Funny some things cycle and recycle, you just never know when to expect them. A recent post by Phil McKinney, one of HP's genuine innovators along the way, recalled some 'forgotten video conferencing lore' that is perhaps worth revisiting (or maybe, for most of you, visiting for the first time).
Here's Phil's LinkedIn post: https://www.linkedin.com/pulse/question-nobody-wanted-ask-phil-mckinney-jtd1c/
Something I keep seeing in boardrooms, innovation teams, and strategy sessions, including my own, is that the most expensive mistakes rarely come from bad ideas. They come from good ideas built on assumptions nobody thought to question.
Not because the people in the room were careless. Because the assumption felt so obvious it stopped feeling like an assumption. It just felt like the plan.
That's what I keep coming back to this week.
─────────────────────────────
✦ This Week — Studio Sessions
Inversion Thinking: How to Guarantee Your Project Fails (On Purpose)
In 2005, I was at HP when we built Halo. Best telepresence system in the world. Life-sized. Perfect audio. Nobody had built anything close.
The team was brilliant. The technology was extraordinary. And we were completely wrong about the one thing that mattered.
We believed quality wins. It felt so obvious that nobody in the room thought to question it. Not once. Not out loud.
It wasn't quality the market was buying. It was access. Cisco understood that. Then Zoom made it a verb. We invented the category and handed it to them.
In 2011, under quarterly pressure, HP sold Halo to Polycom for $89 million. In 2022, HP bought the business back for $3.3 billion. Thirty-seven times the price. To reacquire a category we'd invented.
The failure wasn't hidden. It was sitting inside one assumption the entire time. An inversion exercise, asking "how would we guarantee this fails," would have dragged it into the open in the first planning meeting. The answer was already in the plan. We just never read it from the other side.
Phil is right when he wrote that Halo was 'the best telepresence system in the world." My Collab team at Intel did 'bake-offs' for three years (2003-2006) of all the extant systems. HP was far and away the technology and user preference leader, for several reasons that I'll cover below.
This should not have been surprising. HP was among the first to invest heavily in Computer-aided Teleconferencing. That story has only lightly been told in other venues, and I'll just briefly recount it here. But the sequence is fascinating, not just for HP's prescience and then failure, not just once but twice, but also for other companies (notably Intel and Cisco) trying and failing almost precisely for the same reasons. So there must be something endemic about what Phil has described in terms of overlooking the obvious.
Phil was wrong, though when he says "Cisco understood that it was Access." NOPE, they blew this one bigger than HP, especially because they thought "Access" meant for a select few, not everyone. So they chased a small company that did think "access" and Cisco proceeded to completely miss the target.
First of all, although Phil doesn't mention that the 'chief competitor' to Halo at introduction was Cisco's Telepresence, they were the two "big kahunas". Here are pictures of each, Halo first. Note that Halo includes a big screen at the top, whereas Cisco's room does not. That is one part of the story.
The HP "Halo" 'room'
The Cisco "Telepresence" room
Both systems were backed, actually stimulated even and then strongly promoted, by the respective CEOs, Carly Fiorina and John Chambers. Each system cost roughly $350,000 per 'room' with a hefty ongoing networking and maintenance bill as well. And each system had to connect to a similar systm elesewhere in order to establish a "meeting". This meant that there was a premium on getting many systems installed so that meetings could happen between folks you wanted to have meet. Also turned out hard to call a Cisco room from an HP room, or vice versa.
At that price, not everyone could afford such a proposition, and if you did (Cisco ultimately had 100's of 'in-house' systems for their own employees, and for their extended sales offices; HP's 'in-house' sites never totalled more than 100). Cisco once boasted nearly 1,000 corporations having a system; HP never posted installation numbers, but projections beyond 100 other corporations were scarce.
There were two significant differences between HP's technological solution and Cisco's. First, HP had the additional screen, so that "data" could be shared as well as people's smiling faces. Some Cisco installations had small (21") screens at the far right and far left of the room to try to show the same thing (ZERO visibility for anyone interested in numbers). Turns out Chamber is dyslexic, and he insisted that he wanted to see and feel the 'human' in the communication--data was largely extraneous (which by the way is how he mostly managed).
The more fundamental issue was that Cisco's system was optimized with cameras that all focused on the center seat. This had the unfortunate side effect that someone sitting to the left of the center seat could not address the person sitting directly across from them, nor could they address people seated at the far diagonal screen, without each Virtual group turning their gaze away from, rather than toward, the person addressing them from the side screen. It was the ultimate "command-and-control catbird seat" to be in the center. HP's system, correctly positioned with more strategically thought-out cameras, allowed anyone in a seemingly face-to-face screen set-up to talk to that person.
The spec sheets didn't reveal any of that, but users quickly noticed this aberration. Here are classical specs--unbelievably great and analogous technology for each system, except for these two deviations:
Cisco® TelePresence is an innovative new technology that combines rich audio, high-definition video, and interactive elements to deliver a unique, “in-person” experience—over the network. It is designed to bring users closer to the important people, places, and events in their personal and professional lives. The first application, the Cisco TelePresence Meeting solution, creates a live, face-to-face meeting experience, empowering users to interact and collaborate like never before. The Cisco TelePresence Meeting is purpose-built to be a completely different meeting experience, taking interaction and collaboration to a new level. Because it feels like everyone is in the same room, remote interactions are just as natural and effective as in-person communication, simplifying and accelerating information sharing throughout your organization—thereby impacting your top and bottom lines.
