• by John Capone , June 6, 2012

And other questions from the wild ride of Pinterest 

Some marketers have vented their frustrations with a seemingly fickle audience constantly chasing its own (long) tail.

Fatigue sets in.

“It looks like every couple of weeks, some new social media channel tackles the world, and is rewriting history,” wrote Danny Devriendt, executive vice president and digital and social media strategist, EMEA at Porter Novelli back in March, before ticking off the exhausting list: “MySpace. Google Wave. Facebook. Twitter. Google+. Foursquare. Gowalla. Quora. Pinterest. Yammer. Path. Heatmap. Instagram. Yelp.”

It wasn’t the specter of having to navigate yet another new social tool that had gotten to Devriendt. No. “I have nothing against Pinterest,” he told MEDIA. The panting overreactions of his brethren to the latest shiny new toy coming down the never-ending conveyer belt had driven Devriendt to exasperation. “Every couple of weeks we go through the hype cycle of yet another network, yet another way of sharing content and precious pieces of highly private lives. It’s a bit like the Beatles coming to town: giggles, groupies, T-shirts and lots of screaming.”

By now you can probably recite Pinterest’s numbers and factoids — both hyperbolic and very, very real — by heart: fastest-growing social network; more than 10 million users a month (and climbing exponentially) at last count. An audience that causes marketers to cackle like Mr. Burns on "The Simpsons" — 60-70 percent female, 50 percent Millennial, 50 percent with children, annual household income of $100,000. Four percent of all traffic on the Web comes from Pinterest. Refers more traffic than Google+, LinkedIn and YouTube combined.

Yes, to Mr. Devriendt’s point, Pinterest is bigger than Jesus. And until the Facebook IPO, just about all anyone in the digital space could talk about.

For all its much-lauded ease of use and simplicity, Pinterest and its meteoric rise have left many scratching their heads. But the unbridled exuberance of some marketers led Porter Novelli’s Devriendt to fume that “Pinterest is not a strategy,” and pound out the admonition on his keyboard.

In February, Forrester’s Darika Ahrens similarly cautioned marketers, and told them to all but ignore Pinterest for the time being, writing, “There’s no denying that Pinterest is fun, looks great, and a lot of people love playing with it. That is also true of kittens but no one’s rushing to include them in their 2012 marketing plans.”

Pinterest is not a strategy. It’s true. But it also isn’t a kitten (though you can find plenty of pictures of them there). Ever since it became the nitro-burning funny car of audience reach, it’s been the job of every digital marketer and online professional to figure out how the Pinterest piece fits into their larger strategies.

Stifling efforts has been the closed nature of the company itself. There’s no Pinterest API — well, there is, but it’s zealously guarded and the company refuses to release it, supposedly internally citing Twitter’s API “trouble.” (Nobody at Pinterest would speak to MEDIA at all — about anything, not even kittens — for this article.)

When noted social media analyst (or, as he calls himself, “social media scientist”) Dan Zarrella approached Pinterest asking for limited access to the API for research purposes, he was told basically, he says, “to go pound sand.” Lacking access to the API he did the next best thing: painstakingly assembling together, pin-by-pin, the pieces of it he could access through Bing’s cache. From a dataset of 11,000 random pins, Zarrella was able to draw some initial conclusions.

One of his findings is that taller images are more repinnable. The average number of repins spikes dramatically when the height of the image approaches 800 pixels, and the rate is very low for those under 400 pixels. This is fairly intuitive — anyone who’s looked at the array of blocky images on Pinterest knows that wide images just don’t work, but it’s nice to see breakdowns of repins correlated to pixels nonetheless.

And we had better get used to dealing with these sorts of parameters on content. Jason Amunwa of objective-based Web design and consulting firm Digital-Telepathy says that the way people visually digest information on Pinterest — and that there have been so many of them processing so much of it so quickly -- “shows that the ‘masonry’ layout is ready for prime time among mainstream Web users, where the emphasis is on visually pleasing tiles of content, instead of pages.” Amunwa points to the design direction of the new Windows 8 Metro interface, as well as other content-curation sites such as CircleMe — one of what is sure to be part of a very, very long line of services encroaching on the territory of the Royal Pindom.

Other pretenders to the throne include Gentlemint (Pinterest for guys), SparkRebel (Pinterest for fashion), Chill (Pinterest for videos) and littlemonsters.com (Pinterest for Lady Gaga). In fact, you can expect to hear “Pinterest for blank” as often in the next year as you’ve heard “Pandora for blank” over the past couple, and many of these clones will be ripping off Pinterest’s visual approach as much as its conceptual one.

