMTP, Twitter’s underlying engine) and its user interfaces (Different phone handsets and voicemail systems, different email clients, twitter.com vs. Tweetdeck). If the money is in building communications systems that are “fit for business”, then it helps to be able to control both layers to build the features that enterprises need to efficiently, effectively and securely do business with their customers. Hence another wave of vertical integration may be upon us as SMS short codes, freephone and caller ID evolve in ways that fit with how enterprises do business, and are conveniently packaged by telcos for end users to adopt.

The ubiquity of SMS, voice and email is why these are the preferred modes of customer contact for enterprises today; if Twitter and Facebook want the Web to become a full-blown communications medium, they have to overcome a philosophical blockage: that the Web does not have the feature set, yet, to be a communications medium.