Google Wave

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

~ the obit ~

google wave was announced at google IO in may 2009. the announcement video was famous in tech circles for being unusually exciting. lars rasmussen, who had built google maps, demoed wave on stage. it looked like a new kind of communication tool. you would write a message in a thread that was simultaneously a chat, an email, and a collaborative document. multiple people could edit at once. you could replay the conversation as a timeline. there were robots and gadgets you could embed.

the limited preview launched on 30 september 2009 with invite-only access. demand was high. tech blogs wrote breathlessly about it. friends of mine got invitations and tried to drag me in, but i never asked for one of my own. by the time it would have been open to the public, the product was already losing momentum. this grave is written from the outside.

about a year later, on 4 august 2010, google announced wave development was being stopped. the servers stayed up for another two years for users to export their data. the final shutdown was 30 april 2012. wave is one of the shortest-lived google products of any consequence.

~ the rap sheet ~

~ what it was supposed to be ~

the wave concept was: take everything email does, add real-time editing, add multimedia attachments, add programmable robots, add embedded apps, add a replay feature for new participants, and merge it all into a unified communication platform. it was supposed to be the post-email tool.

each communication unit in wave was called a "wave." waves contained "wavelets," which were sub-conversations or sub-documents. participants could be added or removed at any point. messages were not really sent and received the way emails are. they were collaboratively constructed in real time, with all participants able to edit any message.

the technical design was genuinely innovative. real-time operational transformation algorithms (the same family of techniques google would later use in google docs) made simultaneous editing work. the federation protocol was open. the robot APIs let third parties build extensions. it was an ambitious technical achievement.

~ why nobody used it ~

the conceptual problem was that wave was answering a question nobody was asking. email worked. chat worked. shared documents worked. people did not want a unified blob that combined all three. they wanted email when they wanted email. chat when they wanted chat. a shared document when they wanted a shared document. wave forced them into one mode that was simultaneously all three, and that mode was confusing.

the second problem was that wave was social. you needed a wave-using friend group to actually use wave. without people to wave with, the product was useless. and convincing your existing email-and-chat-using friends to also sign up for an invite-only experimental tool was hard. so most people who tried wave tried it once with one friend and never opened it again.

the third problem was the real-time editing. when two people are typing in the same paragraph at the same time, the result is chaos. wave's UI showed every keystroke from every participant in real time. for collaborative documents this is sometimes useful. for casual conversation it was distracting. you would type "hi how are" and another person would type "i'm good" inside your sentence before you finished it.

~ the shutdown ~

google announced on 4 august 2010 that wave development was stopping. the post acknowledged that the product had not gotten the user adoption needed. the codebase was open-sourced as apache wave. apache kept it alive as a project until 2018, then quietly retired it.

google said wave's ideas would influence other google products. some of those promises came true. google docs improved its real-time collaboration partly based on wave's research. gmail later got some inline conversation editing. but wave as a product was dead.

~ what survived ~

the case study in product-market fit. wave shows what happens when a tech product is that is technically brilliant and commercially useless. every product manager learning the trade reads about wave. the lesson is that a unified blob is rarely what people want. people want specific tools for specific jobs.

the operational transformation tech. the work that google did on wave's real-time editing algorithms moved on to google docs and is now also in apple's notes, microsoft's office 365 collaborative editing, and most modern collaborative document tools. that part of wave's legacy is alive and visible everywhere.

the cultural memory of "the next email." every few years a new tool launches claiming to be "what email would be if invented today." most of them fail in the same way wave did. the wave story is the cautionary tale that gets told every time. people remember it specifically because the demo was so good and the product still failed.

@}-->--

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

« previous  ·  random  ·  next »