Yes, this all may drive Danny Devriendt to purchase a large semi-automatic weapon.

Brevity may be the soul of wit, but there are limits. Zarrella determines from his data set that descriptions of about 200 characters, which basically align with a tweet length, are most repinnable. Images with little or no description and descriptions longer than 350 words do not perform well at all when it came to repins, though Zarrella concedes that there could be other factors not accounted for affecting these rates. Again, though probably no surprise to those who watch Pinterest, the most repinnable words (the word most often included in the description of repins) relate to food. In fact, nearly all the top 10 words could conceivably be related to food (these include “cheese,” “bake,” “cake” and “chicken”). “Recipe” is the most repinned word.

Zarrella also found a much greater correlation between likes and repins than between comments and repins (a .87 Pearson’s correlation versus .55). This adds credence to and explicates a larger theory of Zarrella’s: that greater engagement and discussion doesn’t always equal greater reach. The most controversial content is often the most commented upon, yet not the most shared. In the case of Pinterest, this is clearly true.

RJMetrics took a slightly different tack in its independent analysis of Pinterest data, downloading and analyzing the complete pinning histories of random users on which it conducted a cohort analysis. This analysis found that Pinterest had an attrition rate of zero, which would only be possible if no one who started using it ever stopped, or, more likely, if existing users increased their activity enough to make up for anyone who left. “Pinterest is doing a really good job of retaining its users, and develops very fanatical users at a very good percentage of the people who come through the door,” says Robert Moore, founder and CEO of RJMetrics. “What’s really driving this very high retention rate is the fact that there’s a meaningful subset of the population that becomes hooked and continues to use Pinterest month-to-month not just consistently, but increasingly. So much so that it’s potentially counterbalancing any churn of people who were just there out of curiosity and didn’t stick around.”

The explosion in growth, or what looks like a sudden explosion, is actually a classic organic exponential growth story, according to Moore. Zoom out, and the curve looks like it’s taken off over the past few months. But zoom into the previous few months, and it’s going to look exactly the same, only the scale has changed, like a fractal image. “You go from two people to four to 16, and you keep squaring it, and the curves are going to look like this,” he says. “I think they’ve done an exceptional job of making the incremental user just as likely as the previous user to share through their existing social networks. And a really high percentage of that sharing ends up getting consumed and converted into new users.”

While social in nature, Pinterest might not actually be a social network per se, in that it is not yet really creating meaningful connections between users. Moore says Pinterest is something akin to an unsanctioned Facebook app, away from all the noise of Facebook, but one that relies heavily on the social graph of Facebook and, to a lesser degree, Twitter. “To have something as powerful and well-adopted as Pinterest is,” he says, “you kind of need to live in a different, completely independent domain where the content is consumed but also have some hooks into that social graph.”

“In some ways, it doesn’t really matter where the underlying graph is from. Facebook and Twitter are just proxies for real-life arrangements,” says Mark Johnson, CEO of personalized social graph-powered appazine Zite (acquired by CNN in 2011). “If Pinterest is where people are going to share and talk and have fun, then Pinterest, in some sense, actually owns the graph more strongly than Facebook at that point. Facebook becomes the plumbing.”

“Pinterest is in an interesting position,” concludes Moore. “It’s a social curation platform.”

There’s some evidence that the company’s growth is slowing. AppData, which monitors third-party apps and sites that interact with Facebook, says the number of Facebook-connected users has fallen in recent months, perhaps as much as 25 percent. Still, “every day it’s continuing to reach critical mass,” says Jason Hennessey, CEO of EverSpark Interactive. And it’s using that mass to sustain itself with affiliate revenue. Pinterest uses Skimlinks technology to crawl all the images posted on the service back to their original sources and look for affiliate codes, then inserts its own tracking code “so that Pinterest is actually monetizing the traffic that it sends to these Web sites,” says Hennessey. “It could become one of the largest affiliate sites on the Web in a very short period of time.”

The use of Skimlinks alarmed some users, as has Pinterest’s somewhat blasé attitude toward copyright and fair use. “We’re at this moment in time where there are some terms and policies that need to be tightened up,” says BlogHer’s Elisa Camahort Page. “There are some bloggers who are concerned about how easy it is to really share not just copyrighted content, but so much content that you don’t need to go visit the site.” Also of concern to content producers, says Page, is the way Pinterest repurposes their content. “Where they’re different from many other search and bookmarking sites is in enabling people to review full images within Pinterest and not have to leave to go see the full image.” Though this fear of Pinterest locking people into the site would seem at odds with apparent plans to generate revenue from affiliate links.

