At BoostCTR, Crowdsourcing Brings a Human Touch to Search Ads

4/16/13Follow @wroush

Say you’re a human-resources manager inside a big company and you’re looking for a way to say thanks to an outstanding employee. You go to Google or Bing and you type in the search term “employee recognition certificate.” On the right side of the search result page, there’s the usual column of text ads. The first one says:

Browse Certificate Papers. Create Custom Certificates Using Premium Papers. Hundreds of Styles.

That might be what you want, but it’s not terribly inspiring. But the ad right below it says:

Certificates They Deserve. Imagine Their Faces When You Hand Them Their Personalized Certificate.

Which ad would you be more likely to click on? It’s a no-brainer—the second ad is far more compelling.

What’s interesting, though, is that the first example is a real ad written by a search engine marketing (SEM) professional, while the second was written by a freelancer doing occasional work for BoostCTR, a San Francisco-based ad optimization startup.

It turns out that in the world of SEM, the human element is still pretty important. If your company sells hundreds or thousands of products and you’re creating keyword-targeted ads on Google AdWords or Bing Ads or Facebook, you probably don’t have time to think about each individual ad as a creative challenge. BoostCTR has set out to solve that problem. It’s a Web-based crowdsourcing platform similar in many ways to oDesk, Elance, Amazon Mechanical Turk, or MobileWorks, except that its far-flung workers focus exclusively on generating better search-ad copy, and compete with each other to write the most effective ads.

In measures of click-through rates (CTR) and revenue per impression, ads penned by BoostCTR’s part-time copywriters outperform ads written by in-house marketers by an average of 30 percent, according to David Greenbaum, the startup’s co-founder and CEO. He thinks that’s because outsiders are less immersed in the details of a product’s features and more attuned to the likely audience’s needs.

A November 2012 group picture at BoostCTR's San Francisco office.

“Conveying the benefits over the features and connecting on an emotional level is what makes for a great ad,” Greenbaum says. “Motivation and the impetus to think about things creatively turns out to be more important than deep familiarity with a product.”

More than 200 companies that live or die based on search advertising—including familiar brands like CafePress, 99designs, Bonobos, and Lumosity—use BoostCTR to improve on the performance of their in-house SEM operations. Every time a customer submits an ad that needs improvement, BoostCTR mounts a contest on its Web platform, soliciting ad-copy suggestions from its network of writers. Automated A/B testing algorithms measure which ads perform best, and writers get paid when they win a contest or their ads go live (each one must be approved by the advertiser).

Writing a single ad should take only 3 to 5 minutes, according to Greenbaum, and good copywriters can earn $15 to $25 per hour.

“Say you’re a stay-at-home mom and you have the opportunity to write about a product that you’ve seen on the Home Shopping Network,” says Greenbaum. “I want for them to work in our system for an hour a day, a couple of days a week. It’s an enjoyable way to make some money. And by the way, if you are the Home Shopping Network, that is exactly the type of person you want writing for your product, rather than some fancy agency person who probably isn’t a real HSN shopper.”

Greenbaum co-founded BoostCTR in 2009 with Rob Lenderman and Vince Roche, and the company has collected roughly $3.6 million in venture backing from a group including Javelin Venture Partners, Silicon Valley Bank, Metamorphic Ventures, 500 Startups, Founder Collective, and angel investors.

It’s expanding fast—its team of about 20 people moved into a cavernous office on Sutter Street last fall. But the idea had humble beginnings. Back in 2007, Greenbaum and Lenderman were living in Miami and working for Interval International, a huge operator of time-share vacation properties around the world. As part of Interval’s M&A group, Greenbaum helped oversee the acquisition of ResortQuest, a big hotel chain in Hawaii (it now goes by the name Aston Hotels).

