Showcase Your Business as a Thought Leader - Publish Your Blog, Videos and Events on ALL EC - Save 25% Now
Welcome Guest | Sign In
LinuxInsider.com

Breaking the Vicious Seasonal Development Cycle

By Joe Chung E-Commerce Times ECT News Network
Aug 5, 2008 4:00 AM PT

We've all done it. The holiday season was crazy, the busiest ever. We missed feature lockdown by a mile, cramming in every last bit just as things began to seriously ramp up. The servers went down on Black Friday. A database crash prevented us from processing returns for four agonizing days. Somehow we managed to survive, adding more chewing gum and baling wire to an e-commerce system than has been stuck to the underside of a third grade desk.

Breaking the Vicious Seasonal Development Cycle

We took a well-deserved break, seeing our family's faces instead of our BlackBerries at the dinner table. Finally, recharged and reinvigorated, we found ourselves in mid-February asking what really great innovations we could conceive, design, build, integrate, test and launch before September 1st. Of course, the answer -- despite the best of intentions -- was "not a heck of a lot."

Creating a Timeline and Sticking to It

This year, let's try it in reverse -- working our way backwards from September. For a major Web development project, the timeline should look something like this:

  • Setting up staging servers, testing and bug-fixing: one month
  • System integration: one month
  • Feature development and production: four months
  • Wireframes: one month
  • Mock-ups and prototypes: one month
  • Vendor selection and contracting: one month
  • Business case, ROI (return on investment) analysis, budget approval: one month
  • Conceptualization and prioritization: one month

Now, I know what you're thinking. You've done the calculation -- 11 months, allowing only four months for actual development! Granted, the above plan applies to a major project. But real innovation -- the kind that leaves your competitors choking in the dust -- requires planning of real magnitude. Meaning that even in a "miracle case" scenario, where everything goes exactly right, you should have started, well, last October. Whoops.

Allocating Resources

Doing this back-of-the-envelope calculation of what it takes to build a complex, mission-critical consumer application that reflects directly on your brand and business goes a long way toward explaining why e-commerce innovation is so hard. We are thinking about the next holiday season, but such projects are simply too big to fit in that breadbox. So what is the best way forward? How do we break the seasonal development cycle?

This is the short answer: Divide and conquer. Effectively split resources into two teams -- one focused on the current holiday season objectives and another working on a separate, longer-term development path. Better yet, look at solution providers that provide a services team -- trained engineers who work side by side with your business and technology staff to help define a program plan to outline timing, resources, process and costs.

Sounds simple enough, but budget issues aside (yes, this approach will definitely cost more, but you can't have a return without the investment), let's look at the specific challenges involved and how best to overcome them.

Challenge No. 1: The Mack Truck Test

The first problem most e-commerce teams encounter is not recognizing their resource bottlenecks and single points of failure. Let's call this the "Mack Truck" test. Take a selection of your most valuable people -- be they developers, system administrators, designers, project managers or visionaries. What happens to your two development tracks if one of your key people gets hit by a Mack Truck (or, more realistically, gets recruited by Google)?

If the answer is that one track fails, you have a big problem. If the answer is that both tracks fail, you have an enormous problem! If both tracks are dependent on a single person, you can guarantee that even if you can keep her off the highways and out of harm's way, when crunch time comes (remember October?) that person is going to be overloaded -- and at least one of your tracks is going to come to a screeching halt.

And maybe when the dust settles, your critical resource may be brushing some of it off her resume. If you fail the Mack Truck test, the only answer is to hire and train, i.e., invest in your resources, to spread out skills and create more interchangeability. By improving the skills of your individual team members, you make the entire team stronger and more resilient.

Challenge No. 2: Law of Unintended Consequences

A second common challenge is overcoming the law of unintended consequences. We've all had senior managers jumping up and down, insisting that their pet initiatives take full priority, and we've literally dropped everything to make a short-term deadline. The problem, of course, is that business managers often do not realize that disrupting well-planned development calendars can have gigantic unintended repercussions, such as missing a major milestone.

Often the manager in question will complain that nobody explained how missing one short-term goal would result in the delay of three long-term deliverables. The cure for this is to make sure your development teams have strong managerial air cover. You need to ensure that just as certain managers are empowered to say "yes," other managers are truly empowered to say "no," and all involved can communicate the real cost of new priorities so that decisions are made with both eyes open.

Challenge No. 3: Clarifying Vision

The last major challenge is the lack of a clear vision of where the site is heading. Typically, everyone agrees on what they don't like about the current site, but separate interviews with the key decision makers often reveal wildly varying answers as to how to fix things and what the priorities should be.

The lack of crispness in future vision tends to lead to a pastiche approach; different functions on the site -- e.g., search, browsing or checkout -- are "improved" in isolation, with completely different philosophies and technologies. The result is a site in which each function is arguably better than its predecessor, but the site as a whole is actually harder to use and more confusing to your shoppers.

Creating and maintaining a single common vision of the future is one of the most important tasks that your longer-term development team members take on. Since they are freed from the insanity of attempting to squeeze an 11-month project into a six-month slot, they can think about the site experience as a whole -- from landing page through to checkout and everything in between.

In fact, some e-commerce operations even break out the long-term vision into a separate third team, which feeds development projects into the current season's track and the next season's track. Handling multiple development tracks clearly requires an extra level in project management adroitness, but otherwise our ability to innovate is constrained by what we can throw together in a frenzied couple of months.

Stop the Madness

We've all done it. Let's not do it anymore.

By implementing the steps above, you can begin to move away from "crisis management development" and start an ongoing, cohesive development cycle -- a structural change that will not only yield an increase in immediate returns, but also provide the added value of year-over-year results.


Joe Chung is cofounder and CEO of Allurent.


Facebook Twitter LinkedIn Google+ RSS
What role will technology play in the upcoming U.S. midterm elections?
Advanced security measures will protect the legitimacy of the elections.
Inadequate security will place the election process at grave risk.
Voter technology is so decentralized, attackers can't have much impact.
Companies that manufacture voting machines need to shore up their security.
The upcoming midterms will be no better or worse than others from a tech standpoint.