Wink Communications

» see wink.com as it lived, on the wayback machine

~ the obituary ~

Turkish television in 1999 did not ask you anything. It arrived through an aerial or a dish, it played, and the only thing you could send back was a phone call to a number on the screen. So the product Wink Communications spent nine years building never had a shape here at all. Wink's entire premise was a return path, and a return path is an infrastructure decision somebody else makes for a whole country before you get a say in it.

Brian Dougherty founded Wink in Alameda, California in October 1994, after Geoworks, and the idea was modest by the standards of the era. It was not television on the internet and it was not video on demand. It was a small overlay painted on top of an ordinary broadcast, triggered by the broadcaster, which you answered by pressing a button on the remote you already owned: a car commercial with a brochure request attached, a game show you could play along with. The Wink Engine that decoded the overlay was licensed into television sets and set-top boxes, and the Wink Response Network on the other end collected what viewers had pressed and handed the totals to whoever paid for the spot.

It ran commercially in Japan from October 1996 and in the United States from June 1998, and the list of names attached to it was as good as that business ever got. The networks carrying enhanced content included ABC, NBC, CBS, Fox, CNN, ESPN and HBO, while the distribution side ran through DirecTV and EchoStar on satellite and Charter, Comcast, Adelphia and Cox on cable. The cap table was its own argument for the idea, with Microsoft, Motorola, Toshiba, NTT, Vulcan Ventures and Benchmark all in. On 19 August 1999 Wink sold 4.75 million shares at $16 and raised about $76 million, and it was the first interactive television company to make it to a public market.

~ the case file ~

~ how it worked ~

The engineering was the good part, and it was engineering done under a hard constraint: the bandwidth available for this was almost nothing. Wink data rode along in the gaps of an existing broadcast signal, which meant an overlay had to be described in a few kilobytes and had to survive being sent one way with no acknowledgement. The receiving end was a decoder chip in a box that cost the operator as little as possible. There was no session and no round trip. The television painted the overlay, you pressed a key, and the box remembered your keypress until it next got a chance to phone home over the set-top's dial-up modem.

That last part is the piece worth holding onto. The return path in most Wink deployments was a telephone line running from the set-top box to the wall, dialling out at some quiet hour of the night to report what the household had pressed. An interactive medium with a latency measured in hours. It worked, and it produced real response data that advertisers had never had from television before, and it was completely obsolete the moment a house had broadband and a laptop in the same room as the set.

~ the climb ~

Wink's best years were 1999 to 2001, when interactive television was the thing every cable operator had to say they were doing. Microsoft and Wink announced a joint push on the ATVEF specification in June 1999, which was the industry's attempt to agree on one way of doing this before ten incompatible ways shipped. Charter signed to deploy Wink Enhanced Broadcasting across its digital households. Comcast had it in a million and a half homes by late 2000.

The revenue never matched the footprint. Wink took a cut of transactions and of the enhanced advertising, and the enhanced advertising was a line item that broadcasters treated as an experiment rather than a budget. Meanwhile the same advertisers discovered they could put a URL on screen and get a measurable response with no licensing deal, no decoder chip and no telephone line involved. By 2002 the company was worth a fraction of its IPO price.

~ the day it died ~

Liberty Broadband Interactive Television agreed on 24 June 2002 to buy Wink outright at $3.00 a share, around $100 million for a company that had raised $76 million at $16 a share three years earlier. The deal closed that August. In September Liberty turned around and sold Wink on to OpenTV, which it also controlled, for $101 million in cash, and put ACTV through the same door a few months later. The theory was that one owner holding all the interactive television pieces could finally sell a whole product.

It did not survive the first contract renewal. In November 2003 OpenTV said DirecTV would not extend the Wink carriage agreement. It expired on 22 December, and about 10.4 million enabled set-top boxes were deactivated. OpenTV's chief executive said plainly that the terms on offer were not economically viable, which is the sentence that ends most of these stories. Nothing broke. Nobody lost interest in a dramatic way. Two companies could not agree on a number, and the largest interactive television deployment in the world was turned off from a control room.

~ the next chapter ~

Everything Wink was trying to do happened, and none of it happened on a television. The overlay you answer while watching became the second screen in your hand. The response data advertisers wanted became the whole business model of the internet, at a resolution Wink could not have imagined. Even the mechanism came back: press the red button on European broadcast television and you are using a descendant of the same idea, and the smart TV in a living room today runs an application layer that does what the Wink Engine did, with a WiFi connection where the dial-up modem used to be.

What Wink got wrong was not the concept. It was the assumption that the return path would come through the television. Everything the company built assumed the set was the device the household would answer with, because in 1994 there was no other candidate. The answer arrived instead through a completely separate wire into the same house, and once it did, the television went back to being what it had been the whole time, which is a thing you watch.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »