So, another flurry of stories about fake reviews on Yelp, TripAdvisor, and elsewhere — this time prompted by a Cornell algorithm that supposedly can, with 90 percent accuracy, spot a bought-and-paid-for fake.
(The New York Times and NPR, among others, covered this.)
As I wrote several years ago, it’s getting ever harder for humans — including me — to identify fakery. It’s nice that an algorithm can improve the odds, at least for now, but let’s not get ahead of ourselves:
- Even the best algorithm won’t finger a talented faker, and it also won’t identify a talented programmer who is posting fake reviews algorithmically. (I don’t know for a fact that the latter occurs. But since I can imagine a way to do it, I expect someone smarter than me is already making money at it. Story of my life.)
- Publicity about fake-sniffing algorithms and their methodologies will increase the average “quality” of fake reviews, making them even harder to spot for humans and algorithms alike. We’ve seen this type of arms race before: Algorithms are like antibiotics; they invite the enemy to evolve.
There’s no foolproof way to spot a fake review. However, I do have a foolproof way to spot a genuine review:
Was it written by a friend? If so, it is genuine.
Evolve around that!
Via Andrew Shotland, I recently saw this post. I know I’ll be called naive, but I was surprised at its blatancy.
This guy Stephen Espinosa (whom I don’t know) helps local businesses promote themselves online. His advice is to get your “clients” to post reviews on popular sites — the quote marks are his, and he adds a smiley face in case we don’t get it:
I won’t spell it out fully, since he doesn’t, but this seems like an opportune moment to talk about fake reviews.
You need spend only a few minutes on most rate-and-review sites to understand that they contain fake reviews. There are fake positive reviews posted by the business owners, and fake negative reviews posted by their competitors. Many are amateurish and easy to identify if you’re looking for them, though I suspect that some casual users don’t realize they’re fake.
I’ve never put much store in reviews by strangers. Still, I always thought that out-and-out fakes were a fairly limited and unorganized phenomenon. Now that I see they might be promoted more systematically, I’ve lost confidence that I can even spot a fake.
Furthermore, I expect that such fakery will spread and become more sophisticated. As local search reaches critical mass, it’ll be hard to trust anything.
I used to believe, for instance, that a Yelp reviewer with 10+ reviews and some kudos from friends was almost certainly a real person. That’s probably still a safe assumption — but will it be next year?
If I’m a certain type of SEO consultant, right now I’m probably setting up a network of hundreds of fake Yelpers. They’ll all have real-looking pictures, real-sounding profiles, and lots of reviews (some even genuine). They’ll send each other kudos, enhancing each others’ credibility.
And they’ll exist solely so I can be paid to deploy them for the benefit of my clients.
If done properly, this sort of fakery will be very hard to detect. Probably the only way I’d get caught would be to advertise the service — or to include quote marks and smiley faces when I blogged about it.
And this is just the truly fake reviews. There’s still reviews from friends of the business owner, and “real” reviews that have been solicited directly by business owners, some of whom will give discounts in exchange for posting on … well, on a certain site.
In such a world, reviews by strangers become devalued and personal trust is at a premium.
Not so long ago I heard that we need to see, on average, 20 reviews from strangers before we’ll believe the prevalent opinion that’s being expressed.
What will that number be in the future? 50? 100?
Wouldn’t it be simpler and better to get your advice from people you know and trust?
Via, say, Loladex?
Why would anyone start using Loladex? We get asked this question a lot.
I’ve posted before about the chicken-and-egg issue, albeit in general terms. Probably I should update those thoughts: Since we started focusing on social networks, we’ve learned a bunch.
But for now, let me address Loladex’s specific challenge: How do we motivate people to rate local businesses via a Facebook application? Why would anyone do such a thing?
Well, for many reasons, of course. One day I’ll list them all. But I’d like to highlight one reason in particular, partly because I think it’s powerful and partly because it illustrates a big difference between Loladex and two of its biggest competitors — Yelp and Angie’s List.
Here it is: Loladex believes people will rate local businesses to help their friends.
By friends, I mostly mean actual, real-world friends. People you might have dinner with. For most folks, that’s a subset of “Facebook friends.”
Let’s get specific. Why would anyone use Loladex to rate, let’s say, a plumber? Or a pediatric gastroenterologist? Certainly it’s not something you do on a whim. Loladex won’t be running ads that say “Rate pediatric gastroenterologists!” — and if we did, we wouldn’t expect many clicks.
