Web Site Expert巻頭レポート(英語)

Google Analytics’ Brett Crosby: Moving Customers Down the Sales Funnel

この記事を読むのに必要な時間:およそ 8.5 分

Google⁠s Brett Crosby has two essential pieces of advice for webmasters, especially those running e-commerce websites: know how your users arrived and understand what they did after they got there. What link did they click on to get to your site? What links did they click on within your site? Did they make it all the way through to the checkout section, or did they leave for somewhere else? If they left, where within your site did they depart? Crosby, a senior manager at Google, argues that understanding such user behaviors can lead to more sales “conversions⁠―that is, more visitors who are converted into paying customers.

Crosby came to Google after the company bought his company, Urchin, in April 2005. Urchin started as a website development and hosting company, but the same technology eventually proved useful for analyzing user behavior. Urchin⁠s co-founders were Brett⁠s brother, Scott; a childhood neighbor, Paul Muret, who grew up to be a math prodigy and programmer; and a college fraternity brother, Jack Ancone. With the purchase, the company moved north from San Diego, and Crosby now makes the hour-long commute from San Francisco to Mountain View on a Google shuttle with an Internet connection. When he arrives, he works out of the former office of Google founders Larry Page and Sergey Brin―in a company known for its free food, lava lamps, and dogs who accompany their owners to work.

The company now offers the analytics technology free of charge, figuring that the more people understand about the effectiveness of paid searches on Google, the more searches they will pay for. But Google Analytics extends well beyond Google―it can record any referring site, as well as any activity on any website. And while analytics was once the domain of technical people in the IT department, Crosby says that the technology has gotten much easier to use. The principal users of Google Analytics are now the readers of this magazine: webmasters and other website experts.

