Gary Wolf

March 8, 2010

Last Tuesday I received an invitation from Esther Dyson and
Jen McCabe to attend a small, private workshop on the business side of
user-generated health. The workshop was to be held the very next day. Despite
this short notice, more than 30 people showed up, some of whom flew across the
country to attend. Their question was: how could they make money from this?

I think it’s an important question. Business models are interesting as an informational signal: each model
represents a theory about how the future of self-tracking will unfold. I am
working on a post that maps out some of the commercial territory I see being
explored by the more than 200 companies that have been mentioned at the QS
Show&Tell meetings or here on our blog. But before too many days go by, I
wanted to share some of the discussion that took place at IDEO last week.

The session began with McCabe of ContagionHealth
describing her company, whose first product is a social game that allows people
to exchange exercise challenges. What on the surface looks like a fun diversion
and micro-motivator is actually an insertion strategy for new tools of mapping
human psychology and social influence, McCabe explained. The group is already
experimenting with segmenting users (for instance, into “doms” and
“subs” according to whether they prefer influencing or being
influenced). This is invisible to the user, who sees only an invitation to play
with their friends. Right now, McCabe and her partner Andrey Petrov hope to make money
by licensing the platform to employee wellness programs.

While ContagionHealth has a viral strategy, Vicki Saunders  of Zazengo described a corporate
approach. Zazengo is a set of tools designed to allow large organizations to
orchestrate and track campaigns among their employees and supporters. Zazengo’s
key tool is the “Actvertisement,” – this is a message delivered to
some targeted subset of employees, usually through the company intranet. The
actvertisement urges the employee to take some specific action, associated with
the company’s goals. (As part of a sustainability campaign, for instance,
employees can be urged to turn off their computer when not in use.) By creating
detailed feedback that tracks employee and departmental contributions at any
level of the corporate hierarchy, Zazengo can be used to stage competitions and
orchestrate rewards.

WallMart is currently one of Zazengo’s clients. WallMart
started using the platform for sustainability campaigns, and now is beginning
to use it to orchestrate health and wellness campaigns. One of the interesting
points that Vicki made as that the online component of these campaigns is
strengthened by the face-to-face meetings that WallMart holds with its
employees every day. Individual tracking, here, is understood as part of a
collective contribution to a corporate goal.

While making money depends on having a good business model,
there’s no business to model unless something is happening that people like and
want. Why do people track things? What kinds of systems and messages will be
useful for people tracking their health?

James Park of FitBit has an extremely simple model: he sells
a device that people pay for. The FitBit is a pedometer and sleep tracker, and
though Park acknowledges that there are plenty of other accelerometer-based
trackers out there, his company is trying to take a slice of the market by
offering a very appealing little device. FitBit also has a web site site where
the activity and sleep data can be combined with  food diary and calorie counter. “In the United States
there are 38 million members of weight watchers,” Park said.
“Globally there are 60 million. What people are missing is the input
piece.”

This lead to a short conversation about devices. Nobody from
Philips was there, so I described the Philips DirectLife device, which is very
small. In fact, it was too small for me. I lost mine immediately. And, speaking of
business models: the web site that goes with the Philips device has a monthly
$12 fee, and when I cancelled my subscription after losing my device they
sent me a replacement free.

The accelerometer and chip components of both the Fitbit and
the Philips devices, both of which retail for about $100, cost just a few
dollars. The FitBit, helpfully, has a small display. The Philips DirectLife
does not. Still, these are not expensive machines. If users find value in
self-tracking, there is plenty of room to make money.

Raj Mehta, who wasn’t at this meeting but is an interesting
source of insight into how technology can help people manage the complexity of
their lives, believes that there is a market for a system that can capture any
kind of self-report. His device, the Zuri, contains a voice recorder. You talk
into it, and your reports are captured, sent to a transcriber, and then
coded into a standard format for comparison and review. This is classic way of
going forward by going back: we ought to have fast, cheap, natural language
processing of our voice reports, but we don’t. Transcription is the next
best thing.

Whenever these topics come up, there is some back and forth
between people who guess that a general system will inevitably come to dominate
the market (probably a mobile application that takes advantage of the phone’s
on board accelerometer, GPS receiver, and ready accessibility for other kinds
of self-reports), and others who guess that there will be a large set of
different tools united by some data standards and privacy guarantees.

I suspect the latter answer is correct. Generality is going
to be necessary at some level, because what matters to us shifts over time, and
we may want to go back and query our data in light of new ideas or new conditions.
But this generality will probably be delivered as another layer on top of a
rich diversity of services and devices. If this vision is correct, there are
probably good businesses to be had in building the infrastructure that will
support the rise of personal data systems in health and elsewhere.

