If I were to create a statistic called “Talk of the Town”, and with it try to gauge which restaurants were in the news, I think most folks would have a good instinctive idea what such a statistic would measure. It shouldn’t be limited to a particular forum. It shouldn’t be limited to a particular group. It should have some sense of who is talking about what, and how wide spread that conversation actually extends.
Urbanspoon provides such a statistic, and it’s called Talk of the Town. In general, they weight contributions by mainstream media, alternative media, and to some extent, their own top 10 bloggers in creating this list. I’m in the top 5 bloggers as ranked by Urban Spoon, and I noted one day that if I reviewed a restaurant, and another top 5 blogger reviewed a restaurant within a few days, that restaurant would end up in the bottom half of the top 10 for a week or so. It was something I noted. It didn’t really disturb me. The metric, as it is currently implemented, is however notably imperfect.
There is an ‘elite’ status that Urbanspoon can confer to an active user of the system, and they are called Primes. Some primes are well known bloggers. Marie Let’s Eat, Foodie Buddha, Chow Down Atlanta and your truly all are Primes. Some Primes, and some of the best Primes, have no personal blog and yet wield, in my mind, considerable influence on the Atlanta food community. Barney (who also has a handle on 285 Foodies) comes to mind.
Over the past months, there has been a push within Urbanspoon, by a certain set of Primes, to decouple US’s Talk of the Town stat from any meaningful connection with any influence outside of Urban Spoon and make it dependent purely on what Primes think. This is one proposal. Another is to make it dependent on what “Top Contributors” think. Understand that US is a web and phone application whose parent company is small, with few employees. They depend heavily on Prime contributions to do dirty work for them and keep their database fresh. If a particularly important subset of Primes push hard enough, and no one speaks up, things will change, and not necessarily for the better.
The subset pushing for decoupling is noted for another couple peculiarities. They are envious of the mainstream media, and they are jealous of bloggers. Every time they open their mouths, bloggers like me are depicted as giving a less than sincere or “genuine” contribution to Urban Spoon. I’ve seen a ton of blogger versus Prime debates, and this group of Primes disappoint me routinely with their myopia. The notion that a major fundamental difference between Yelp and Urbanspoon might be US’s investment in its blogger community does not occur to them. The notion that Yelp might be more read than Urbanspoon – much less TV or print media – never occurs to them either. And it’s exactly this subset leading the charge to remove any “external” influence from the Talk of the Town stat.
Metrics at times don’t entirely encapsulate the influence of certain people on the food scene. Take Mr Jones of Eat Buford Highway as an example. Take Sean, of Take Thou Food, as another example. These folks have, on occasion, turned the whole of the food community’s opinion on the reputation of various eateries, and yet are not to be found on Urbanspoon’s Prime list, nor among their top 10 Atlanta bloggers (Sean, though, is still the top ranked Athens GA blogger). That said, thinking about how to measure their contributions leads to the abstract notion of reputation.
So what is a reputation? How do you measure it? Why are we concerned at all? It’s because if we’re in the shoes of a small web company with limited resources, we really can’t measure a Talk of the Town. We don’t have the resources to, say, interview millions of people continually. We, at best, have to estimate it. And if we can estimate reputation, and find a subset of folks who have plenty of it, then the “Talk Score” for a restaurant becomes, in pseudocode:
Score = Sum(Mention(i)*Reputation(i)/(Age_Of_Mention + 1 ))
So, who should be counted? How do we measure their influence? But simply put, about the worst solution I can think of is to total a few select members of the US community. Questions that come to my mind are: why should Primes or Top Contributors have any more say than a regular US member, and if the answer is no, don’t we already have the results of what they seek in the regular Urbanspoon restaurant ratings? If not in the rating itself, just look at new restaurants with high approval rankings (< 3 months old, circa 20 votes, over 90% positive).
What’s wrong with a US statistic that is what it claims to be? Call it Prime Picks, or Top Contributors Recommend? That’s the appropriate forum for the “elite” users. It’s otherwise a travesty to call such a stat Talk of the Town, because the notion that I’m paying as much attention to my local Top Contributors* as much as mainstream media in town, is asinine.
Now, much of my points of view come from someone living in a city with a large number of Urbanspoon users, a large blogger community, a large and active collection of media talking about food. In communities served by one or two newspapers with an indifferent attitude towards food (I’m thinking about Shreveport LA, near where my dad lives), an extended definition of “Talk of the Town” may create a metric that serves the community better than the current implementation. But a *cough* Talk of the Town *cough* metric that basically counts only what Primes think is in my mind, exclusive, degenerate, too reminiscent of the top 100 list already, and in many way, overly empowers the vote of those Primes.
* One problem is that it’s too easy to become a Top Contributor. I can become one by making 10 reviews, taking 40 photos of every restaurant I review, and uploading each photo into Urban Spoon. This could be done in the span of a week, and suddenly, I’m a “Top Contributor”.
Update: fixed the formula to add an age component. Older mentions should eventually disappear.