Dealing with Technological Change in a Political Environment
Richard E. Zeller
Florida Highway Patrol
This presentation provides an update on new technologies being implemented by the
Florida Department of Highway Safety and Motor
Vehicles. Three major changes have been
initiated in the last year: installation of mobile data computers in the Florida Highway Patrol’s
patrol cars, outsourcing the data entry of crash reports and switching from microfilm to
electronic imaging for arc
hiving them, and developing a data warehouse for department
databases. Each of these processes has involved the political influence and reaction to external
events that is typical of large
scale government procurement processes. Each also has presented
he department with challenges involving re
engineering business processes and the effective use
of existing department personnel.
Mobile Data Computers
Those of you that attended last year’s presentation will recall that we were just embarking
on a proces
s of installing mobile data computers. That process has taken many strange twists
and turns over the past year. The computers are intended to be used for dispatching, ticket
writing, and report writing. They are also used to check licenses and registrat
ions. They are
linked to the computer aided dispatch system and to FCIC/NCIC via CDPD.
Funding was provided by the Legislature based on the need to collect traffic stop data in
order to study racial profiling. An invitation to negotiate was issued and t
he award was made to
a systems integration company and a reporting software vendor. Unfortunately, neither of these
was the company which had done most of the lobbying to get the funding. The State Technology
Office then decided that the department had r
elied too much on price and not enough on quality.
The award was withdrawn and given to two other companies, including the one which had done
the lobbying. The original winner sued over the cancellation and a settlement was reached by
which all but the o
riginal software vendor would participate. As a result, we have a systems
integrator, a software vendor, and a hardware vendor participating in the project. (Lesson 1:
Never use multiple vendors unless one of them is clearly in charge.)
The next event
to affect the project occurred on September 11. In the panic budget
cutting which followed that event, the Governor’s Office would not release the funds for the
project. This resulted in a three month delay and put the department and the vendors under a
great deal of pressure to complete the first phase of the project by the end of the fiscal year in
Project management was assigned to a person in the Division of Administrative Services
who had purchasing experience, but limited technological knowle
dge. This resulted in the
vendors being able to cut corners and, in some cases, not provide the capabilities that were
originally intended. Design changes also resulted from the under
funding typical of the
When a product was finall
y presented, some of the reports had been changed by the
vendor to the point that they could not be accepted by the department. After the first installation,
troopers were in a position of having a computer with report writing capability, but still having
written reports. This aspect of the project was salvaged when an FHP Captain with
a high level of computer expertise was assigned to manage the project.
Overselling of technology also occurred. When the project was first envisioned, the
department was assured that the use of CDPD would be no problem. Since the dispatch
capability and the link to FCIC/NCIC are dependent on CDPD, this is a critical element of the
project. As the project was being started, the department was informed that
a large area of the
Panhandle does not have CDPD coverage and will not for the foreseeable future. (Lesson 2: Do
not believe all that a vendor tells you.) The State is investigating the possibility of adding a data
channel to its 800 MHz radio system to
overcome this problem.
The good news is that there are over 900 computers installed and the system is functional
in six of the ten FHP troops. Dispatch capability has been delayed by the need to develop an
interface to the computer
aided dispatch system
(vendor number 4). The next task is
implementing a system for extracting the data from the report and transmitting it to headquarters.
This was one of the selling points with the legislature. Having the computers would eliminate
the need to pay for dat
a entry services.
As with all technology projects, the department is faced with the issue of restructuring its
workforce. Clerks who had primarily processed paper documents are now expected to be able to
process and retrieve electronic documents. Field
staff are required to be able to provide
computer support and operate fairly complex equipment.
The needs of the users of the reports tend to get lost in the process. From the troopers’
standpoint, the computer is there to help them write their reports.
They are perfectly happy to
have the computer generate a paper report and submit it, not recognizing that improvements can
also benefit those receiving the data electronically. Being former law enforcement officers, the
software vendor’s approach has tra
ditionally been to generate a paper report which looks like it
was typed on the old pre
printed form. It has been a challenge to reorient them toward moving
data instead of creating documents. Concern with things like being able to create a file formatte
so that it can be loaded into existing databases has been a low priority.
Funding for the second phase of the three year project has been provided and all of the
computers will be installed by the end of this year.
Outsourcing of Crash Report Data Entry
On July 1, 2001, the department contracted with PRIDE (prison industries) to image and
data enter all crash reports. The reports are received by the department, checked, batched, and
then sent out for imaging and data entry. The images are placed in an
line image vault, where
they can be accessed by department employees through an internet connection. A back
of the images is provided to the department along with the data. The data are transferred to our
mainframe computer. Once on the main
frame, the data are processed exactly as they had been
before. The data and images also are provided to the Florida Department of Transportation for
its internal use and uploading of commercial vehicle data.
