I don’t trust Yelp any more. And that’s not a conscious decision. I’ve largely ignored the well-publicized allegations of how the ad side manipulated ratings and reviews to drive sales, instead continuing to turn to the site for recommendations on everything from restaurants to plumbers to airport parking. But they’ve nonetheless seeped into my subconscious and tarred my view of the service. Indeed, Yelp is learning that trust is a hard thing to win, but amazingly easy to lose. And that’s why it needs to be protected with the corporate equivalent of the Praetorian Guard.
Here’s what happened: The other day I was looking for a decent long-term lot at San Jose Airport; Yelp’s recommendation page was near the top of Google, so I clicked over to check it out. As usual, I scanned the top 3-4 results, then read through the reviews of the most likely suspects. Quick Park SJC, ranked No. 1, seemed to have everything I needed -– a rating of over 3, pretty decent reviews and a nearby location.
But then I started looking more closely at those reviews -– and noticed an interesting pattern: There were a few from the last month or so, then nothing for nearly a year. And the last of that group was complaining about how the lot had recently raised prices and had a bit of a surly shuttle driver problem.
So why the gap? I didn’t know, but suddenly I wasn’t so sure this was such a great place to park. So I clicked over to TripAdvisor to look for airport parking information there. Since I couldn’t find any — and I was admittedly in a hurry — I ended up selecting Quik Park SJC after all. But I was far less sanguine about my choice. Yelp, alas, was no longer in my inner circle of trust.
Which leads me to three key trust rules:
Got a Problem? Deal With It Quickly: I learned that the hard way during my early years running the test lab at PC Week in the 90s. We had a columnist, Will Zachman, who was an ardent proponent of an early Windows competitor from IBM called OS/2. Microsoft, in those days, advertised incessantly in tech magazines, and Zachman felt that his editor was shaping his OS/2 diatribes to please Microsoft. So he publicly declared independence from the magazine on July 4th, accusing members of the business side of leaping over the “wall” and smacking down the EIC until he censored Zachman’s opinions to appease their biggest advertiser. Zachman had been kvetching about his supposed “censoring” for some time leading up to his Independence Day action, but the editorial team just ignored him — until it was too late.
Notably, I never saw any evidence supporting his accusations. Which leads me to my second tip:
Manage Not Just the Facts, But the Perception: Lack of evidence aside, just the merest whiff of perceived bias was enough to tar us with a wide brush. It took us a long time to cast off that perception. And that’s why, about a year later, I immediately fired a junior lab staffer who falsified test results, and not for money or influence, but because he was being run ragged by an overbearing manager. He was young. Impressionable. He probably only deserved a warning. But he violated a trust, one that, had it become public, would have been harmed us even more. I had to take quick action to preserve the trust that our readers had in our reviews.
Trust has to be carefully nurtured and ruthlessly defended. It’s why TechCrunch fired the intern who asked for a MacBook in exchange for covering a new company. It’s why IDG moved quickly to cut off any association with Randall Kennedy and why Facebook has such a big problem on its hands with the hacking allegations against CEO and Co-founder Mark Zuckerberg. It’s because:
If You Trade in Trust, Everything You Do or Say Is Relevant: At our live Diggnation show last weekend at SXSW in Austin, the wacky folks at URDB.Com convinced the crowd of 3,000-plus people to perpetrate a huge hoax — that Conan O’Brien was onstage and coming over to Revision3. Hundreds of partygoers tweeted out the “news,” and it quickly became the biggest Twitter hoax ever.
It was all fun and games, up to a point. Some pretty influential and trusted people tweeted and retweeted the hoax to their followers. Unfortunately, more than a few journalists saw tweets from people they had come to trust and were subsequently convinced that they had to be true. And once they discovered they’d been punk’d, they lashed out. In the end, more than one social media “expert” damaged their credibility by engaging in a little bit of pranksterism.
In Yelp’s case, that delicate tissue of trust has already been perforated — perhaps fatally. For even those of us that have willfully ignored the allegations against the site are ready to go elsewhere.
Jim Louderback is CEO of Revision3. He was previously vice president of Ziff Davis Media and Editor-in-Chief of PC Magazine and PCMag.com.
Music subscription services promise unlimited access to enormous libraries of songs, typically on the order of 6-10 million tracks. And while a few superstar artists are famously absent from streaming services as well as Apple’s iTunes -– the Beatles and Garth Brooks among them -– my experience testing out several services has left me frustrated in other ways.
Indeed, there are plenty of empty trays at the all-you-can-eat music buffet, though some will leave you hungrier than others. Use one for awhile, and the gaps in its catalog soon become apparent. Try two or more, and the inconsistencies among them become downright baffling.