Where did the idea of analytics―and Urchin―come from?
We started out as a Web hosting service. Back then, bandwidth was extremely expensive and we needed some way of measuring it so that we could bill clients appropriately. We originally did this by using log files―which record visitor interaction with a website. Today we use a modified version of this technology. We put a little snippet of code on your site that allows us to create a log file on our server.
The broader idea of analytics started when people realized search engines were sending traffic to their websites and therefore wanted to raise their ranking in the search results. So the search engine companies got smart and said: why not let people pay for ranking.
So once search engines have driven traffic to your site, analytics tells you…..what?
On Google, there are two common ways that people get to your website. The first is what we call an organic search, in which Google crawls the web and, through algorithms, determines which sites are most relevant to a search. The second is paid ranking, in which companies bid on key words. If those key words are sending you traffic that converts well, you will want to reinvest in those keywords. If not, you⁠ll want to reinvest in other keywords that work better. Analytics helps you understand how well both of these searches―organic and paid―are working for you. It does this by tracking the URL of the link that people used to get at your site.
How does Analytics know that?
Your browser passes that information on once they arrive at your site. It essentially says: “here is where I came from to get here.” They might arrive from Google or another search engine, or via a banner ad, email, an online video, even a blog.
What should a webmaster do with this information? What value is it to know where people have come from?
The bottom line is that you want quality visitors coming to your website. You determine that by considering ROI [return on investment] metrics―how many visitors are coming from a referring source, and how many of those visitors are converting to paying customers. When you know that, you can begin to target the places that are referring you the most valuable traffic. That⁠s one of the reasons that search engine marketing has become a big phenomenon over the past few years: it⁠s accountable. In addition, analytics can also take better advantage of the free referrals. Getting mentioned on a site like Digg.com, in which articles are submitted and voted upon, or Slashdot, can create a tremendous amount of referring traffic. Web analytics can help you see what types of things are picked up by these communities.
What does the actual Google Analytics report look like?
We provide a lot of information. You can see where people are coming from, and if they are coming from Google, the pages within Google that are sending the traffic.
You can see how many visitors came from a given site, how much revenue that reference drove, what the ROI calculation was, the percent of total visitors that came from there versus everywhere else, and whether the traffic from those sources converted on various goals on your website. You can see that different referring sources may send different types of traffic―that people who come from different sources may tend to do different things on your website.
And once they arrive at your site, does analytics track every link they click on?
Yes. As people go through your website, we can see each page that people click on. It⁠s like footprints in the sand: we don⁠t necessarily know who made the tracks, but we know where the accumulated footprints went. You can drill down to an individual level: often times that⁠s valuable if you are trying to understand the behavior of a single individual. But more often, it⁠s interesting to look at aggregated general traffic patterns through your website.
Once you understand user behavior, how do you know what actions to take in response?
We make a huge effort in making the Google Analytics data be extremely actionable. But it⁠s important for people looking at the information to have a familiarity with the website and understand how they are going to drive traffic, and what they are trying to accomplish. So when they see weekly visitor or sales figures, they understand whether the numbers are good or bad. That means understanding the benchmarks and history of the site.
You talk about user behavior in terms of a “sales funnel.” What is it?
A sales funnel is a way of looking at how people move through your website. At the rim is the landing page where people actually enter the website. At the spout is the checkout and the “thanks for buying” page. Google Analytics allows you to examine how the sales funnel is working for you. You can watch how people enter the funnel and see where they exit. Hopefully they exit after they have bought something, but if not, you can see where they left the site. From there, you can “fix” the holes in your funnel. The sales funnel is really about sales conversions, and that⁠s one of the first places we recommend people look when they are ready to improve their site.
There are few different ways to look at funnels. Maybe you just look just at the results―whether people converted or not. Or maybe you examine the steps in your shopping cart, where customers specify the size and quantity of their purchases, credit card information, color, size and quantity of purchases, and shipping preferences. The level of detail is up to you.
To help understand all this, we have a highly visual report called Site Overlay that provides a visual look at user activity on any page of your website. Next to each link, we overlay two small bar graphs. They indicate where people clicked―thereby showing the most popular clicks―and how many of those clicks led to conversions. Sometimes the results are unexpected. You may have a link where few people click, but those that do have a high rate of conversion. There⁠s a lot of additional information, as well. The tool is especially for people who are creative and visual, rather than intensely analytical.
You seem to be implying that different people within the company now use analytical data.
This used to be more of a technical thing done by the IT department. But as online marketing became more of a real business, people started caring much more about tracking activity. The biggest change occurred around the time of the “dot.com bust.” That⁠s when people stopped throwing money wildly online, buying banner ads and just hoping that something good would work. Back then, people didn⁠t care about conversion rates, they just cared about the number eyeballs viewing their ads. But once marketing budgets shrank, people needed better ways of tracking their online marketing dollars. They needed to be more accountable in justifying what they spent.
