Tag Archives: web analytics

Web analytics databases keep getting bigger

Who has the biggest database? Due to the increasing amount of behavioral information tracked during a web browsing session, some internet properties are starting to rack up some pretty hefty databases.

Ebay has a 6.5 petabyte Greenplum warehouse and a 2.5 petabyte Teradata warehouse. This system ingests hundreds of billions of new rows of data every day.
Facebook has a 2.5 petabyte Hadoop system
Yahoo has more than 1 petabyte running on their homemade system

Traffic measurements vendors differ widely in results

In an article from Hitwise today, Hitwise used their data to determine that Craigslist had made it to the top of the weekly search results list beating 3 year champ, MySpace.
However, Techcrunch has a very different view based on a look at Google Trends. They rank Craigslist third with Facebook having a huge lead over MySpace. With different sites using different tracking methods, it’s no wonder we’re seeing such varied results. Keep your eyes open.

How to get more info on who is visiting your site

In addition to any on-site web analytics tracking that you might do using Google Analytics, Coremetrics, Omniture, or other solutions, there are a number of vendors who use other methods to provide information about who is visiting your website.

Recently Robert Scoble at Fastcompany TV sat down with the CEO of yet one more vendor, Quantcast. Check out their interviews here: Quantcast Part 1, Quantcast Part 2

Omniture implements Twitter tracking

Check out the blog post from Omniture about their new Twitter tracking capabilities.  And is a sample of how a developer could implement the solution.

These are the 4 key questions the solution seeks to answer…

  1. How often is your company mentioned on Twitter?
  2. Is there ever a spike (positive or negative) in brand-related terms (in a week, day or even hourly)?
  3. Who are the people most often mentioning your company on social media tools and who are they communicating with the most?
  4. When are people on social media tools mentioning key product/service features that your Product Managers should know about?

Beware of the costs of free web analytics software

While many web site operators are using free web analytics tools like Google Analytics to monitor visitor behavior, many sites lack the resources to get sufficient value out of the tools, research and consulting firm Web Analytics Demystified says in a new report issued this week.

?Web analytics is hard, and when companies aren?t fully invested in understanding what analytics data mean, they?re likely to make wrong decisions based on it,? says Eric Peterson, founder and CEO of Web Analytics Demystified. Peterson co-authored the report, ?The Problem with Free Analytics,? with Zori Bayriamova, a former research analyst at JupiterResearch who is now research manager for Time magazine. ?One person?s interpretation of data can give a thumb?s up or down to an entire project, but that person may not be getting the big picture and making decisions only on what he thinks is important,? Peterson adds.

Peterson contends that use of free web analytics tools often coincides with a failure to compensate for the lack of support services routinely offered with licensed tools to help interpret analytics data, leaving the users of free tools with less useful information. ?Site operators need to realize that nothing is really free, so if they save on the technology, they need to reinvest it in people to manage the free technology and data analysis process,? he says.

The study found that among users of free web analytics tools, 35% deploy them on only an ad hoc basis, compared to 19% for users of licensed tools, and that 25% use them for tactical and strategic decisions, compared to 39% for licensed users.

34% of study respondents using free web analytics tools, compared to 55% for users of licensed tools, said their analytics tools answer 50% to 89% of their questions, the study found.

Peterson notes, however, that Google Analytics is a valuable tool with an easy-to-use interface for users who can properly interpret its analytics data. ?Google Analytics is good at data visualizations and it has an especially good interface,? he says.

Google reported recently that it had added improvements to the Google Analytics interface, including a new ?Segment? menu that lets users organize data pages by attributes including referral source, keyword and visitor type.

Although the report does not mention other brands of free web analytics tools, Peterson notes that others are available from vendors of licensed software, and that Microsoft Corp. is expected to release soon a beta version of a free analytics tool called Gatineau.


Making sense of web analytics data

Although web analytics data can be overwhelming and difficult to interpret, retailers can take steps to understand it, apply it to improving web site operations and share its value across departments, experts said at the Shop.org conference.