Carol McCall of Tenzing Health, and formerly a Vice
President of Humana, has deep expertise in health costs, and she tried to give
the group a quick understanding of how large the savings from behavior change
could be. The health care costs of being overweight are $19.36 per year for
each overweight pound, she said. The costs of other significant conditions –
hypertension, high cholesterol, and diabetes – are also very high. (Here is Humana
press release from December 2008 with these health cost numbers.)
McCall mentioned a diabetes prevention trial that showed how aggressive
lifestyle interventions could prevent what would have been an inevitable onset
of diabetes. Provocatively, she said that the interventions by traditional
means do not pay for themselves, because they are so “high touch.”
But if by supplementing human interventions with semi-automated systems of
tracking and motivation the cost of behavior change could go down, the health
benefits to the individuals and the savings to society could be very large.

Bernie Plishtin of Sagebrook Partners made a brief comment about business models
that charge companies a percentage of the amount saved going forward. I made a
note of this as a potent suggestion, and did a bit of looking around afterward.
A good sense of this model in the energy sector can be gleaned from looking at
Metrus Energy.

Dave Marvit of Fujitsu Laboratories of America mentioned
that in Japan there was legislation pending to set fitness targets for
corporate employees, with penalties if the average size of an employee grew too
large:

“The law that went into effect April 1 forces companies and government,
the two sources of health insurance in Japan, to measure the body fat of
employees between the ages of 40 and 74.” (From Workforce Management, 7.3.2008)

Marvit pointed out that
while the focus on costs of unhealthy lifestyles often addresses only direct
costs of health care, the indirect costs to companies (productivity,
absenteeism) are also high, and may even equal or exceed the direct costs.

 In the context of talking about what motivates people to
exercise and to adopt more healthy behaviors, Este Solomon Gray talked about
her experiences in the fitness industry, with a focus on women. She explained
that for many people self-awareness is not only a method to reach a goal, it is
itself a goal. “Paying attention to how you are feeling is a daily,
simple, profound thing. For women, they don’t want scales yelling at them, or
exogenous voices in their lives. They say, ‘I already know what I’m not doing,
but the issue is that I’m not the only person in the world. I’m responsible for
all these other people. I don’t need to be told that Istopped working out, or
forgiven for not doing it.” There is a brief summary of some of Gray’s
thinking in this slide show called Sociality Inside from the Texting 4 Health conference.

(This contrast between tracking as a means to an end and
tracking as a mode of self-awareness is something I’ll return to below.)

At some point I asked whether anybody had looked at
government and non-profit focused business models. At first there was a general
shaking of heads, a  few comments
about the difficulty of finding money in government work. But then, almost as
an afterthought, a few people spoke up with a different point of view.

David Fore described Lybba.org, a nonprofit whose mission is to
alleviate the suffering of people with chronic health conditions. Lybba is
partner on a multi-million dollar NIH grant to Children’s Hospital Medical
Center in Cincinnati. I quote from the research grant here, because it is
inspiring to think about from the perspective of QS:

Chronic illness kills too many and costs too much. Reducing
this burden is possible, yet within our current health care system, Americans
receive about 50% of indicated care and patients follow doctors’ orders about
as often. What if there was a way to create a vastly better chronic illness
care system by harnessing the inherent motivation and collective intelligence
of patients and clinicians? What if this system allowed patients and physicians
to share information freely, collaborate to solve important problems, and use
their collective creativity and expertise to act in ways that improve health? A
system for transforming chronic care is possible – we intend to build it by
designing, prototyping, optimizing, and evaluating a clinical collaborative
care network (C3N)

People like interacting with nonprofits, Fore said,
“because they believe that you don’t have nefarious reasons for gathering their
data.”

Khaled Hassounah of MedHelp was standing in the back of the
room for the entire time and may not have been much noticed, but he has a lot
of data on self-tracking and health. MedHelp has millions of users, mostly
doing very basic community sharing and mutual help around specific health
conditions. For the last couple of years, MedHelp has been offering tracking
software to its users. I will have some stats in my upcoming New York Times
magazine story about the culture of the Quantified Self, so I’ll save them for
now, except to say that the numbers of people using MedHelp trackers is quite
large. MedHelp supports itself through advertising.

