Thursday, September 8, 2011

Google Buys Zagat

So Google has bought Zagat. Here are some thoughts:

Impact on Zagat

Aside from the boatloads of money they were surely paid, this rescues Zagat from a slog far longer than the one I endured with Chowhound. Tim and Nina weren't merely slow in extending their brand to the Internet; they were downright luddite about it for years. And even after drawing hundreds of millions of dollars in venture capital funding during the Internet boom of the late 1990's - from investors who could see their brand's online potential, even if they themselves couldn't - not much seems to have been done with it.

Every step Zagat's ever taken in the tech world has been sluggish and tottery. The books were great. The branding was great. The distribution was pure genius (bookstore sales were surely dwarfed by sales via the countless unconventional channels they opened). But tech + Zagat, even at this late date of 2011, has yet to fully happen. They've been waiting...and waiting...and waiting...

Impact on Google

Google is all about data. Lots of data brought into their computational fold toward the goal of "organizing the world's information and making it universally accessible and useful". Tech nerds, who traffic in algorithms and such, prefer their data normalized (or at least normalizable). That means it's presented in a format which can be organized, standardized, crunched, analyzed, and generally be easily digested by computers.

Zagat offers highly normalized data on restaurants and more, and it's a natural for Google (so what took them so long?). Contrast with Chowhound, a data hairball where lots of people fling information at each other in chaotic plain language, all within only the loosest organization. Try extracting all references to Mel's Diner, much less parsing a nice clean number reflecting its quality relative to, say, Bob's Diner. It's not going to happen (and, for lots of good reasons, it must never happen), and that's why I didn't even talk to Google when I was selling the operation.

Add in the key issue of locality - which Google's previously been slow to seize upon - and it's easy to see why Zagat's a good fit.

One problem, though, is that Zagat's data, while normalized, is highly subjective. And subjective data is not Google's specialty. Moreover, this is not particularly high quality data. It's always been insanely easy to stuff Zagat's ballot box.

Posters to Chowhound offer a rich trail of information about themselves, allowing users to decide who's smart, ignorant, or a likely shill. Still more of that trail is available to moderators, helping them remove dishonest postings and monitor patterns of subversion. Chowhound's not completely honest, nor could it ever be. But it's damned good, and that's because the hairball approach supplies a thick enough trail to keep it that way.

But anonymous individuals plugging ratings into survey forms offer little evidence as to their honesty - much less their savvy. The aggregated result may be sleek - assigning each restaurant a clear rating, which Google's processors can chew on and repurpose to their kernel's content - but it's shallow and iffy.

On the other hand, when it comes to data, the perfect is very much the enemy of the good. In the large view, tons of iffy data is super useful. So load it in, boys.

Impact on Chowhound

Mild positive. The pitfalls of Zagat's algorithmic approach make an alternative like Chowhound more attractive for more dedicated enthusiasts. A few of many reasons:

1. Restaurants are too dynamic for a numerical rating to mean much.
Consider how many of your dining opinions - even your most confident ones - are for restaurants you haven't patronized in months or years. Chowhounds (who by the nature of the forum, strive to offer specifics and currency) often report recent meals, while Zagat raters often lazily report stale opinions (or simply regurgitate conventional wisdom, which is staler still).

2. Averaging restaurant ratings shaves off too much value
An otherwise mediocre restaurant with one mind-blowingly great dish may be a perfect ten for a savvy eater, but will always be underrated in an aggregate. Same for a restaurant with one great cook and one poor one. In both cases, you need to strategize, and Zagat won't help you much with that. Similarly, aggregated ratings lead to distorted results. Should a very good street lollipop vendor rate as highly as a very good French bistro? My answer, of course, is a loud "yes", but I'm hardly Zagat/Google's target.

3. When ratings clash with actual experience, confidence plummets
For reasons above, plus lots more, restaurants don't lend themselves to hard ratings. So when Zagat tells people unambiguously where to eat and their experience drastically differs, that erodes confidence in the brand. Chowhound's hairball approach offers nuanced opinions from individuals, rather than a branded rating. We convey a sense of a dynamic dining scene requiring plenty of strategizing and intrepid exploration - quite a different thing than issuing official pronouncements re: "What's Good".

4. He who controls the agenda controls the outcome.
Zagat's playing field is determined by the listing of restaurants offered for rating, and there have always been (and will always inherently be) limits to the currency and breadth of that list. I like lesser-known places, so many of my favorites aren't even on that list. Chowhound's the place for learning about off-radar venues.

This isn't a sales pitch for the superiority of Chowhound. They're apples and oranges, and the vast majority of people, not so picky, will of course find Zagat's neat, clean orange more useful than Chowhound's fuzzy, sprawling apple. (I actually had a clever plan for polishing Chowhound's data without changing its culture. It would have worked, but CNET rejected it. I'll be recounting it as the final installment of my tale of the Chowhound/CNET merger, so stay tuned).

My point is that for all the attractions of standardized data, much is lost when rich, dynamic, and subjective realms are hammered into machine-readability. And the stuff that's lost is the stuff Chowhound specializes in. Which is one reason I built it in the first place.

But Google + Zagat is a natural, and a great move for both of them. Shoot, I'll make good use of it myself. The standard rap among food writers and aficionados for decades has been that Zagat makes a sensational address book. Well, we're about to enjoy a much more souped-up address book.

With, surely, APIs. Hmmm.....

No comments:

Blog Archive