Your web analytics report is filled with clues on how to boost sales.
If he were alive today, Luca Pacioli would probably love poring over a Web analytics report. Pacioli, a Franciscan monk who published the first balance sheet in 1494, understood the power of data to transform the way companies make decisions. Today, thanks to the Internet, the amount of data available to any company with a website dwarfs anything Pacioli could have imagined. Web analytics programs track nearly everything that happens on a website, keeping a real-time record of how potential customers find the site, how they behave on it, and why (or why not) they buy stuff.
The wealth of data is so great -- you might learn, for instance, that customers in Milwaukee typically spend three minutes on your site and use Firefox -- that many entrepreneurs find it overwhelming. "If you try to look at every piece of data, you'll go crazy," says Andy Beal, an online-marketing consultant and co-author of Radically Transparent, a book on the power of new-media tools to build a brand.
To get a sense of how to use analytics to uncover your website's flaws and make more money, we asked a fast-growing start-up to let three online-marketing experts review its analytics reports. The panel included Beal; Reid Carr, CEO of Red Door Interactive, an online-marketing firm based in San Diego; and Avinash Kaushik, from Google's (NASDAQ:GOOG) analytics division.
Our guinea pig, Bonobos, is an apparel company founded a year ago by Brian Spaly and Andy Dunn, both Stanford M.B.A.'s. Based in New York City, Bonobos sells fitted men's pants that, according to Dunn, avoid "the khaki diaper butt problem." The company sells its products on its website (it has no retail distribution), and prices typically range from $110 to $190.
By September, when we looked at Bonobos' analytics data, the company had already gotten off to a fast start. The company was booking about $170,000 a month in sales. What were the keys to Bonobos' success, and how could the site be improved? Here's what the numbers told us.
A look at your traffic mix can help you measure the effect of marketing activities, both online and off. Normally, the biggest chunk of a site's traffic comes from search engines, followed by roughly equal streams of traffic from referring websites and from direct traffic -- that is, people who type in a Web address. But Bonobos receives less traffic from search engines than from either referring sites or direct visitors. This happens because the company invests heavily in public relations. Look at the site's daily visits: On August 11, Bonobos was mentioned in The New York Post, and the site logged 2,238 visitors, up from 652 the day before. A similar jump occurred on August 26, when Bonobos was on local TV.
These spikes are nice, but PR can be fleeting and does not always result in long-term traffic. And if you look at Bonobos' daily conversion rate, the percentage of visitors who buy something, you see that on August 11, it fell to about 1.8 percent. The site may sell more even with lower conversion, but, moving forward, Bonobos should complement its PR strategy by going after more traffic from Google and Yahoo (NASDAQ:YHOO) searches, which typically converts to sales at a higher rate than referring-site traffic.
Our experts noticed the high average number of page views -- nearly eight per visit -- that Bonobos.com generates. You may think the more page views, the better. That's certainly true at big entertainment sites, such as YouTube and MySpace, where clicks directly translate into more ad revenue.
But for an online store or an informational company site, it's not that simple, says Reid Carr, whose clients include Overstock.com and Intuit (NASDAQ:INTU). Given the site's bounce rate, which shows that nearly a third of all visitors leave after only one page view, Carr expected an average number of page views of six or fewer. The fact that the average is 7.75 "suggests that people can't find what they're looking for," he says.
To figure out how to simplify and improve the site's navigation, Carr recommends that the company engage in "A/B testing," whereby it serves test pages to a sample group and tracks how they perform in comparison with the current site, to see what tweaks reduce the page-count totals.
A bounce rate in the 20 percent range is good. Bonobos' rate tops 30 percent and, because a company blog is featured prominently on the site, Avinash Kaushik guesses that the effective bounce rate is even higher, since some visitors are probably reading the blog and then leaving the site.
If Dunn and Spaly can bring down their bounce rate to 20 percent and maintain a decent conversion rate (the number of visitors who make a purchase), their sales could increase by as much as $10,000 a month.
How to do this? Marketing consultant Beal believes that Bonobos' homepage devotes too much space to PR hits, to the blog, and to goofy product names such as Snapdragons and Jive Cats. Meanwhile, different styles of pants are not grouped by categories such as fabric or style. "They should make it easier for a visitor to figure out where to find corduroy pants or weekend wear," Beal says. "A new customer is not going to know what 'Snapdragons' means."
As one might expect for a New York clothing company, most of Bonobos' traffic comes from North America. However, the site's map overlay shows that it attracts a significant number of visitors from Australia. Beal recommends that Bonobos capitalize on this fact by adding a currency calculator, touting the availability of international shipping, and eventually building a separate international site that sells Bonobos "trousers," reflecting the most commonly used term in Australia to describe pants.
The Goal Funnel tracks customers during the sales process. This funnel shows that among users who put items in the shopping cart, 63 percent (or 1,222 people) failed to complete a purchase -- and 511 of them left the shopping cart and exited the site altogether. The problem, Kaushik believes, is that the site asks for a coupon code early in the process. "That's an invitation for someone to leave your site to look for a coupon," he says. "And if they don't find it, they may not return." By taking steps such as moving the coupon field to the end of the sales process, Bonobos' rate of abandonment could fall 10 percent, adding about $20,000 a month in sales.
Keywords are the search terms that most often lead people to a site. Bonobos' top keywords show that nearly all of its customers search for it by name; few people stumble upon the site while looking for "men's pants" or "pants that fit." This makes sense given Bonobos' aggressive PR strategy. It also suggests a way to drive more traffic. A basic search-engine optimization program could begin to draw in people who have never heard of the company but are looking to buy pants.
Another number caught Kaushik's eye: the site's bounce rate for "content targeting" -- meaning ads placed through Google's AdSense or similar programs -- is, at 55.1 percent, quite high. "It doesn't look like the traffic is converting well," he says. To lower that rate, he recommends that Dunn and Spaly try tying their ads to a different set of keywords or rewrite their advertising copy.
"Our reaction is, Hallelujah!" says Bonobos co-founder Andy Dunn. "We couldn't agree more." Dunn concedes that the low conversion rates and high bounce rate are problematic but says that makes sense, given that the goal of the start-up's website has been to tell the Bonobos story. The founders decided to play up the blog and their PR because they didn't want visitors to think they sold existing apparel labels such as Polo or Levi's. "Our problem is that we want to emulate large e-commerce sites like Zappos and Amazon (NASDAQ:AMZN), but we also want to be a brand," he says.
The next iteration will probably take steps to improve conversion and bounce rates by organizing the pants by fabric and style and by adding a navigation bar similar to what you find at traditional online stores. Dunn loves the idea of offering coupon codes later in the sales process but disputes the suggestion that the high number of page views reflects customer confusion. "My read is that our site is engaging," he says, citing an online contest that attracted close to 200 entrants.
And about that mini-boom of Australian Web traffic? Dunn believes that was the result of the company's being mentioned on the blog The Cool Hunter, which is popular in that country. Bonobos will eventually create international sites -- at that time, Dunn promises, the all-important pants-versus-trousers question will be revisited.