Cisco TelePresence lets you improve productivity and responsiveness by focusing on more interactions with the right resource at the right time. It speeds time to market and enables faster decision making. At the same time, you can improve quality of life by giving employees back valuable time and, more importantly, control over their time with solutions that are as high-performance as they are. Imagine bringing more resources into an account for sales or support, and extending your reach and relationships without ever leaving your office.
Critical to delivering this experience are breakthrough innovations in delivering very high-quality audio and ultra high-definition video at very low latency with limited bandwidth utilization, in addition to the meeting environment itself. The Cisco TelePresence Meeting solution delivers true life-size images, ultra-high-definition video (both 720p and 1080p), and spatial audio, which creates the dynamics of voices coming directly from the participants. This scenario creates a “room within a room” environment that makes users feel as if they are at the same table with people in remote locations. In addition to direct eye contact, the full-duplex, near-zero latency allows users to communicate in real time, catching every comment and every nuance of the conversation. Every expression, every gesture is now clearly visible, whether you are meeting across town or across time zones.
Another user experience was notable for both HP and Cisco 'room administrators' (which were a necessary side operational cost).
THE WRONG PEOPLE are using the systems. Both Carly and John believed that these were intentionally built for their key executives, to minimize travel costs and times, and to maximize their exposure time to key account executives. Instead, between 80% and 90% of the users were 'low-status' folk, from the quality department, or personnel, or manufacturing logistics folk, or even folk who had no official standing but just wanted to talk to a friend or colleague elsewhere. The common characteric of these users was that they usually could not get approval for company travel wherease the top execs had no trouble getting travel budgets approved. Moreover, the top execs, despite their avowed stated beliefs, LOVED the travel perq, and were loathe to give it up. Excuses like "It's just not the same if you're not there to press the flesh" abounded. HP learned this in 1985; and had to relearn it in 2004, because everyone who learned it the first time had retired, left, or died.
The ultimate ACCESS test was, akin to Phil's observations, when Ciscc bought WebEx, in order to capture a larger user community. WebEx at the time was a subscription service, with a clunky interface that Cisco managed to mangle even worse. Eric Yuan was VP R&D for WebEx, and he kept explaining, almost ad nauseam, to Cisco officials what was needed. Totally ignored, he left and started Zoom, with major funding from Cisco's SVP and legal counsel, Dan Scheinman who coincidentally quit Cisco the same day. Scheinman's cousin provided VC funding later, and became a long-standing Zoom Board member. Zoom's first customer was Stanford, where MediaX and Terry Winograd in CS had both been advocates.
Note: I participated in three 'firsts' here actually. A. Running MediaX where we did a fair amount of Zoom pre-practice in 2011. B. I was actually Webex's first user-customer for Dialogic in 1997--when they still sold the software as a product. In fact, I wrote three papers about Video conferencing systems and a major C-Suite study (see below). C. I helped define HP's first efforts, and thereby authored Cisco's first commercial sale, using their early router to develop our HP Internet in 1985. The story here (which I did NOT know for another 25 years) was that not only was my signature their first sale, but HP was their ONLY commercial customer for 27 months (when Boeing, thanks we think to Dave Packard telling Ted Wilson and Frank Shrontz how well it worked) became Cisco's 2nd commercial cusstomer.
Here are several studies that I published later via Intel and InnovaScapes Institute
(2) https://www.researchgate.net/publication/287205434_Confronting_Culture_Clash Published in Research Technology Management, 2015. The specific HP Internet development in 1984-1985, and the cisco systems role (early, the name was not capitalized) are well covered here. This paper describes the same "Wrong people using the system" phenomenon already discovered by 1987.
(3) https://ieeexplore.ieee.org/document/5230814 "It depends on your point of View" January 2009, IEEE Signal Processing Magazine, and IEEE Xplore.
(4)
https://doi.org/10.1109/HICSS.2005.3 January 2005, HICSS 38, "Information worker Tools selection, adoption, and evaluation, Chuck House. This paper explicitly mentions WebEx by name, and shows extensive graphs re value of various capabilities.
A couple of other historical points could be made. Intel, with Pat Gelsinger as the lead designer and Exec, along with Andy Grove, bet heavily on Workstation-based Video teleconferencing with Team Station. See the clear article by John James W. published June 11, 2024
https://www.linkedin.com/pulse/intel-teamstation-visionary-leap-collaboration-john-james-worrall-1rane/
I bought several of these for Dialogic in 1998/1999, and they were incredibly painful to use compared to the G-job terminals we had buiit for our HP system in 1985. Not the least issue was the "look 'em in the eyeball, not the eye" problem of the off-axis camera. But they did work, and though pricey, at $12,000 per terminal, they were lots cheaper than a $350,000 'room'. This was a partnership with PictureTel, who while the leader, had some primitive tools at the time (and a number of the old HP guys). The goal was to extend this into cheaper PCs, and ubiquitous 'telephone directories'. Alas, the networks weren't yet up to it, and in particular, the advantages of IPv6 network protocols weren't yet available. Nor were multi-core chipsets, so that any slow node could be isolated from others rather than dragging all concurrent users to a depressingly low level.
One more sidenote--we now know that the off-axis camera and lack of 'seeing you eye to eye' has largely ceased to be a problem. It built mistrust in early systems, but today most people really don't look you in the eye--its body language, or room topology, or voice tonality that win the day
Okay, okay, so much for rambling through a long and tortured history. But think where we'd have been in the COVID days if Zoom hadn't existed.
No comments:
Post a Comment