Most of the services claim to provide access to the full digital catalogs of all four major labels and a slew of independently distributed recordings, most delivered via aggregators such as IODA and the Orchard. But why is MOG missing the first two Tom Petty albums, while Thumbplay has them all? Why does Spotify – at least the preview version I’m testing here in the U.S. – have only two or three Bob Dylan albums, when its competitors have dozens? Why is Rhapsody the only one that has Grizzly Bear’s “Veckatimest,” an acclaimed independent-label album that entered the Billboard chart at No. 8 last June?
As I’ve learned from conversations with subscription providers, obtaining a complete and stable catalog of music is hardly as simple as working out a contract with a label or distributor. Songs and albums are constantly blinking in and out of view as ownership rights change hands, reissues are prepared, and songwriters and performers change their minds as to where they want their songs to be heard. Some labels handle their own distribution rather than going through aggregators, meaning that individual deals have to be struck in order to make their catalogs available. Geography can be a factor, as licenses vary from country to country. What’s more, a glitch in something as simple and unsexy as the file metadata that identifies a song –- a missing capital letter here, a misspelling there –- can render a track invisible to the consumer, even if it’s properly licensed by the subscription service.
Filling holes in the catalog is time-consuming and labor-intensive. (As MOG’s director of content licensing, Buzzy Cohen, told me, “Finding the holes is harder than filling them in.”) Companies with deeper financial resources and more personnel will have the upper hand when it come to chasing down rights holders one at a time in an effort to maintain a more complete catalog, so it makes sense that the older companies are more successful at it than the new ones — and explains why Rhapsody’s service, which has been around for more than eight years, satisfies my searches more consistently than any of its upstart rivals.
Though the causes are manifold and the companies’ efforts to fill the gaps are admirable, it’s frustrating to music fans when our searches aren’t satisfied, and even more irritating when songs in a playlist disappear without warning. And as consumers choose from among several services — or choose not to subscribe at all — holes in the catalog can ultimately be a dealbreaker.
Post and thumbnail photos courtesy of Flickr user samsmith
Related content from GigaOM Pro (sub req’d):
Rankings: Spotify Leads the Streaming Music Scene
FCC Chairman Julius Genachowski has spoken to a YouTube team about the National Broadband Plan and tried to respond to questions from Internet users across the country. The answers were, to put it politely, nuanced at best. Regular reader Brent Glass is in the video asking questions about WISPs.
Both Stacey and I are wading through the 400-odd pages of documents and will be offering our impressions after we’re done. Up until that point, you can watch this video. And if you need to learn more, check out these posts to get a better understanding of the National Broadband Plan.
In addition, you can follow our coverage of FCC.
“The web is becoming a live medium — sales and auctions happen in time, product launches, chats with celebrities, live video events and audio, games, events in virtual worlds — there are a huge amount of scheduled things taking place,” says Nova Spivack, who recently sold off his semantic web search engine Twine to Evri. Now he’s setting out to create a massive programming guide to those live online events, called Live Matrix.
Spivack last year brought on Sanjay Reddy, who’d done corporate development at Gemstar-TV Guide and led its sale to Macrovision for $2.8 billion. Reddy knew from experience that TV watchers spend 10 percent of their time watching TV on the interactive programming guide (per set-top box data). The web has much more going on at any one time, but nothing like an IPG.
Does that metaphor carry over? That’s what Spivack, who’s serving as Live Matrix chairman, and Reddy, the CEO, want to find out. They plan to launch in May and gave me a look at the pre-release site under the condition that I not release screenshots. I’ll tell you a bit more about what they’re doing and describe in words what it looks like.
First of all, there’s the prototypical grid interface — though instead of listing channels due to preset numbers like on TV, they’re dynamically ordered based on demand. Log in during March Madness? That’s going to be on top. If a web metaphor is more your style, you can look at all the events in a Digg-like interface. Users can pull events into a personal playlist and receive reminders when they start. If you actually want to watch something, Live Matrix sends you to the host site rather than framing the content.
These ideas do need a bit more nuance — for instance, personalized ranking rather than global popularity will often be more helpful, and on days like the Obama inauguration there are going to be tens of duplicate or near-identical feeds running on various channels. Spivack and Reddy do say they’ll provide recommendations as well as social filtering based on what your friends are planning to watch. They’re also planning to provide widgets for syndication around the web.
I totally see the need for what they’re doing — in my time at NewTeeVee, we developed the “where to watch” franchise of posts after seeing the amount of people coming to our site trying to find the actual URLs for live video streams of major sporting and pop culture events.
