The so-called widget economy of the Web 2.0 era seems to be declining. A few years ago, social bars like Meebo and Wibiya as well as info bars like Snapshots were popular across the web. Based on our recent experiences browsing the web, these widgets have largely disappeared, though a few popular widget types survive: live chat add-ons (e.g. Olark) and feedback tabs (e.g. UserVoice) in particular.

Although in theory the widget economy could work based on a symbiosis model, in practice it rarely works that way. Meebo’s CEO, Seth Sternberg, even declared that the widget economy was a lie. Our guess is that there has been growing awareness of A/B testing among web developers, and most widgets prove to be conversion killers, despite possibly providing ancillary benefits. As to why most widgets detract from site goals, we hypothesize that the mode of deployment (throwing a box with extraneous content in the middle of the conversion funnel) practically annoys users, changes browsing mood, and derails the conversion process. Of course, certain modes of deployment annoy more than others (e.g. pop-ups begging you to sign up for a newsletter vs. static chat box on the right bar), so manner of deployment is also a relevant factor.

The common trait across those widgets that remain successful (e.g. KissInsights, GetSatisfaction, SnapEngage) seems to be that they focus on making the site owners happy instead of ostensibly making end-users happy through social functionalities or enhanced browsing experiences. Also, site owners no longer swallow gibberish like “increase site engagement” when shopping for onsite solutions; they know too well that “increase site engagement” means increasing engagement with the widget solution while reducing the amount of attention that end-users would pay to the main website. There also has to be a way to measure ROI that is tied directly to sales, not vanity metrics like number of followers, traffic, or even “leads,” whatever that means.

These are lessons that we have learnt from building our flagship widget product, Zuupy FriendAdvisor, which we have tried to apply consistently when building our second widget product, Zuupy CrowdDeals. Perhaps the best way to sell widgets is to please site owners first to get past the barrier of deployment, and then work on pleasing the end-users to avoid causing disillusionment in the site owners, or perhaps the best thing to do is not to build things that sit in a box in the first place. Share with us what you think.