So in a way analytics represents the maturity of the business model―and of the maturity of the web itself.
Exactly. The dot.com bust was the spark. At the same time, analytics shifted from IT to marketing. These days, the technology makes it so easy for people to track what⁠s happening with their sites. The reason the on―demand model Google uses is so popular is that it is extremely easy to set up.
How easy?
For each page that you want tracked, you add a bit of JavaScript code. You can put it in a header or footer file or you can use a server site―there are several techniques. Then you set up your goals and funnels. To do that, you copy the URL of the pages, identifying the steps of your funnel leading to the goal page―that is, the desired path you want people to take for a sales conversion. The first page may be a landing page or the home page of your site; the second page could be any product page―all the way through to checkout.
The steps can be more general. You can say “anything within this directory,” for example, or you can use filters. There⁠s a lot of customization, but you can also keep it very simple and still get all of this information.
What does the snippet of code actually do?
It sets a cookie on the visitor⁠s browser which allows you track if that person comes back to your website using that same browser. The idea is that if a person eventually buys something, you can still attribute that sale to the original keyword or banner ad or whatever drove them to the site in the first place. The code also makes a call to our server where all this activity is stored. We use that information to create the reports.
Typically, what kind of time investment is involved in getting set up?
That⁠s one of the things we work very hard on with Google Analytics. The setup time, really, is in tagging every page on your site. That could be a matter of a few minutes if you know what you⁠re doing.
Is the real investment then in analyzing the data and figuring out what you do in response?
Exactly right. Setup was a lot more cumbersome a few years ago. Now, we want people to spend their time analyzing the data and making decisions based on the data so they can optimize their marketing and improve the user experience on their website.
What kind of changes do people make in reaction to the data?
People make all kinds of changes. The most obvious one is they optimize their marketing spending―they buy what works and they cut what doesn⁠t. They cut out the advertisements that don⁠t work, keywords that don⁠t work, banner ads that aren⁠t converting for them. They get rid of those or modify them and try other things. People are always testing new approaches.
What kind of changes should they be considering to their websites?
The first thing to look at is where people fall out of the funnel. To do that, you can pull up the Site Overlay and see if a lot of people are clicking on something that doesn⁠t result in a conversion.
It is also important to see if people are clicking in, then immediately bouncing off your website. If so, why is that happening? Are they looking for one kind of content, and finding something completely different? You can drill down and see a bounce rate for that page. Or perhaps the page is poorly designed or you are sending the wrong kind of traffic to that page. For example, imagine a visitor to a shoe website who gets there while searching for open―toed sandals―only to be directed to the catalog⁠s general home page. That⁠s a bad user experience. You can quickly increase the conversions by sending them to the exact page they are looking for.
Any specific advice to Web Site Experts in Japan?
One difference is the language factor. If you are targeting people beyond Japan, you will want to have a multilingual website. We can help you with this by telling you in the reports which languages people have as their default in their browser. The other thing is that if you are buying keyword traffic, and you only care about Japan―specific customers, you should indicate that somehow in your ads. That way, people don⁠t click on your advertisement and waste your budget if they are not ever going to become a customer.
And of course, Japan is well ahead of the West in using mobile devices to search the Web. So in Japan, and Korea too, it is very important to have a mobile version of your website. That is something else we can tell you in Google Analytics: the type of platform and browser people are using to visit your site.
Finally, I want to ask about privacy concerns. I know people who frequently erase their cookies because they don’t want to be tracked. Is that a problem for websites? Is it a legitimate concern for users?
The people who are clearing their cookies will definitely make the reporting slightly less accurate. We think that of all the people that visit a site, about three to six percent will have cleared their cookies. That leaves some small margin of potential error.
As for the people clearing their cookies, I think that⁠s fine. There⁠s a certain amount of healthy paranoia on the Web. I don⁠t think it⁠s ever going to go away, and it keeps people honest. We recognize that and we⁠re huge advocates of user trust and user experience on the web. Sometime it⁠s a balance between the two because if you delete too many of your own cookies then it will harm your own user experience. For example, if you go to your banking site all the time and you want it to remember you, first party cookies could be a very good thing.
So maybe the question for users is: how much do you trust the site?
You definitely want to be ethical in the way you treat visitor data. That means posting privacy policies that explain what you do with your data and respecting user privacy. You can design your site to help instill that feeling―by qualifying for a Trust―e or Verisign seal, for example. The more people feel confident about how you treat their information, the higher your conversion rate will go.


Bart Eisenberg

Bart Eisenberg's articles on the trends and technologies of the American computer industry have appeared in Gijutsu-Hyoron publications since the late 1980s. He has covered and consulted for both startups and the major corporations that make up the Silicon Valley. A native of Los Angeles and a self-confessed gadget freak, he lives with his wife Susan in Marin County, north of San Francisco. When not there, he can sometimes be found hiking with a GPS in the Sierra, traveling in India, driving his Toyota subcompact down the California coast, or on the streets of New York and Tokyo.


1980年代後半より,『Software Design』や『Web Site Expert』などの雑誌に,アメリカのコンピュータ業界のトレンドと技術に関するレポートを執筆しています。シリコンバレーで,スタートアップ企業から大企業まで幅広い分野でコンサルタントを務めました。

ロサンゼルス生まれで,自称ガジェットフリークです.現在,妻のSusanとともに,サンフランシスコ北部のMarin County在住。また,SierraのGPSを携えてハイキングしたり,インドを旅したり,カリフォルニア海岸をドライブしたり,NYや東京の街中を歩いたりしています。