The Customization Epidemic

Good news: The IT boom is here to stay. For four years,  India's IT
market has grown at 27%. The recent Dataquest-IDC Megaspender survey shows that
over 200 top IT-spending enterprises invested an average of Rs 34 crore each on
IT last year.

Better news: IT has spread. Across verticals, well beyond  just banking
and telecom; in government, in PSUs, in smaller businesses.

This special issue profiles projects for PCQuest's IT Implementation Awards.
From e-gov to insurance, from meter-reading to tele-medicine, there's a huge
range in these 250 projects.

I see this vast variety with mixed feelings. Yes, it spells a booming
domestic IT market where 40%, well over $6 billion, was services. On the other
hand, there's a problem in there, one that affects rapid deployment and scaling
up.

The customization epidemic.

It's so visible in e-gov projects. Each Indian state creates projects from
scratch, from consulting to pilots...India is a graveyard of hundreds of pilot
projects. Some worked, but few were scaled up, let alone replicated.

That's also rampant among midsize businesses. The larger ones have been
there, done that. They deployed huge custom projects, and figured there's no way
to maintain them, or enhance features three years later. They learnt the value
of standard platforms, modules that inter-operate.

Smaller businesses often can't find, or afford, the solutions they think they
need, and walk the custom route. Often, that's in-house development. And then
you have projects that are person-dependent and undocumented. There is no easy
way to scale, upgrade or even migrate later.

What do you do if you're a small or midsize business, looking for a
technology solution-and an SAP or Oracle or Microsoft doesn't seem to fit your
budget, or even your needs?

Look harder. The big names have SMB products today, suited to midsize
companies in India. You may find even those expensive. Yet...a from-the-scratch
custom project may appear cheap, but can end up costing three times more (and
taking years longer) than first estimates.

Prasanto K Roy Chief Editor

Compromise on features. Feature-creep kills deployments, where stakeholders
get greedier, and keep demanding features. If an off-the-shelf package will do
50% of what you want it to, but will be off and running in three months, take
it, over the 'perfect' project that will take two years-and may be obsolete by
the time you get it going. The package will also have a larger user base, and
hence support, peer reviews and inputs. Future updates and

scalability are bonuses.

Third: change your business processes. Adapt to the packaged application, for
it may capture current best practices. Don't insist on customization to your
business process.

A 'package' won't fit every need, especially in a country as vast and diverse
as ours. But it's a great starting point.

There is learning in a long custom project. But it's rarely captured. One
project covered in this issue was an open-source workflow/ERP built within a
college over five years. The students' learning was great, but it would be rare
for this to be retained in the institute, documented, and built upon. In a
business, the opportunity cost is horrific: the thousands of man hours spent
could be put to productive use with an existing platform to start from.

So look really hard at what's available in the market. Even if it only partly
addresses your needs. And stand on the shoulders of giants.