Can biz software really be crowdsourced?

Replies to This Discussion

Permalink Reply by P.J. Hinton on January 10, 2009 at 7:50pm

I don't think it makes sense to have a complete business application developed within the context of a time limited competition. But the reasons vary depending on the situation. Let's look at the question from two angles:

1) a business aspiring to become an independent software vendor which licenses the target application either as a shrink wrapped product or as a service

2) a business needing a custom application for use in-house

For case (1), the company's long term survival hinges not only on the release of a single version but also the ability to extend and improve the product for years to come. Software is a competitive marketplace. If one company releases a product that is wildly successful and makes lots of money at it, you can be sure that other companies will attempt to enter the marketplace to get a slice of that pie. One release, no matter how good, will one day look stale and inferior when competitors start add innovative features.

If you crowdsource your application through a competition, your relationship with the developers is probably over once the competition is done. The developers will be coding with a much shorter term mindset, namely get the job done by the end of the competition. In that state of mind, the code usually isn't written from the standpoint of robustness or maintainability. That drives up the cost of training developers who later have to maintain or improve the code base.

If you've got an idea for an application and have a vision for building a business that will grow and endure, you're better off building up a team of smart-and-get-things-done people who have a vested interest in its success. You don't get that from the one-night-stand nature of competitions.

Regarding situation (2), chances are that an in-house business application will have to integrate with an existing infrastructure that is heterogeneous, quirky, and in some cases sensitive. Developing non-DOA deliverables will be a long shot if some of the points of contact are legacy systems or proprietary products for which developers cannot easily obtain tools. Testing may require databases which, in production, contain personal information. Creation of mock data sets may not be feasible.

The integration specifications may require divulging pieces of the business' IT architecture, information that you probably don't want falling into just anyone's hands. When the competition is global, the field may contain entrants whose personal ethics do not share the same concern for secure handling of confidential information as your company does.

Having said that, I do think crowdsourcing might have a place in business application development. Suppose there is a component that can be adequately isolated to be developed outside the full context of the application. Perhaps the component hasn't been developed yet, or is an old piece of code that is underperforming, and there is no one on staff who has the time or expertise to develop a good implementation. I think in a case like this, the scope is narrow enough that the a short term engagement can bring about a superior result at a reasonable price.

▶ Reply to This