|
[Dallas,
Aug 20th 2000]
With
GPRS on its way, the wireless internet industry is gearing up to
switch to a pay-for-what-you-download style of billing. This model
is being pioneered in Japan by NTT DoCoMo's i-Mode service. As the
mobile phone carriers rub their hands with glee at the prospect
of increased billing, what does paying for what you download mean
for usability?
Once
again usability is under threat from the suits and their bottom
lines. With current WAP systems there is still very little revenue
being generated from transactions. In fact a recent Gartner
report has suggested that there may be little revenue to be
made by content providers for the wireless internet. So the only
revenue being generated with WAP is the call revenue from actual
phone usage while browsing.
When
the switch comes to GPRS, a packed switched system which is on all
the time and allows you to make voice calls without interfering
with internet browsing, then it is no longer possible to charge
by the minute or the second. The DoCoMo i-Mode system already works
this way. The Japanese subscribers are paying for the data that
they download, calculated in a "so many cents per kilobyte"
fashion. This model is working for DoCoMo, the carrier and service
provider - at least for now. Why might it fail in future?
Perceived
Value
When
I make a 10 minute call to my mother, I perceive that there is a
value in connecting us together by voice. I also perceive that the
length of time that connection is made has a value. I have a perception
that I am using a resource that others could be using. I rent that
resource for a period of time. If my call runs to 20 minutes because
I have more to say or more to listen to, then I perceive that I
must pay for that added value.
It
is marginal, at best, whether I have the same perception of a WAP
internet connection. Currently, I am paying for the time of the
connection, but I have no concept of exactly how that works, necessarily.
Does the time start when I commence browsing at the service home
deck, and does it finish when I somehow logout, terminate the call?
Probably. Why I might ask doesn't it connect, download a deck, and
disconnected invisibly and I pay only for download time? Indeed!
Now
let's look at the perceived value of a packet switched, "on"
all the time connection. Suddenly I have the impression that I have
my own personal channel on to the internet. I presumably pay some
monthly subscription. If this is "on" all the time, then
why might I have to pay more when it's in use versus not in use.
I have surely rented it for the month? This is the first problem
that the carriers will encounter. There is no perceived value for
data downloading.
You
might try to use an analogy or metaphor. You could say, well it's
like being connected to the water main, you pay a subscription for
the connection and you pay a metered rate for the water that you
use. Well that might work. However, water is a very very tangible
commodity. How do you get users to perceive data as a tangible commodity?
Try
asking a typical man in the street with a WAP phone, "How many
bites are in a typical web page?", "How many bytes are
in a typical WML Deck?" How many people have even the remotest
idea of these quantities? In truth it takes an expert to examine
a web page and know within a reasonable error, just how many bytes
it might contain. Exercise: How many bytes are in this web page,
including graphics?
Useful
data, useless data, all the one price
Now
let's consider the issue of charging for the quantity used when
you are metering a data stream and not a water stream. With water,
you know what you are getting. In most countries it is drinkable
and it is all the same, from one day to the next, one minute to
the next. With data this is not the case. Not all data is drinkable,
some you might want to discard as dishwater.
For
example, let's say that you want to book a holiday. If you walked
into a travel agent, you might browse through some brochures, study
some maps, talk over with the agent and others in the store about
possible destinations. You might discuss airlines, airports, rental
car companies. You hear both good stories and bad. You hear about
the drinkable and you hear about the dishwater. All of this leads
you to make a decision. The agent then makes his money from the
transaction. The agent takes a chance, a calculated business risk
that you will make a transaction and that his time will not be wasted.
Now
let's say that the agent put a sign on the door saying, $10 per
month subscription for store entry, $0.10 per brochure page read,
$0.20 per opinion sought, $2 per flight, hotel, car, availability
checked. And so forth... How then are you going to deal with booking
your holiday? And how does it feel as an experience?
You
are going to be very very cautious about what you do and what you
ask, and you may well miss out on a great deal, great location,
or better airline, because you failed to ask. The value of the service
is depreciated.
However,
by charging by the byte, this is exactly the kind of experience
that the operators want to deliver to you.