But suppose you were asked directly by a friend whose kid needed a medical specialist? If you knew of a good gastroenterologist, would you take a minute to make the recommendation? If you were seeking such a specialist, would you value this sort of recommendation?
We think so. Such recommendations are an everyday part of friendship, and numerous surveys tag them as a more powerful force than the Yellow Pages, a $14 billion industry.
With Loladex, we want to provide a channel for these person-to-person recommendations.
Contrast this to Yelp. I always say I like Yelp — and I do — but Yelp isn’t about helping your real-world friends. By and large, the people who rate businesses on Yelp do it for reasons of (a) self-expression; and (b) social standing in an online community that may overlap with their real-world friends, but doesn’t have to.
These mostly twentysomething Yelpers provide a service for us all, God love them. But it’s almost never a person-to-person transaction. Also, the motivation to rate something on Yelp fades quickly outside its core realm of restaurants & other social venues.
Or consider Angie’s List. I’m not a fan of Angie’s List, simply because it’s a subscription service. If it were free, I’d love it. They’ve built something that’s clearly valuable to their users — and they’ve focused their brand admirably, defining it around home services.
Again, though, Angie’s List isn’t about helping your real-world friends. It’s mostly a community of cooperating strangers who share ratings because they understand the value of the site’s virtuous circle. There’s an implicit quid pro quo.
Both Yelp and Angie’s List have powerful models. Loladex aims to tap many of the same motivations; we’d be silly not to. But mainly we’re about recommendations from your friends. We’re trying to bring this everyday personal interaction into your online world.
OK, so much for the theory. How’s the “help a friend” strategy working for us, specifically on Facebook?
To be honest, it’s a learning experience.
More in Part 2.
In my last post I did quick sketches of 14 Facebook apps with a local-search element. I’m not considering Loladex at the moment, because we just launched.
OK, time for a reality check. (You may be depressed by the following.)
The most popular app of the bunch, TripAdvisor’s Local Picks, has fewer than 2,000 “active” users = daily users, more or less. Per Adonomics, it’s the 859th-ranked application on Facebook right now, lagging behind things like Vibrating Hamster at #673.
Like many Facebook apps, Local Picks rose fast and fell fast. At one point in December 2007, it clocked more than 100,000 active users, but it’s fallen below 10,000 for most of 2008 — below 5,000 for the past two months.
More worrisome is the fact that, despite its collapse, Local Picks still has more daily users than the other 13 apps on my list combined. Almost twice as many, in fact.
The second most popular app on my list is Restaurants by Hungry Machine with more than 500 daily users, down from a peak of more than 10,000.
And that application, in turn, has as many users as the remaining 12 apps combined.
In short, local-search usage on Facebook is loooooow right now. And it’s heavily concentrated in a couple of apps, both of which have cratered in 2008. Based on Adonomics charts, I suspect that only a few other apps have any life in them:
• DoYa? may be building a bit, perhaps based on its gift-cert giveaway
• My Restaurants looks to have a small core of regular users that isn’t shrinking
• iEat seems to be growing somewhat
But obviously, the numbers here are fairly inconsequential. DoYa? is the biggest of the three, with 215 active users.
So — yikes, right?
Yes and no.
On the one hand, it’s easy to see how Jon Carder of MojoPages concluded that local search isn’t worth doing on Facebook. (See earlier post.) On the other hand, I think we can learn some things that’ll help us crack the code:
• Recognize the potential. Local Picks is a good app, and it grew nicely in November and December to top 100,000 active users. That’s a number worth noting. OK, in December it suddenly crashed. Maybe someone can tell me why?
• Leverage success. The two biggest apps have more popular “sibling” products on Facebook; none of the others do. Hungry Machine explicitly presents its apps as part of a family, using a toolbar to link between them. TripAdvisor doesn’t do this, but it does some lesser cross-promotion from Cities I’ve Visited. Most of us can’t draft off an earlier app, but there are other ways to apply the lesson.
• Attract repeat users. Most local apps either never took off, or peaked and then fell off a cliff. But several — Eating and Hangouts, for instance — took off and then went into a slow fade. While a fade isn’t as good as a rise, it’s better than falling off a cliff: These apps didn’t lose people immediately. Interestingly, they share a focus on broadcasting “status”-type messages, which may be a key to keeping users engaged.