The first thing Greenbaum noticed after the purchase, he says, was that “their online marketing efforts were absymal. They hadn’t optimized their campaigns in any way, and they were only taking advantage of a fraction of the opportunities open to them.” He talked over the problem with Lenderman, who had experience doing SEM for Lendingtree.com, and the two realized that they needed to rewrite the ad copy for almost every group of search keywords that the chain was bidding on.

Lenderman “had the really smart idea of setting up a game where everyone in the company, from the administrative assistants on up to the VP of our group, would write an ad. The first thing we noticed was that it was really fun—everyone was talking smack about how their ad was better,” says Greenbaum. More importantly, about two weeks into the game, Lenderman’s data showed that the ads Interval’s staff had dreamed up substantially outperformed ResortQuest’s existing ads.

Greenbaum says the lesson he and Lenderman learned from the exercise was that “if you take people who are motivated and interested and aware of what a product is, and have them match wits against professional marketers, the ordinary people can write better ads.” At that point, Greenbaum says, “A bell went off in our heads, and we said, ‘Hey, there is a really cool, powerful dynamic that we’ve hit on here.’”

To fast forward through a lot of company history, Greenbaum and Lenderman incorporated BoostCTR, recruited Roche as a third, technical co-founder, moved the startup to New York, started building a network of copywriters, attracted the attention of ex-Googler Satya Patel and other angel investors, moved again to San Francisco, and eventually won $1.6 million in seed funding from Javelin and other backers.

At first, according to Greenbaum, the company pitched its services mostly to small and medium-sized businesses, thinking that it would be better to work the kinks out of its platform with less demanding customers. But as the startup won introductions to a few larger customers, it became clear that big companies were getting a larger performance boost from the crowdsourced ad copy than small ones. So in 2012, the company switched its focus to the enterprise market.

Greenbaum’s theory about why big companies benefit more has two parts. First, as companies get bigger and do more search-based advertising, they usually don’t scale up their marketing departments accordingly. Even a giant retailer or online catalog company might have only half a dozen people managing campaigns across hundreds of thousands of products, keywords, and ads, he says. These people simply don’t have time to think up appealing ads for every single product, so they end up using lots of generic, templatized ad copy, along the lines of “Buy ABC at XYZ retailer today. Get yours now!”

Second, Greenbaum says, full-time SEM jobs tend to be filled by spreadsheet jockeys and other people with a quantitative, analytical bent. These types of thinkers tend to fixate on a product’s features, rather than trying to imagine how it might benefit its users. In other words, they aren’t very good at applied psychology. BoostCTR’s freelance copywriters, on the other hand, don’t need to apply anything—they’re consumers, meaning they’re often the very same people for whom the ads and the products are intended.

But they’re consumers with a proven ability to write effective ads—and to adjust their approach based on data from BoostCTR testing platform. “They get very quick feedback based on how much better or worse their idea is, relative to the baseline ad and relative to other writers, which is pretty unique,” Greenbaum says.

In fact, Greenbaum thinks this is the key ingredient that differentiates BoostCTR from other crowdsourcing platforms and makes it possible for non-professionals to generate effective ads. “It started off with Mechanical Turk, which was generic and had low-skill participants,” he says. “From there you had Elance and oDesk, which had higher skill workers and more specific work flows. Then you have 99Designs, which is even more verticalized. The evolution we are driving is, we are the first company to put a performance feedback loop on a flexible labor market.”

Right now BoostCTR optimizes ad campaigns for Google, Bing, and Facebook, but eventually the company hopes to be able to produce better ads for more platforms, including Twitter and Linkedin. And since the operators of performance-based ad platforms make money when their advertisers make money, they have a vested interest in seeing BoostCTR succeed. “We are going to have tighter and tighter relationships with each of the ad platforms, and the ultimate objective is to have them offer us to their customers as a service,” says Greenbaum.

If its network writers can keep the whole cycle going, BoostCTR may just be in the market for a few recognition certificates of its own.

Wade Roush is a contributing editor at Xconomy. Follow @wroush

By posting a comment, you agree to our terms and conditions.