Lowering The Risks In Developing Do-It-Yourself Software Projects

Mike Dunville* had a decision to make. As the newall-time high. A poll of 200 MIS managers conducted by
operations manager for Alpha Staffing, he wanted toInformation Week revealed that 63 percent of them
make a difference at the dynamic staffing firm. Thehad either eliminated or rebid an IT service contract
principals of the firm had charged him with making thewithin a year. This was typically due to unacceptable
day-to-day operations run smoothly and efficientlyperformance or failure to deliver as promised.Why the
while they concentrated on growing the business inlow success rate of home-grown "Do-It-Yourself"
today's challenging economy.With his background insoftware projects?With a success rate of only
operations and finance in the insurance industry, Mikeroughly 40 percent, you might be tempted to accept
had confidently and successfully translated some ofthe premise that the very complexity of custom
those service practices into Alpha's daily staffing andsoftware would have a bearing on the success or
recruiting operations. Now it was time to take the nextfailure of project. To a degree, this is true, however,
step, and modernize the hodge-podge of computerthe three main causes of software development
systems that until now had been reasonably adequatefailure, as determined by industry surveys, are not
for their three-office operation.But where to start? Herelated to the technology itself, but to experience,
needed a real staffing software package. Mike knewmanagement, and politics. They are:1. Inexperience:
that the custom system his former employer, anTechnologies and programming methods change
insurance giant, had implemented the year before hadrapidly. This means that both business-side and/or
done wonders to improve efficiency and overallcontract programmers and program managers are not
profitability. It had also taken three years to develop,always up to speed on the latest development
and another to implement. In fact, when he had left,technology. What is more important, IT-oriented
they were still making "customizations" -- a process heprogram managers and programmers are rarely totally
figured would go on forever.Mike didn't think that Alphafamiliar with the business issues to be addressed, and
could invest that kind of time or money. Could he couldso may not be able to make the connection to the
bring in some business software consultants and workbest technology needed to address them.Consulting
with them to develop a tailored package in a shortercompanies, including the largest and most well-known
period of time? Mike knew that the continued successof the accounting/IT consulting firms, often use novice
of Alpha Staffing, and its entrance into new marketstalent, fresh out of school, to handle programming and
and business lines would be greatly dependent on amanagement chores. This inexperience leaves clients
successful staffing software systemopen to potentially massive cost-overruns or, worse, a
implementation.This reinforced Mike's doubts that onesystem that never works even after years of
of the "off-the-shelf" staffing software packagesdevelopment.You probably won't hear much about
would fit their needs, even if it were quicker and easierthese consulting firm failures though; both the firms and
to implement. The more he thought about it, the moretheir clients have a vested interest to keep such
Mike felt that his decision on business automationadmissions of failure quiet. Only when it reaches the
would be his biggest, and if wrong, his last. He reached"public scandal" or "write-off" stage does such
for a bottle of antacid.Mike's business is unique, likeinformation become public knowledge.2. Management
yours. Special. Unlike any other competitor in the wholemis-objectives: Any firm trying to develop a staffing
wide world of staffing. Your business practices aresoftware project for itself must first fully and clearly
equally different. As a result, your staffing businessdefine the objectives for the project. Few do this to
requires specialized information technology thatthe extent needed.Software development and
addresses those unique needs. You are convincedimplementation is an incredibly complex process, even
that with the right IT infrastructure -- the right staffingfor a moderately sized firm. Senior management must
software -- you will be more efficient, productive, andbe involved from the get-go. End-user considerations
profitable. If you subscribe to this common belief, you,are paramount if the technology is to be fully
like Mike, have three options to achieve that end:1.accepted. Questions regarding business practices,
Develop your own custom staffing software andmethods, and future business or technology possibilities
technology infrastructure;2. Hire a consulting firm tomust be answered.Critical to the success of the
develop the staffing software and IT for you;3. Buyproject is both established accountability and an
from a staffing software/IT vendor familiar with yourunderstanding that software development is inherently
industry.Of these three options, which do you think isrisky due to rapidly advancing programming technology.
