iPerceptions : web analytics, attitudinal predictive customer feedback
Turn Up The Silence

Oct 09

Everything New is Old Again

By Christopher Pam and Dany Di Tullio

EvE Online is space-based massively multiplayer online game, or “MMO” as the industry jargon goes. Ever year, the CCP team (EvE’s developers) produce 2 free expansions for their game, chock full of new features, items and ships to be bought, sold, flown, traded and destroyed. While the player base appreciates all the shiny goodness that comes around expansion time, they grudgingly acknowledge that with awesome new toys comes an army of bugs.

While a game like EvE Online (with over 600,000 lines of code) may be a more complicated animal than a website, the general principal remains the same: big releases generally have big bugs. In order to counteract this common ailment, web developers typically work extensive testing into their development schedule prior to releasing their updates to the live servers. Testing takes man-hours, and the fewer people you have testing and debugging, the longer the process takes.

Working backwards, lets assume that WIDGET inc. has a major release going live in December. With a December launch date, that probably means they have been testing and debugging for a couple of months (September or October). When you factor in programming time, this may mean that “feature freeze” was in July or August. WIDGET inc. is leveraging intelligence gained from their clickstream and survey data, which naturally predates their feature freeze. Their reports are based on data gathered from April, May and June. Phrased another way, their “New December Release” is based on data that is 6-9 months old.

In the offline world, 9 months is the time it takes to create a new human. In the online world, 9 months is an eternity. Moore’s Law states that computing power (CPU transistor count) doubles every two years… 9 months is a long time! What may have been innovative this week has become standard by next month. YouTube, social networking sites, RSS feeds, blogs, 24-hour news networks (and their websites); all focus on distributing information as widely as possible, as quickly as possible. Once “nugget A” is sent out, the focus is on “nugget B”. People no longer care about “what is”… they care about “what’s next”.

So often I hear the criticism from clients that Voice of Customer data doesn’t change month over month. But the question I have (which I typically don’t verbalize) is: “why would you expect it to?”. If your company follows the long development process illustrated above, websites are essentially stagnant. Why would you expect metrics to be any different?

The ability of a website to improve and satisfy their customers is directly tied to their development cycle. More and more companies are adopting a micro-release philosophy, using customer feedback data to make their sites incrementally better. Micro-releases may contain a couple of minor changes, but sometimes a little goes a long way. Smaller changes mean less time debugging, testing and require less programming time overall. Not to mention fewer man hours; a positive for companies where the web team is either understaffed or overworked.

This is akin to the proven open source software development approach: “release early, release often”. Most open source projects are an ever changing, ever morphing response to the constant customer input that this approach brings. The tightening of the feedback loop is a substantial benefit here that can also be leveraged in a website redesign context. For instance, you will quickly know if you misinterpreted something and are going down the wrong path which allows you to fix these issues with minimal time loss.

Shortening the development cycle by adopting a micro-release philosophy allows webmasters to meet the needs of visitors while keeping the site relevant. Patches can easily be released, and websites are better suited to adjust to what is the current “flavour of the month”. The quest for the “best” website becomes a series of steps, as opposed to a giant (and sometimes blind) leap. Websites are no longer optimized for “9 months ago”; they are optimized for 2-4 weeks ago. Not ideal… but significantly better than the alternative.

Listening mechanisms like Voice of Customer surveys and clickstream data are tools that provide real-time feedback, and they are tools that are too often left in the toolbox. Website administrators need to shorten development cycles and leverage these tools to ensure that their visitors are satisfied. They otherwise run the risk their shiny new website be “old news” before it even launches.

Post a Comment