658 Part IV • The Information Management System
there was a mandate that the system had to be imple-
mented during Year 2000 and the decision from
Ciberion was “Look, we are not going to test the code
on any other machine besides this handheld unless
you want to defer the implementation until later,” we
had no choice. If that was the decision, we had to go
with it. I explicitly communicated that decision to the
project stakeholders. This was, in the long term, going
to create a problem. But, there was this mandate that
we had to implement regardless of the robustness of
the hardware. And guess what, we have an average of
10 handhelds each week that need to be replaced!Upgrading to newer, more rugged handheld devices
was more complicated than it might have appeared to
Gracia. Microsoft released Windows CE 3.0 during the final
stages of MaxFli development. Because the new version was
a major upgrade from Windows CE 2.1, it made fundamen-
tal changes in how MaxFli could interact with the hardware.
The new CE 3.0-enabled handheld devices would not run
MaxFli in its current version. Ciberion directed their efforts
toward making the shift from CE 2.1 to CE 3.0. The Jornada
690 in use by Chiletabacos, BAT Colombia, and BATCA
could not run CE 3.0 and was thus obsolete. Neil Coupland,
a senior vice president for marketing and sales in Ciberion,
argued that, “because of the fact that hardware reliability
was an issue, we had to support new hardware, [and] the old
hardware was now obsolete. You couldn’t buy it. Therefore,
the code was obsolete too. The focus was put on actually
producing a working version on the new hardware.”
Complicating the difficulties with the Jornadas, the
service-level agreement with the supplier who provided
and supported the Jornadas in all six countries was unsat-
isfactory. The average repair/replace order required 6
weeks instead of the contractually agreed 2 weeks. The
delay required an additional 20 percent reserve supply of
handhelds. Eventually, Gracia switched to a vendor in
Costa Rica, who agreed to support all six countries cen-
trally and manage the shipping to the sales reps in each
country.
A third technical issue important to BATCA but large-
ly unaddressed by Ciberion was the “suggested order” func-
tionality. The suggested order routine is intended to provide
the salesperson with a suggested order of brands and quanti-
ties at each outlet. This functionality was one of the major
selling points of MaxFli in the minds of the trade marketing
department. By giving the salesperson an accurate, up-to-
date order history for each outlet, BAT hoped to increase
sales and market share. However, due to technical problems
with the Siebel and Oracle systems, this functionality was
not implemented in initial MaxFli versions. This frustrated
the project teams in Chile and Colombia as well as in
BATCA. Kruger repeatedly pressed Ciberion about the
importance of the suggested order routine. As of June 2001,
the suggested order functionality was still not implemented
in MaxFli 1.1.
BATCA installed multiple servers to manage MaxFli
across six countries. Because of the communication infra-
structure issues discussed previously, the management
team chose to put the servers in Honduras and the IT sup-
port team in Costa Rica. The rollout occurred in three
phases. Because Nicaragua was the strongest market in
Central America, they were selected to go first. Costa Rica,
Panama, and Honduras went live in the second phase of
implementation, and finally El Salvador and Guatemala in
the third phase. Like BAT Colombia, the BATCA MaxFli
project team also chose a phased implementation strategy
for the rollout. Instead of implementing the full MaxFli
functionality, the project team chose to focus first on the
basics of ordering, invoicing, and receiving payments. The
value-added features of MaxFli, like brand coverage, com-
petitive information gathering, and cycle planning, would
be integrated one at a time every other month.
As in Colombia, this implementation decision created
perceptual problems for MaxFli. The system had been billed
as a panacea for Central America. Unfortunately, the enthusi-
astic expectations for MaxFli surpassed its initial capabili-
ties. Because the initial system only handled the basics, it
was essentially a replacement for the relatively low-tech
handheld system previously in use. Without the value-added
features of MaxFli, Central American users felt they had the
worst of both worlds. They had given up the highly
customized sales force automation system they were used to,
without gaining the information benefits promoted for
MaxFli. The morale and enthusiasm for MaxFli waned
quickly. This sentiment was expressed by Morales, who said:I think we oversold this as a tool.... Our business is
very, very simple. And now we’re utilizing this big,
big, very complicated tool. It is silly because we
treat big complicated outlets like Shell, Exxon, and
others the same as the little mom and pop shops. It is
easy to sell this kind of system to Shell and Exxon.
But when it comes to the mom and pop shops, this is
not important. What they’re looking for is that our
salesperson will be there every week, be there on
time, provide the right product, and so on. They
don’t care about all of this information gathering.While initial user satisfaction was lower than in
Chile and Colombia, there was reason to be hopeful. One
of the technical challenges faced by the MaxFli team had
become one of the technological strengths of BATCA. The
MaxFli project team successfully negotiated with several