While there was a lot of cynicism – much of it deserved –
about advertising models, and in particular about taking pharmaceutical
advertisements, one good thing about advertising models is that they force you
to develop “organic traffic.” That is, people have to actually like
and value what you are doing. (That’s leaving aside the question of black hat SEO, an obsession of mine at the moment, but a topic for another time.)
Khaled also said that his company had recently been working with the state of
Washington to do something related to health and wellness for state employees.
He didn’t go into details, but he reminded us that the government is an
employer, too, and a very big one.

The day ended with a short discussion of privacy, ethics,
and the future architecture of personal data. John Mattison, the Chief Medical Information Officer of Kaiser Permanente Southern California talked
about the difficulties of designing a personal data architecture in the face of
both institutional requirements and regulatory ignorance.

Mike Kirkwood
of Polka tried to give a quick
philosophical “reboot” on the privacy topic, saying that we should be
able to hold on to our data, but allow others to ask it questions. You don’t
have to give your health care provider unlimited access your bits. Instead, if
things are stored in a standard format, you can issue permission for somebody
to query your data, without copying everything (or anything). You can even
cancel your permission later, and the reports that had been generated about you
will shrivel for lack of data. This is a very important notion, quite hard to
get your mind around at first. I think of it as a “personal data server in
the cloud,” and I think of it that way because of a long conversation I
had with Kirkwood many months ago. I don’t know of any single document where he
lays this out in detail, but for the technically-minded, here’s a recent essay
he wrote about “Identity, Health Care, and a Higher Realm of Credentials.”

From a VC perspective it is still very early to make
investment decisions in this sector. Andrew Braccia, of Accel Partners, acknowledged
this when Dyson pressed him to respond to the assembled entrepreneurs as a VC.
He was their target audience. What would he invest in? Braccia did not express
enthusiasm about any particular approach. (Of course I wouldn’t
necessarily  have expected him to tell all,
in this setting, even if there are some ideas percolating at Accel.) But he did
say something that connects with Mike Kirkwood’s perspective on security and
privacy: The rise of personal informatics takes place in the context
of the rise of cloud computing. He mentioned Accel’s investment in Cloudera. Where the focus is on the
rational deployment of capital in the face of an interesting, early stage trend
whose final “consumer friendly” form is still very unclear, the
enabling technologies are always going to seem the most attractive to VCs.

(There are VentureBeat news stories about Cloudera with some background here and here.)

This meeting overlapped quite a bit with QS topics, but of
course not completely, as many QS folks track things other than health data;
while health folks are interested in getting people to make better choices
whether or not this leads to (or stems from) greater self-knowledge. So, since
this blog is about self-knowledge through numbers, let me make a few editorial
comments from that perspective.

After thinking about it for a few days, it seemed to me that
most of the health ideas presented at this meeting could be placed somewhere on
a two dimensional grid. The x axis runs from individual to social. The y axis
runs from consumerist to utopian.

 Here’s what I mean by consumerist: products and services
that emphasize clear, understandable value over exploration and insight. The
good thing here is that they can be effective, easy to use, and have to
potential to become wildly popular. The drawback is that there will inevitably
be a lot of junk, potential burnout, and, more importantly, a lack of ability
to solve problems that aren’t definable in marketing terms.

Utopian things, on the other hand, address human needs that
are still poorly defined but strike their inventors as urgent for human
survival and happiness. They appeal to our sense of discovery, and can open
entirely new prospects and opportunities. The downside is that they are hard to
understand in terms of current markets, and may languish as mere eccentricities
or geek fads. 

Utopian technologies derive their energy from
consumerist ones. Consumerist technologies tend to be bullying; whether this
bullying takes place through conventional marketing, through employer mandates,
or by social network enthusiasts who never miss an opportunity to tell their
friends “oh, you mean you don’t twinglefish?” Everybody can sense that there is a bit of a pyramid scheme going on here, enriching the inventors of these technologies and the early adopters at the expense of the latecomers who expose their habits to a system well tuned to push them around. In this context, Utopian ideas about personal informatics function as a kind of self-defense strategy: how can I use this data to understand myself at least as well as the people who are collecting it understand me? Self-awareness is a counter force against social manipulation.

But consumerist technologies depend in their own way upon Utopian ones.
In the case of the Quantified Self, the promise of a new form of
self-understanding, a frontier of human optimization that we might steer
towards, makes it seem worth paying the price (in risk and in work) of using
the tools of today. It’s worth it because we are all trying
to peer ahead into tomorrow, and get a hint of what kind of people we’ll be. If the Utopian promise rings hollow, the consumerist products will struggle.

Most business models run
along the bottom of grid: consumerist products serving a range of individual
and social needs.

There is still plenty of room at the top.