Open
to abuse
A
system which charges you by the byte is heavily open to abuse. Many
web designers now spend a lot of time optimizing their pages and
content. HTML is compressed to remove white space; HTML tags are
minimized - there is currently a lobby against the <font>
tag for example; graphics are compressed; use of graphics are minimized;
style is often flat to facilitate better compression; information
architecture and interaction design are done to minimize numbers
of screens and forms to be filled. In short, a better experience
has a lot in common with less data, rather than more.
Already
we are seeing abuse of these "less is more" principles.
Online magazines are increasingly opting for shorter and shorter
pages. Some have reduced articles to as little as 5 paragraphs per
page, forcing the reader to navigate through 5 to 10 pages to complete
the whole article. The reason for this is simple. The website owners
want more chances to push advertising at the reader as they are
being paid per page impression. Page size has nothing to do with
it. 1 page of 100K is only 1 page impression. But 5 pages of 20K
is 5 page impressions. This might be fine if advertising used 0
bytes of data, but of course, it doesn't. The 5 pages of 20K each
will probably consume 125K in total to deliver the same amount of
content. That's a 25% premium.
As
a user, do you want to pay 25% more for your phone bill because
your favorite online magazine wants to exploit you with more advertising,
and as a side effect your mobile carrier makes a bonus 25% revenue?
Probably not!
So,
I hear you cry, the user simply switches to sites which use less
data. Hmmm!
Switching
to data-lite providers
Earlier,
I pointed out that most people have no perception of how much data
is involved in any given web page or WML deck. So how are they going
to know.
For
example, two web travel services, astronomer.com and sword.com offer
flight, hotel and rental car bookings online. It turns out that
astronomer.com has a really great web designer and the average page
is only 50K and the average number of pages to complete a transaction
is only 15. Meanwhile, over at sword.com, they saved money and got
a junior web designer, the average page is 100K and the average
number of pages to complete a transaction is 20. The sword.com users
will pay more than twice as much to achieve the same goal. However,
they have almost no way of determining that they have been cheated.
A
scenario which might facilitate detection of sword.com's hidden
costs is hard to imagine. Two people sit together on a plane, both
have GPRS enabled PDAs. One says to the other, "Do you use
sword.com?" The other replies, "No, I use astronomer.com.
How do find sword.com?" and so on until they say, "Let's
compare our monthly bills". Not very likely is it?
So
now let's just make one last little step of conjecture and say that
the carrier who signs a deal with the service provider to put sword.com
on the home deck, says "We want the average transaction to
require no less than 1 Megabyte of data. That way, we can guarantee
to make $x.yy per transaction" and they make that a condition
of the home deck contract.
Or
reversing the scenario, sword.com sales people approach the carrier
and say, "Our service uses 125% more data than the competitor's
to complete a transaction, that's more revenue for you guys!"
Which provider is the carrier going to pick?
Language
Discrimination
Finally,
we have to look at language issues. Anyone who develops help text,
manuals or internationalized UI designs, will tell you that other
germanic-roman languages use more space than English. German is
often quoted as the example, and can need up to 50% more space.
So German speakers receiving their content in German will pay more.
This
example raises some interesting points of law. For example, in Canada
will it be acceptable that French speakers pay more? In the United
States will it be acceptable that Spanish speakers will pay more?
In Switzerland, is it acceptable that German, French and Italian
speakers will all pay different rates for the same data?
Where
does that leave usability?
It
is the opinion of this webzine that "by the byte" charging
models are doomed to failure for nonessential or non-compelling
content. The end user will curtail usage to save on the bill. They
will, however, continue to use essential and compelling services,
just in the same way that everyone continues to use water, even
when it's metered. In order to prevent abuse of such systems, consumers
associations will need to be vigilant and consumer reporting magazines
such as Which? will need to
act for the consumer, trying out services, metering the bytes and
reporting the true costs and value for money. The end effect of
this is that it will raise awareness amongst the general public
of design and interaction issues. Issues which they never needed
to be aware of. With adequate reporting and public information,
usability may indeed make a comeback, until then packed switched
mobile internet will continue to "byte your cents off".
|
