Microsoft HealthVault

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

~ the goodbye ~

healthvault launched in october 2007 as microsoft's flagship consumer-health product. the bet was straightforward in 2007 terms. every american would eventually have an electronic medical record portable across hospitals, and there would be a market for a neutral consumer cloud where patients could pull all of their data into one place. healthvault was meant to be that cloud.

healthvault was a us product designed for a specific kind of patient: the kind who has a primary-care portal account and a continuous record of digital prescriptions. that is not a turkish patient. our health system was paper-and-stamp through almost the entire decade healthvault existed, and the turkish e-health portal that finally digitized prescriptions came years later with its own stack that never spoke to anything microsoft made. healthvault is therefore one of those products i know about as an engineer who reads health-tech press, not as an actual user.

healthvault never reached the user numbers microsoft was banking on. by the early 2010s the consumer health space had moved toward mobile-first wearables and apps, where apple and google were shipping platform-level frameworks that captured the same data with no upload step. healthvault required you to actively maintain it. you logged in and you typed your blood pressure into a form. that workflow does not survive contact with people. microsoft announced the shutdown on april 5, 2019, gave users seven months to export their records, and turned off the servers on november 20.

~ the chronicle ~

~ what it was ~

the actual healthvault product was a personal-cloud database with an api. you could connect a withings scale and watch your weight populate the dashboard. you could connect myfitnesspal and watch your meal log show up. some hospitals built migration tools into healthvault as a way to give discharged patients a copy of their record. microsoft band data went there too, until microsoft killed band in 2016. the api side was the part that actually saw use. the consumer-facing dashboard never had real engagement.

~ the day it died ~

the april 2019 shutdown notice was unusually clean by microsoft standards. it gave users export tools and a partner data-migration option through allscripts and get real health. data not exported by november 20 would be deleted with the servers, and that is what happened. there were no protests and no last-day memorial threads. the user base was mostly hospitals and partner companies that had been migrating off for two years.

~ what it became ~

healthvault is now a footnote in the longer story of why nobody has built a successful consumer health record. apple's health app effectively took over the same data layer with hospital-portal imports and continuous device sync, while google fit took the wearable side. microsoft moved its own remaining health work into the cloud-and-enterprise side of the business under nuance and dragon, which is the dictation product that hospitals actually pay for. the consumer never wanted to be a database administrator for their own body. that was the lesson, and it took twelve years to learn.

@}-->--

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

« previous  ·  random  ·  next »