The web we all know and love is a web of sites. From YouTube to MSN, Google to Directgov, these exist to facilitate the exchange of content between humans. While the web will continue to fulfil this purpose, it is also becoming the conduit for machine to talk to machine. The implications of this will be as profound for society as for marketing.
To those with the tools and willingness to listen, the web has opened up a new dimension in understanding consumers. By tracking their actions in the spaces we occupy, we can learn about what motivates them, their buying habits and favourites.
We can tell what keywords they use to search a few days before buying something from us and create a statistical relationship with that outcome to give a value for that point of influence. We can tell which affiliates are creating real value and which slip in at the end of the process and claim the credit for all the hard work done by our other marketing activities. We can see how many times consumers visit before purchase and tie together their activity on our site with a CRM programme, so we reactivate abandoned baskets and sell accessories subsequent to a core purchase.
All this - and much more - is possible (although still unusual). Yet all of it relies on data collected from consumers' activity online.
What if we could build that out to the rest of the world? What if we could collect behavioural and contextual data from the interactions of machines with the world, consumers and each other?
The embedding of increasingly cost-trivial sensors into all sorts of devices is starting to create a world where millions of machines are reporting information back to servers.
It started with hackers and, on the early web, a flurry of coffee machines being connected to the internet, starting with one at the Cambridge University computer lab in 1993. Now though, hackers are competing with serious players to connect anything they can think of. Facilities managers are remote-monitoring buildings via feeds of temperature, weather and energy consumption over the web. Consumers are using their iPhones to connect to their home-security systems remotely, enabling them to view the security-camera footage wherever they are. Many of these are sharing their data through platforms like Pachube - a kind of YouTube for data feeds.
Nokia and the University of California showed how mobile phones could be used to observe road traffic patterns, by sensing the transit of handsets through the cellular network. This approach allowed traffic conditions to be monitored outside the limited reach of fixed sensor areas, and fed back information to users over the web.
Apple's iPhone is a bundle of sensors - accelerometers, hygrometers, compass, microphone, camera, proximity, ambient light - providing information that can supplement location data, which the phone's GPS can report back. Earlier this month, Apple registered a patent for a sensor to detect cardiac activity, allowing users to be authenticated via their heartbeat's 'fingerprint' - or perhaps funeral parlours to deploy rapid-response teams.
The creation of services based on data generated from sensors and machines around the world will be a boom business over the next five years. As these sensors become pervasive, and the feeds from them more public and shared, the noise of this data flood will be daunting. However, for marketers who can make sense of it, a new insight into consumer behaviour is going to emerge, creating businesses based on a dynamic connection with consumers that seems impossible today.
Andrew Walmsley is co-founder of i-level
30 SECONDS ON ... The Trojan Room coffee pot
- The origins of the pot feed pre-date the creation of the web. In 1991, researchers in Cambridge University's old computer lab building had just one filter-coffee machine, outside the Trojan Room.
- The machine was in heavy demand; while it was slow to fill, it always emptied quickly, and those working elsewhere in the building missed out. XCoffee - effectively, the first webcam - was created to minimise the resulting disruption. Those on the network could see an icon-sized greyscale image of the pot, updated every 20 seconds.
- When the image-grabber broke, the tool fell into disuse, but was resurrected in 1993 and the images fed onto the web. Thousands of users viewed the feed before it was axed due to the move to a new building in 2001.
- The feed's final image, of the server being switched off, can be seen here: .
- The last, and longest-serving, pot used was auctioned on eBay. The online operation of German magazine Der Spiegel bought it for £3350.