To be clear, Pinterest is no danger from copyright lawsuits whatsoever. After reviewing Pinterest’s latest posted Terms of Service and copyright policies, Derek Bambauer, a professor of law at Brooklyn Law School who specializes in intellectual property and Internet law, tells MEDIA, “Pinterest has complete protection as long as they stay within the DMCA safe harbor. Every UGC site survives based on that protection. The cost of DMCA compliance could be significant, but it can be automated almost entirely.” As long as this is the case, any potential suits would be dismissed as soon as they came before a judge. “Pinterest would have to pay to defend initially,” says Bambauer, “but most plaintiffs’ lawyers know those suits aren’t meritorious.”

Page feels that Pinterest has a greater responsibility to bloggers and its users, though. “What it’s trying to say, is ‘Hey, we have no legal liability and the users have all the legal liability,’ ” she says. “Considering that its whole reason for being is that it’s an image-sharing tool, that’s a pretty cavalier attitude to have toward your users.” To assuage some concerns, Pinterest did introduce a simple “no-pin” snippet of code so sites could opt-out, but in reality, not many sites are using it. At least, not those interested in getting traffic.

Pinterest and its many aforementioned hell-spawn and whatever follows them through the gates are surely becoming a part of the digital ecosystem, and trying to banish them would be like a farmer trying to banish microbes from his soil. Both can be beneficial, of course.

One way Pinterest can be beneficial to marketers, besides the obvious targeted referral traffic it generates, is that, “you’re getting SEO value from the links you’re continuing to build on Pinterest,” says EverSpark’s Hennessey. Each of those shares that links back to your site is another pathway for search ranks to take into account. In a bit of circular logic, your rank on Pinterest and how likely people are to come across your content are also controlled by an algorithm. “The more times people click on images and start pinning them, the more popular they become. Unlike Twitter where if seven people retweet your tweet, in another 20 minutes it’s still gone,” says Hennessey. “It’s kind of like Google in a sense, where the more popular your images become the more visible they become over time. So if the images stay popular, they’ll have a much longer lifespan.”

There are other ways well beyond this that the data and metadata generated by Pinterest-like services can become useful. It can play a role in reordering and defining all the content on the Web by adding an extra layer of metadata.

One of the major drawbacks to the near-utopian dream of a semantic Web has always been that it somehow asks that all this metadata be added to existing and future content. “A definition doesn’t necessarily come from an intense tentacled description of something; it comes from usage,” says Mark Johnson. “And that’s where many of the algorithms that have made a lot of money, like the Google search algorithm, have succeeded. There were no extra semantics that Google put onto the Web. It was actually the internal structure of the Web that made the algorithm so powerful.”

The next step could be in further defining the millions of bits of content we produce every day through, as Johnson suggests, usage. “If you can encourage people in ways that are fun to add a lot of metadata to the Web, then you probably have something there. The challenge is mining it,” he says. “Pinterest is, in fact, adding a ton of metadata to a swath of the Web that never really got a lot of it. You imagine that there would have been enough metadata on pictures already that you could have done something Pinterest-y with it. Though it’s really about drawing things together with a human editorial eye.”

A smudge pot — used by farmers since the 1930s to heat their fields when crops are in danger from frost — is generally a diesel-burning rusted hunk of tin that looks like it dropped off a World War I battlefield. There are many of these in the area where I live, and I find their rust patterns and dents visually interesting, so I created a pinboard full of images of them. Now, other Pinterest users come to the board and repin that content, and when they do they add their own descriptions and tags. One, placed on a badly banged-up smudge pot slouching in a field, caught my eye: “Lovely.”

A million monkeys typing tags on that image for a million years might never come up with that, but it somehow fits. Posits Johnson: The “Google for social media” that uses the social layers of data overlaid on content and filters and organizes our modern world, might make serious hay of this sort of human element.

He offers the example of the classic linguistics problem of figuring out humor and sarcasm through algorithms. “It could be hundreds of years before we do that. On the other hand, it might happen very quickly … instead of having to read the documents themselves, suddenly you’ve got this extra data set that’s being created out there. What looks like a hard computer science problem becomes a much easier problem if you look at all this metadata.” Social streams and sharing tools, with their lolz and hahas are all providing a rich mix of metadata that, when layered on top of stories, could provide a more simple and straightforward solution than previously thought possible.