most likely to work the best, cost the least, and getIt's useful to remember that the billions of dollars
implemented the fastest?If you, like the high-profileallocated to correcting Year 2000 computer problems
public staffing firms of Norrell (now Spherion) andwere the result of management and programming
Manpower choose to develop your own softwaredecisions made, in some cases, in the 1960's.At a time
solution, you too may have a very expensive futurewhen CEO's often focus on the performance of the
write-off to enjoy. In fact, Manpower's write-off for itsnext quarter, trying to plan and manage something that
failed three-year software development effort in themay take years and millions of dollars is difficult at
late 1990's was pegged at $57 Million.The right answerbest. Even with long-range planning and management,
to this perplexing question is to buy from a staffingmis-steps are fairly typical. And it takes unusually
software vendor who is intimately familiar with thestrong and pragmatic business leadership to admit that
industry.Surprised? After all, you're special, unique,they've taken the wrong development road and make
different, right? Well, not exactly. Yes, it is very trueany needed course changes.3. CYA CIO's: The CEO
that you and your competitors differ in a number ofwho does not want to get involved in software or IT
important areas, such as management styles andprojects is often tempted to place all of his eggs in the
objectives. However, you and all of your competitorsbasket of a trusted CIO or similar IT manager. Doing
also share a wealth of common traits, such as trackingso may set the firm up for a humpty-dumpty fall.Years
employees and candidates, performing the businessago, the MIS department was often referred to as the
transactions of orders and assignments, making payroll,"glass house", where massive mainframes resided,
and billing customers.These shared business issues arepresided over a technical priesthood whose mystic
at the core of your staffing and recruiting business,incantations were the interface between the "big iron"
and it is these issues that a staffing software vendorand the information needed to run the business. Today,
familiar with your industry is best qualified to address.business-savvy CIO's and networked computer
After all, such a vendor has researched the issues tosystems are the rule.However, the firm's CIO and the
develop the necessary technology. TheseMIS department may still have a vested interest in the
industry-specific software vendors can oftentechnology and methodology with which they are
implement a solution that can handle 80 percent to 90already familiar. They may want to protect their turf at
percent or more of your business operational needs. (Itall costs. As a result, their tendency is to do things "the
should be noted however, that vendor solutions mayway we've always done them", only bigger and more
vary greatly in quality, efficiency, andexpensively.At the other extreme are sometimes
cost-effectiveness, depending on their expertise andvisionary CIO's who see a new IT project as an
previous success in the marketplace).But let's say thatopportunity to bring in the latest and greatest
the "90 percent solution" that an industry-familiartechnology, without fully understanding it or its
vendor has to offer just isn't enough for you. Youapplicability to the business. Here, they may apply an
want it all. Or you want some features or functionsattitude that if the business sotware is "not invented
that the vendors don't have.Your options then are tohere", it won't work right.Either way, the firm loses,
either do the job yourself or hiring the project out to(although the CIO and MIS department staff may gain
outside consultants. Both are similar in that you, thesome new programming skills they can use on their
client, assume the overall project management risks,next job).Can you ensure the success of any
and, in the case of doing the job yourself, the"Do-It-Yourself" home-grown software project?With
programming chores for your unique staffing software.so many variables -- technology, management,
If you decide to take either of these routes, do so withbusiness practices, internal politics, competitive issues,
extreme care -- as industry surveys indicate that thereand more, it is virtually impossible to predict the
is nearly a 60 percent probability of its failure.Accordingsuccess of any custom software development
to one survey of thousands of software projects,project. At best, you've only got a 40 percent chance
conducted by the Standish Group of Hanover,of success. You can, however, improve your chances
Massachusetts, four of 10 software projects failedsomewhat by following a few simple guidelines:1.
outright. To make matters worse, an additional 33Establish goals. The most important thing is to define,
percent of software projects were completed late,exactly, what you want to accomplish. What is your
went over budget or were completed with fewerfirm's business plan? How does IT relate and
features and functions than originally specified. Cancontribute to it? What should its objectives be? Be
your staffing business afford to make that kind ofrealistic in setting the goals, then document them. Get
risky investment?In an even more unexpected finding,buy-in on these goals from senior management all the
the study also revealed that the use of IT consultingway to end-user.2. Establish budget and time
houses -- even highly respected companies such asspecifications. No matter how you cut it, software
Andersen Consulting (now Accenture) and Lockheeddevelopment is all about time and money. Working
Martin -- increased the risk of a project's failure. This isfrom your goals, consider the cost to develop the
astonishing because such consulting firms have stakedsoftware, AND those areas where budget and time
their reputations on and are hired for their claimedrequirements can increase exponentially. This includes
expertise at developing or implementing enterprisetesting, documentation, implementation, and end user
software from such giants as SAP andtraining. And don't forget ongoing support, development,
PeopleSoft.But the reason for their poor results is notand "bug" fixing.3. Establish project milestones.
so surprising -- as it pointed to business practices thatHand-in-hand with establishing your budget and
leave a lot to be desired, frequently at the considerabletime-frame -- determine what goals you have to
expense of their clients.It should be no wonder thenachieve in the development process to complete the
that dissatisfaction with IT contractors has reached anproject on-time and on-budget.