• Keep working the problem. It’s striking how many Facebook apps are abandoned, more or less, once they start losing users. The charts tell the story: No secondary upward blips as new solutions are tried. The users are allowed to melt away. This isn’t limited to local, of course. Building a Facebook app is an experiment rather than a strategy for many developers, and it shows. But given the potential of local — and its importance to some of these players — I’d expect a bit more dedication to figuring out what works.
That’s the end of this little series of posts. Soon I’ll tackle the question of how Loladex can avoid the fate of its Facebook competitors.
So who’s doing local search apps on Facebook? What follows is my brief impression of various local apps I’ve used in the past few months. I’m sure my list is incomplete. If I’ve missed something obvious, feel free to add it in the comments.
In my opinion, no one on this list competes with Loladex’s combination of functionality, comprehensive content and usability. Oh, and style. But of course I would think that.
This list is in no particular order.
MojoPages: Based on Jon Carder’s comments (see my earlier post), MojoPages is’t pushing this app anymore. It’s a bit of a shame, because (a) they put some thought into it, improving it since it launched some months ago; and (b) the MojoPages.com site continues to push some social aspects that are, in my opinion, better suited to Facebook than to their standalone version. It’s true that the FB app never got any traffic, and the execution still needs work, but I thought it deserved more promotion than they gave it.
Local Reviews: On its splash page, this app by eSesame claims to be “Facebook’s FIRST and LARGEST Local Reviews App!” Even if this was true once, by some measure, it isn’t anymore. It seems downright moribund, in fact. I find the interface a bit confusing and scattered, and the content is limited to a tight selection of categories and cities. Even in these areas, they don’t have comprehensive data — searching for Chinese restaurants in the Washington, DC area yields a single result, for instance. I guess they have just what their users enter?
Restaurant Reviews: As the name implies, limited to restaurants. This app comes from SuperPages.com, which is certainly a credible source, but it never got off the ground. I’m not sure why they restricted it to a single category since they have the whole phone book at their disposal, literally. The app is presented mostly as a way of browsing places that have reviews: You’ll need to click to uncover a search box. It’s mostly a window into user comments from the regular SuperPages site, so there are few associations with actual Facebook users. Not very social.
Local Picks: From TripAdvisor, the king of hotel reviews. Limited to restaurants, despite the generic name — base data from the worldwide but long-in-the-tooth ChefMoz database, perhaps, supplemented by members? (Someone correct me if I’m wrong.) Users appear to be griping that their additions and deletions aren’t processed quickly enough by TripAdvisor, which is also my experience with the TripAdvisor hotel database. Still, it must be said that this is a very nice Facebook app: Deep, engaging, well-designed. They are making an effort, and they have more users than most local apps. TripAdvisor also makes the “Cities I’ve Visited” app, which is far more popular but not nearly as useful.
iVouch: I recall that this app used to be a local thing, where users “vouched” for businesses, but lately it’s about users vouching for each other. The local aspect still exists, but has been subordinated. It’s not particularly useful, in any case, as I can’t search for places that others have vouched for. Or I don’t think I can, anyway. Strange. Either the app isn’t fully conceived or I’m missing something.
Search Local: This is a fairly straightforward Yellow Pages app, except that it doesn’t seem to have much data. (Or maybe its search doesn’t work properly.) Its main distinction is that it wants to broadcast all your searches to your friends’ news feeds. This may or may not be good for virality — not, I’d guess, based on its traffic — but either way it’s a bad idea. No one wants this stuff to be public by default. They have an opt-out checkbox, but it’s easy to forget and I can’t set it to opt out permanently.
DoYa? This app has some of the same ideas as Loladex. They’re trying to jump-start virality by offering $10 Amazon gift cards to people who “share” five recommendations with 10 friends. Depending on how many of those friends become DoYa users, it might make sense. I’m not crazy about the interface, but that’s mostly a matter of taste. They’re on the right track.
Eat It! Is this app working as designed? It’s a restaurant-only product by big ol’ CitySearch, and someone spent time on it, but it lacks all of CitySearch’s deep content for each listing — no information about restaurant hours, no professional reviews, no user ratings, no photos, nothing. Just an address. Whole features like “QuickRate” appear to be pretty much broken, too. Weird. I have a vague memory that it used to work, but until today I hadn’t visited it for months. Maybe its creators haven’t visited lately, either?
Eating: Yet another restaurant-only app, this one by Menuism. It doesn’t take a local-search approach, exactly; it’s got a less utilitarian, more social vibe — more like one of the various “bookshelf” applications on Facebook, where people participate to express themselves and share their interests. Or like Yelp, except with more bona fide social networking. That’s my take, anyway. Like Loladex, it incorporates Facebook networks, which is a good way to open up more friend-like content. Data is far from comprehensive, however, and the sync to the “real” Menuism site seems imperfect.
