|
The
mantra of the last 15 years has been "leave the mainframe alone".
The only legitimate excuse for changing legacy systems was Y2K compliance.
Now that we're safely into the 21st century, conventional wisdom
would have it, that you can put your old COBOL programmer out to
grass. Maybe not! Some legacy systems are seriously inhibiting the
information age user experience. It's time to change the big iron!
The
last ten years has seen the emergence of network computing systems
with global reach. Collaborative systems, groupware, email, and
more recently intranets, extranets and eCommerce on the internet
have all come of age. For existing corporations these new technologies
have been grafted on top of existing legacy systems. Often the joins
between the new and the old are more like cracks, patched with a
plaster of middleware and message queuing software.
As
we move to an age of wireless internet information appliances, the
middleware wrappings around the legacy transaction processors will
no longer be sufficient. Businesses who want to compete for information
age customers must bin the big iron and start again.
The
problem lies with the ( not unreasonable ) user expectation of instant
gratification. The user believes that if they can contact anyone,
or anything, live over the internet, anywhere in the world then
they should be able to get answers and information, live, anytime,
any place. Afterall, if the data is there, why is it not available?
Billing
Systems
With
older batched, overnight, transaction processing systems, instant
responses just aren't available. Batch processing systems were often
the only practical solutions for large volumes of transactions in
the late 1970s and early 1980s. If you were processing phone calls,
or banking transactions, it wasn't possible for live on-line transaction
processing, or if it was then it was too expensive and arguably
unjustifiable on pure business need.
In
the web world, that overnight delay becomes problematic. The user
believes that when they do something from their web browser, the
server at the other end receives and processes it immediately. That
isn't always the case. Sometimes, it doesn't happen until tomorrow.
For
example, let us imagine that a wireless telephone company has provided
a customer website for viewing phone bills, perhaps checking airtime
packages and viewing promotional giveaways such as free minutes.
The quite reasonable expectation is that the user can make a phone
call to a friend, then a few minutes later log on to the web site
and check the number of free minutes remaining. The user expects
the call they just made to be part of that calculation. It won't
be. The free minutes displayed will be the number of minutes remaining
after the last call they made the previous day, before the arbitrary
point in time when the batch feed was made to the billing system.
The user, of course, has no concept of this. No concept that the
call routing, mediation and recording systems are separate from
the billing system and that the billing system is a batch feed system
and not an on-line transaction processing system. All they know
is that the system doesn't change instantly after they make a call.
It's
a bad experience for the user. It leads to superstitious behavior.
Airline
Systems
Now
let us consider another example. This time it isn't batch processing
that is the issue but the old problem of on-line storage, disk space
and RAM. Many older airline seating systems were only capable of
working on flights departing within the next 24 hours. So seats
had to be allocated at departure time. This is still common with
many airlines. Those who have made a multi-leg journey around half
the globe will recognize this. Often the final leg of the journey
is on a flight more than 24 hours in advance. The airline cannot
issue a boarding card for that final leg at initial takeoff time
because the final leg is still more than 24 hours away. If they
issued you a boarding card, it would be for the current day, 1 day
too soon. The reason for the 24 hour limit was simple - storage
space, or rather the lack of it.
It's
clear to see that airlines can improve the user experience by issuing
boarding cards further in advance and further they can improve business
responsiveness if they could issue seat allocations at the time
of booking (or at least several days in advance).
Again,
in this example, there is no reason why a better user experience
isn't possible, if only the back end systems could support allocating
seats against any flight, on any day. All it needs is a better data
model and more storage. Both of which are easily possible with today's
technology.
By
hanging on to legacy systems with an out dated architecture, large
corporations are affecting the user experience with their web presence
and affecting the customer experience with their business. In industries
which have traditionally been the big customers for relational database
technology such as banking, telephone services and airlines, the
legacy systems have to be replaced.
Times
are changing
There
is evidence in the United States that some of the big businesses
are beginning to "get it". Big legacy systems are being
phased out or have already gone. For example, I flew with American
Airlines last week on an electronic ticket complete with a seat
allocation which was issued electronically, delivered by email,
over 1 week in advance. That is a good customer experience with
technology. It is also good customer service. So, ultimately, it's
good business.
There
are also rumors that the big telcos in the US such as AT&T and
Sprint, are working to replace their billing systems. There was
another motive for replacement. The legacy systems could only cope
with 2 decimal places i.e. they could bill you to the nearest cent.
The need for micro-payments primarily in the wireless web world,
has necessitated a move to a new system which can cope with more
accuracy, at least 4 decimal places i.e. 1/100th of a cent. The
upshot of the billing system replacement is that it is possible
to move to a real time on-line transaction processing system, simultaneously.
What does that mean for the end user?
Firstly,
the earlier problem described above goes away. You want to know
how many minutes are left on your monthly plan. Well now you can.
Just query the web site and ask it. The answer will be calculated
based on the last transaction you made and will be up-to-the-second
accurate. That's a better user experience. That's better customer
service. That's better business.
Secondly,
by finally taking the plunge and replacing the old box and the old
software, there is a far greater chance that other maintenance functionality
can be added. For example, how about pooling of mCommerce transactions
whilst roaming in an area which would normally incur a foreign exchange
conversion on each transaction. With a new system, designed for
expansion and almost certainly built on object technology such future
changes are much more likely. So future improvements in the user
experience are more likely.
Interaction
Design must influence back-end architecture
As
more and more eCommerce and eService moves onto wireless devices
and the age of ubiquitous, pervasive computing is upon us, the need
for instant response to the user who seeks instant gratification
will be greater and greater. If businesses in the banking, investing,
credit card, communications, travel and leisure sectors want to
impress their customer base with a better user experience, then
it is time that they dusted down the COBOL programmers and set them
the task of jazzing up the big iron. Or better yet, just bin the
digital dinosaurs in the climate controlled glass house and put
in all new systems instead. New systems facilitate flexibility.
Interaction
Designers as a group have tended to shy away from conflict when
back-end systems are involved. Wishing to pick their battles carefully,
they have focused on grafting as elegant an interface as possible
onto the middleware wrapped big iron. Well, it's time to drop the
silence. A better user experience needs better interaction design
which in turn needs real time transaction processing, across larger
data sets than has ever been seen before. Nothing less will satisfy
information age man, preprogrammed with attention deficit disorder.
In
the Information Age, the user experience IS the customer
experience. Make your customer experience better - put the big iron
in the bin!
|