The outsourcing occurred in conjunction with a
revision of the crash report form. The
new form began to be used on January 1, 2002. The change required some modification to the
mainframe data file and new data entry software.
The original intention was to use a private company for the process inste
ad of PRIDE.
However, the company hoped to make its profit by selling the reports for eight or ten dollars
each. When our General Counsel informed them that, as our agent, they could only charge the
two dollars that the department is authorized by statut
e to charge, the company quickly lost
The whole project was an attempt to contribute to the goal of reducing the size of
government and getting it out of providing services which can be provided by the private sector.
seven data entry po
sitions were eliminated and the employees moved to other vacant
positions. As with the mobile data computer project, the job skills required of the remaining
employees have changed. People who had been handling paper reports and retrieving documents
microfilm now have to develop a level of computer skill which allows them to download
and print images. The system of transferring data from PRIDE to the department’s mainframe
requires monitoring and technical computer support. Where the department had
been getting by
with borrowing programmers from the data center when problems arose, it is now necessary to
have more constant attention.
One area that has not changed is FARS. The processing of fatal crash reports for the
purpose of entry into FARS has
house. As batches of reports are prepared to be sent
to PRIDE, copies of those involving fatalities are sent to the FARS unit. If an update arrives
indicating a delayed fatality within the 30 day window, the FARS staff can retrieve the elect
image of the original report.
The biggest problem was generated by PRIDE not thinking through the whole process
before making a price offer. Once they got into the details of the department’s requirements and
the data quality checks needed, they re
alized that they needed resources not available to them
and more time to develop the system. The result was that a large backlog of 2002 reports built
up as these problems were overcome. The department has just begun receiving data and the
contains reports on crashes up through about the middle of March. The file
contains a little over 40,000 long
form reports. For comparison, there are about 280,000 long
form reports done in an average year.
Data quality continues to be a concern. As w
ith any new system, there is an on
dialogue between the department and PRIDE regarding data editing and process improvement.
Our next adventure will be to incorporate data directly from the mobile data computers to
eliminate the need to print paper
copies and data enter from them.
An additional outcome of September 11 has been to develop a department
warehouse. Interest in tracking down terrorists quickly led to the need to be able to identify
when an individual got a drive
r license or ID card, what kind of cars they own, when and where
they were stopped by the police, and any number of other things that might be gleaned from the
department’s records. Existing databases were designed to support the department’s business
cesses, such as issuing driver licenses, titling and registering vehicles, and dispatching
troopers. Extensive use of the files for other purposes would put a strain on the department’s
Money was added to the existing contract wit
h HN Bull that supports the computer
dispatch system for them to develop the data warehouse. A Terradata computer was purchased
to hold the warehouse and the query software is Hummingbird’s BIQuery. The warehouse will
contain elements from the FHP
aided dispatch files, along with information from the
driver license and motor vehicles databases. Traffic crash data may be added at a later date.
When completed, it will allow the data to be queried without interfering with the transactions
ing conducted on the “live” databases. Frequent extracts will be made during off
populate the warehouse with recent information.
This project illustrates another aspect of doing strategic planning for government. Often,
programs must be develop
ed quickly to respond to a hot topic. This whole project will be
completed within a year of the need being first realized. FHP encountered a similar problem
when the racial profiling issue surfaced and staff was given six months to develop an entirely
w reporting system and database. These quick reactions create an additional problem since
they must be financed from the existing budget. The department is now in the final stages of
developing its budget for the fiscal year which begins on July 1, 2003.
The proposals it contains
were developed without a clue about what the hot topic of January 2004 might be. While the
mobile data computers and outsourcing of data entry were planned well in advance and went
through the normal budget process, projects li
ke the data warehouse, which require quick
turnaround and diversion of resources from other activities, can be expected to continue to arise.
The three projects discussed here illustrate a number of things about planning for
e. First, new technologies are often over
sold or offered before they are
completely developed and tested. Computer software developers in particular have a tendency to
promise that they can do almost anything. Second, when looking to technology to cut
the cost of
government, the public tends to take a short term approach. Immediate savings are expected.
Investment in long term improvements is limited. Third, government programs are driven by the
issue of the day. Fourth, and finally, when new techno
logy or outsourcing is offered a way to
reduce the size of government, a reduction is expected. It can be very difficult to obtain the
resources and people to maintain the technology once it is in place.