Vanno is live again. We’re still the same Company Reputation Index that we always were, but the way we go about building our rankings has changed significantly. We think the changes have resulted in a more powerful set of business tools for our subscribers. So if you’re interested in how companies treat their customers, employees, communities, the environment and society in general, sign up for a free trial.
For those of you who were familiar with our free site, here’s what happened between then and now. The free site relied on “crowdsourcing” to generate the raw material – stories, votes and comments – for our analysis. While we count many serious and thoughtful people among the thousands of users of our free site, we discovered two things that led us to question the viability of a crowdsourced model for company reputations.
First is level of interest. Surveys find that most consumers say they are very interested in how companies behave, but there is much recent evidence (independent of our experiences) that points to a big gap between sentiment and action. Getting consistent and serious participation by a critical mass of engaged and informed online users on serious topics like company reputation is evidently non-trivial, and I suspect we’re not the first site to discover this truth.
Second is independence of voice. As we’ve discussed in our blog and white papers, credible crowdsourcing requires that the opinions being collected are independent. We discovered that user behavior was much more correlated, leading us to coin the phrase “mobsourcing” for much of what we were observing. We suspect that most social media news and microblogging sites are in fact seeing primarily mob behavior. And the highly correlated nature of “mobsourced” data greatly limits the ability to make reliable predictions and inferences. We’ve found, in fact, that the most useful thing about social news and microblogging sites is that the mobs and influential users seem to point back to a rather consistent (and relatively small) set of primary sources of stories and commentary, particularly when it comes to companies.
So the bad news, then, is that crowdsourcing really doesn’t work for us because the online crowds are mostly mobs with ADD. But there is much good news too. First and foremost is the extent to which business people – particularly investors, consultants, marketing professionals and corporate executives – want more insight into the subjective aspects of company reputation that we track, and consider these insights to be a very valuable complement to financial performance data, particularly for larger companies whose behaviors are talked-about the most online. And our experience with the online mobs and their ringleaders taught us how to find the sources that contained the stories and commentary that had the most influence on company reputations. This allowed us to build an engine that collects and orders company-related online stories and commentary so that human editors can quickly and effectively assess the effect the stories and comments have on various aspects of company reputation.
So the result is the new Vanno: a corporate reputation tracking service that follows the most talked about companies on the Web, and analyzes online stories and commentary to rank the top 250 companies based on how they treat their customers, employees, communities, the environment and society in general.
A final note about pricing. This is a labor of love for us, but it’s also a business. We struggled to find a reasonable price point, and concluded that we’d feel like we were doing something useful if a subscriber was willing to buy us a beer or a coffee once a week in return for our efforts. That’s where the $9.95 a month came from, in case you’re wondering.
Let me be among the first to subscribe. Nice work on the site and strategy. That was a quick turnaround, well worth waiting for. I’m in.
Welcome back. You’ve been missed. I understand the decision to change the inputs to avoid “mobsourcing.” That said, I’d appreciate a little more daylight around how you now create ratings. How much of the process is quantitative and how much is qualitative? What degree of precision and confidence do the new ratings carry? I need to be able to answer these questions prior to recommending Vanno to my clients.
Thanks for coming back to Vanno. Your question about the mix of qualitative/quantitative in our rankings is a very good one. Broadly speaking, our approach to determining which companies are “most talked about” is quantitative (primarily automated), while the way we determine which stories and commentary are most influential is qualitative (primarily editorial).
We use keyword and semantic tools to track most every business and company-related news site, blog and microblog (e.g. Twitter) for stories and commentary that relate to the aspects of reputation we track. We actively monitor upwards of a thousand companies – but the exact number of companies in the Index (250 right now) is determined by a lower “buzz” threshold that we set. That makes our Index rather unique, in that membership is determined by magnitude of online buzz, not market cap, sales, geographic reach, etc.
The fire hose of company buzz is organized in a time and source-correlated manner and then scrutinized by members of our staff, who use their editorial judgment to assign aspects of reputation and weights to stories and commentary. So an editor may look at dozens of stories about Toyota, for example, and select the ones that appear to have the most overall influence on the online dialog about Toyota’s reputation. Those stories are assigned specific aspects of reputation (e.g. customer or employee satisfaction, community involvement, the environment or social responsibility), and given weights (strengthen or damage reputation).
We designed this approach based on what we learned from trying to crowd-source company reputations. The mob nature of online users introduced significant biases in the flow of stories and commentary (most talked about) and the assessment/assignment of importance (influence). We think this system significantly improves the reliability and accuracy of the rankings relative to our two principal criteria – “most talked about companies” and “most influential stories”. Of course, there are always biases in the judgment of human editors, but we think the combination of automation and considered human judgment is the “least imperfect” way to approach company reputations that we’ve been able to come up with so far.
I go along with you actually, It looks like! Should that become workable to get your blog translated directly into French? English is my second language.