The main way Live Matrix expects to list events is by partnering with content providers, plugging into publisher APIs and using a small editorial team. Then, it will sell advertising on programming based on demand expressed by its users. It will also maintain a page for every event after it’s over, pointing to archived video (which is excellent; so often people who do live completely forget about on-demand). And Spivack notes that pre-recorded videos released at a certain time will be considered “live” as well.
Live Matrix has a good problem but, like its on-demand cousin company Clicker, the web TV guide, it will need to aggressively build an audience for its site and through distribution partners, including hardware companies. The key is to make an interface that’s a significantly better option than searching for such events.
The company, dually based in San Francisco and Los Angeles with offshore development, has raised money from angel investors including Allen Morgan at Mayfield Fund.
March Madness is kicking off with its opening round game today, and CBS is getting ready to live stream all games online starting with the first round on Thursday. The network has already made $37 million and sold all of its ad revenue for March Madness On Demand, AdAge is reporting. However, other sites could also profit from the NCAA Men’s College Basketball Tournament, as almost half of its audience will follow games and scores online, according to a new report from Unicast.
As it turns out, March Madness is very much a two-screen experience for many viewers, with live streaming only being one of many parts to the puzzle. Eighty-three percent of its audience will watch it on TV, 44 percent will follow the tournament online and 10 percent will use a mobile device. Sports sites like ESPN.com and Yahoo Sports should see a lot of traffic from fans trying to keep track of the scores, but the official CBS live stream at NCAA.com will see its fair share of users as well: 54 percent of the March Madness online and mobile audience actually wants to watch games in real time.
Unicast’s new 2010 NCAA Basketball Tournament Fever Report reveals that 58 percent of the March Madness online audience want to use the medium to keep track of scores while 49 percent want to watch game highlights online. Catching up on complete games after they air is something that only 19 percent are interested in. Most users said they will flock to ESPN.com (69 percent), followed by Yahoo Sports (42 percent), FoxSports.com (34 percent) and CBSSports.com (29 percent). The official CBS-powered NCAA.com site was on the mind of 26 percent of the respondents.
51 percent of the respondents don’t care about and won’t follow March Madness at all, which still leaves 49 percent ready to take work a little less seriously in the weeks to come. Out of those, 44 percent will go online. Put those pieces of pie into context, and you end up with the estimate that around 11.6 percent of U.S. adults are planing to tune into March Madness online. That’s a lot of eyeballs, even if not everyone follows through. Last year’s March Madness online coverage was seen by 7.52 million people, which generated a total of 8.6 million hours of live streaming video, according to CBS.
AdAge reported yesterday that CBS has sold as many ads online as on TV this year. Of course, TV ad revenue still outweighs online, which was estimated to be around $619 million for last year’s March Madness. Still, online revenue is up 20 percent, which validates CBS’s strategy to stream the games in real time without any of the restrictions NBC imposed on its Olympic coverage this year.
Related content on NewTeeVee: Where to Watch March Madness Online
Related content on GigaOM Pro: Why Is the Plasma HDTV Market Dying? (subscription required)
In a world of web-based services that depend on various other services, like Twitter or Google Maps data, your product will only be as strong as your weakest API call. I’ve found it’s actually a fun topic to discuss. For developers and web-based businesses, thinking about and managing federated flows of information has a big impact on the end-user experience.
I moderated the Can You Run a Serverless Business panel a few hours ago, and two of the panelists brought that up as an issue, with Jim Louderback of Revision3 (an occasional columnist for GigaOM) saying that at one point he ended up slowing down that company’s video delivery because it had relied on too many services located in the cloud, a sentiment echoed by Ethan Kaplan of Warner Music. It’s a topic I discussed yesterday with Sam Ramji, at Sonoa, which offers a service that helps track the health of APIs.
But after surveying the audience at the panel via a show of hands as to what type of different web and cloud services they use, it was clear that most were using a mix of software, platforms and infrastructure as a service, with some also using traditional hosting or running their own data centers. Information technology has always been a confusing mix of gear and services, but it seems like that complexity is only expanding in terms of how and where you can build an offering as well as the “partners” whose APIs you might use to build a product.
Making all of those elements work together, ensuring a good user experience, and eventually delivering service-level agreements seems inordinately complex, but if done right we should see the emergence of new ecosystems of data, maybe built around cloud providers like Amazon or Google, or maybe popular APIs such as Twitter’s. But like any ecosystem, the effects of small breakdowns will ripple throughout the whole, something we’re for which only now beginning to react and build tools and contracts.
For the GigaOM network’s complete SXSW coverage, check out this round-up.
Related GigaOM Pro Content (sub. req’d): Report: The Real Time Enterprise
The GigaOM Network is a leading provider of publications and events for the technology and entrepreneurial markets worldwide.