Restaurants: Bulit by our neighbors in the nation’s capital, Hungry Machine LLC, creators of Visual Bookshelf. Once again, restaurants only. Once again, data is far from comprehensive. Annoying popups that ask me to invite friends. I don’t have many friends on the app, so it’s not clear to me whether their recommendations count more than those of strangers. Be nice if they did. (Anyone know?) Generic food photos are used to illustrate specific restaurants, which can be misleading: Our local chain Ledo Pizza, for instance, is known for its square pizza; they probably aren’t thrilled to be represented by a round pie. I’m not a fan of the overall design, which is busy, but the underlying functionality seems solid.
Following are some other apps that I haven’t used much yet, but that may be worth a look. None are “big” by Facebook standards.
[Added after initial post] Hangouts: OK, so this is Yelp’s application. It’s not a local search, just a way of broadcasting what you’re supposedly doing this evening, with a search that sometimes helps you link to the right listing. It launched fairly soon after the Facebook platform launch, and has the air of a side project rather than a committed effort. Functionality is shallow and requires going to Yelp.com for detail on listings. I could be wrong, but it seems to have been abandoned by Yelp. I assume that if they ever return to Facebook it’ll be with a different approach.
My Restaurants: A modest app that seems like it’s doing many things right.
Eat-a-Rama: Bills itself as the “top-rated restaurant application on Facebook,” whatever that means. At a glance I’m not a fan: Overdesigned & gimmicky. Someone is working hard, though.
iEat: This app doesn’t really appeal to me, either, but it’s a bit interesting because it has a two-level search. If it doesn’t find a restaurant in its database (a likely scenario, it seems), it does an “Internet search” and returns matches from … well, somewhere or other. I enjoyed the stock photos shown for each restaurant; see the one for Domino’s below:
While I’m at it, I should mention a few notable absences:
Yelp: As far as I know, Yelp has no app on Facebook. A few third-party apps try to track or broadcast your Yelp.com activity. See above; thanks, Jon.
Zagat: Zagat has a Facebook app, but it’s for use by restaurants rather than regular folks — the virtual equivalent of putting a “Zagat rated” sticker in your restaurant’s window.
Angie’s List: Not terribly surprising that AL doesn’t have an app; as a subscription site, it doesn’t really translate.
(Continued in Part 3)
Responding to our TechCrunch coverage last week, Jon Carder of our competitor MojoPages said the following:
Local search just isn’t gamey or sexy enough to make a local search Facebook App worth the time.
I disagree with this analysis on several levels. It’s true, so far, that many popular Facebook apps have a game-like aspect. And it’s true that their success can teach us some lessons.
But such apps are just the baby steps of social networking; the meaningful stuff lies ahead, when we’ve learned to walk. And run. And climb. And that’s definitely “worth the time.”
Indeed, I think it’ll make Facebook as a whole “worth the time.”
Social networking is already more than games. It allows our online actions to be informed and enhanced by the participation of our friends. The Web has become more like the real world.
In the case of local search, it means we can finally get advice from our friends — something we’ve always done offline, but that’s new on the Web.
Say you’re looking for a good Italian restaurant. Offline, you can consult the Yellow Pages. Online, you can do the same. You’ll find a restaurant, but not necessarily a good one.
Offline, you can see what’s recommended by the local newspaper or magazine. Online, you can do the same. They review a very limited number of places, however. (And beyond restaurants, virtually nothing.)
Offline, you can consult Zagat. Online, you can do the same. Or check Yelp. You’ll get the wisdom of a certain crowd.
Offline, you can seek recommendations from your friends. According to every poll, these are the recommendations you’ll actually trust — the advice you’ll actually take.
Yet until recently, it’s been pretty hard to do the same thing online. Facebook has made it easier. And increasingly, Facebook is where your friends are.
To me, that means Facebook is a great fit for local search.
Of course, we’re still figuring out how local search should work in a social environment. Jon observes:
The existing apps [on Facebook] are all fairly good, well designed and user friendly yet none of them is gaining any sustained traction.
I don’t agree with the “fairly good, well designed and user friendly” part, but it’s absolutely true that no one has gotten traction.
Why? I decided to survey the current apps to see if I can draw any conclusions.
(Continued in Part 2)