With all the excitement around user-generated content, and so many new sites launching that depend in large part, if not exclusively, on it, relatively little thought seems to have been given to the question: what if it works?

We were early fans of TripAdvisor a 2004 Model of Excellence award winner. Yet TripAdvisor's phenomenal success came with a hidden price. It is now routinely held up as an example of a data provider that became a victim of its own success. Look up a particular hotel on TripAdivsor and you can find dozens of user reviews. Do they provide information? Yes, lots of it. Do they provide insight? Not really, because they are invariably and passionately contradictory. One person loves the hotel; another had the worst experience of his life. How does a user of the TripAdvisor site extract value from such extreme inconsistency? It's not easy, and it can be frustrating even to try. That's not a good user experience.

Some of the pioneers in user-generated content are acutely aware that user-generated content can quickly turn into user-generated information overload. Yet, discouraging user-generated content is contrary to the business model and editing or policing it creates costs that make the whole model less attractive and more complicated. What's a publisher to do?

One excellent example of how to encourage user- generated content while maintaining its value is provided by review site yelp.com. View a typical hotel review on Yelp and you can click on a "rating details" link that elegantly and instantly summarizes all the dozens of full text reviews that appear below for you to read -- or not.

There's an important underlying lesson here: value is not created by amassing the greatest amount of stuff online. It is created by making information accessible, digestible and actionable. These are some of the tasks routinely performed by data publishers on a daily basis, and it's one of the key reasons that the data publishing business is thriving right now. We are in the right place at the right time with the right skills.

Okay, valuemetrics is a made-up word. But maybe it should be a word. After all, anyone who has ever sold a data product, whether advertising or subscription- based, knows the number one renewal objection from current customers revolves around value, specifically, not enough value. For ad-driven data products, that means advertisers don't believe their advertising generated a high enough level of response. For subscription-based publishers, that means that subscribers don't feel they used the product enough to justify the subscription.

This is very frustrating to publishers, who know that there is often a gap between reality and what customers believe, especially when it revolves around an advertising or subscription renewal decision. Why the big gap? That's because for the most part, these customers aren't doing any systematic response or usage measurement. Instead, they are heavily influenced by specific, often isolated events, both good and bad. To those events, they add in perceptions, gut instincts, stray comments made by others, and mix them all together to create what in their minds seems like objective certainty. How do you argue with a customer who is certain of something that's not true?

The obvious answer is to develop proof of your own to shoot down these misperceptions. Yet despite the obviousness of this answer, at this late date too few data publishers are providing credible and useful value metrics to their customers. And we still encounter too many publishers who don't themselves have any firm grasp of how much value they deliver. That we are now all online, working in "the most measurable medium," makes this all the more remarkable.
For advertising-based publishers, education and context are key. Generating online usage metrics is not all that difficult. But if the sales staff doesn't understand what the metrics mean, and if they lack benchmarks to know if results are good or bad, they will use this information reluctantly, if at all. Keep in mind that salespeople need more than a superficial understanding of usage metrics, because in most cases they'll be educating their advertisers as well, a situation in which a little knowledge can indeed be dangerous.

For subscription-based publishers, take the most obvious step: track subscriber logins and other usage metrics. We continue to be amazed that the relatively small percentage of publishers who do this at all are primarily interested in preventing password sharing and other abuse. Indeed, the bigger the subscriber, the less tracking we tend to see. For enterprise subscribers with IP-based database access, it's a rare publisher indeed who tracks anything at all. We're uniquely capable now of responding authoritatively to that oh-so-common objection, "we really didn't get much use out of your database," yet only a tiny percentage of the industry is leveraging this huge opportunity to improve renewal rates.

It's been said that data publishers have long been resistant to measuring usage and response because they themselves weren't sure they'd like the answers. Unfortunately, this new environment doesn't allow anyone to hide anymore. Besides, if you're not in fact delivering all that much value, wouldn't you want to know that while there is still an opportunity to fix the problem?

Labels: metrics