Web analytics can bring more questions than answers, said Kim Weller, senior manager of web analytics for multi-channel consumer electronics retailer Circuit City Stores Inc., who appeared on a web analytics panel. Weller was joined on the panel by retailers Dylan Lewis, senior manager of web analytics for the Intuit Turbo Tax unit of financial software retailer Intuit Inc. and Michael Fried, director of strategic analytics for outdoors sports gear retailer Backcountry.com.

Indeed, 80% of retailers in a recent study noted that they have doubts about the accuracy of their web analytics data, said Eric Peterson, CEO and founder of consultants Web Analytics Demystified, who co-moderated the panel with Patti Freeman Evans, senior retail industry analyst at JupiterResearch.

Retailers on the panel said it’s important to organize data to make it more useful, but to also get cooperation from multiple operating departments in interpreting and applying the data.

You can lose focus on the true goal of serving customers by concentrating too much on isolated key performance indicators without considering the overall context of a web site’s activity, noted Lewis.

We put the same focus on web analytics data as on financial data, said Fried. The retailer breaks down analytics data so that it can compare apples to apples in viewing changes in site activity, and it also combines analytics data with the results of direct surveys of customers about their shopping preferences.

At Circuit City, a careful review of analytics data revealed that shoppers were struggling with the checkout process, which enabled ther retailer to identify where to make necessary changes. Likewise, Intuit made improvements after it realized that many shoppers apparently weren’t noticing certain text intended to help make purchase decisions.

It’s a challenge, however, to make sure that all people within a retail organization charged with using analytics data actually know how to interpret and use it, Fried said. We teach people how to use it and what the value of it is, he said. Then we hold them accountable to actually use it.

To get started, it can be helpful to focus on one particular metric of site operation, such as conversions of visitors to buyers in particular product categories. Then create a flow chart to show with analytics data what leads to either an increase or decrease in that metric, Lewis said.


Web metrics that actually add up

By: Rebecca Lieb

Everything on the Web is measurable in so many disparate ways, it’s easy to see why many marketers despair they’ll ever get a real grip on Web metrics and analytics.

Anyone who’s ever compared two analytics solutions side by side knows even the most basic stats can differ considerably, often by a wide margin. And of course, there’s no all-in-one solution. Online marketers and advertisers rely on an array of packages and solutions to measure and analyze their Web sites, e-mail campaigns, blogs, RSS feeds, branding, conversions, and anything related either to online-to-offline or offline-to-online activity.

Web measurement is strikingly similar to counting in Japanese. I recently hit a wall in my efforts to learn that language when we reached the chapter on counting stuff.

It doesn’t suffice to learn ichi, ni, san, shi — not by a long shot. No one warned me that Japanese has one counting system for round things, another for flat things, one for floors of a building, one for big animals, another for small animals (watch out — birds and rabbits have their own category, fish are in yet another).

Think that’s complicated? It’s the tip of the iceberg.

So wacky and complex is the Japanese numeric system it’s why scholars believe Japanese is unrelated to any other language. Web metrics, the way marketers keep score, not to mention allocate literally billions of dollars, is sounding more and more like Japanese.

Analytics experts agree marketers would do well to use the analytics tools they acquire from vendors, assuming, of course, they’re adept at using these incredibly complex measurement tools (no mean feat). For research spanning the entire Web, things rapidly become much murkier.

Several days ago, the Interactive Advertising Bureau (IAB) very publicly challenged online measurement firms comScore and Nielsen//NetRatings (NNR) to open their kimonos and come clean about their respective methodologies. IAB CEO Randall Rothenberg says he’s aghast both firms still rely on antiquated panel-based measurement and what he terms “flawed media-research methodologies.”

The IAB is asking for transparency, and its demand is backed by the other major industry organizations, including the Association of National Advertisers, the American Association of Advertising Agencies, and the Online Publishers Association. The IAB claims online advertising growth will suffer if advertisers can’t count on reliable, accurate site audience and ad impression reports.

NNR and comScore measure the same things; yet to advertisers’ frustration, it often seems one set of numbers is metric and the other imperial. A recent Forrester report cites NNR reporting online search grew 39 percent between 2006 and 2007, while comScore reported 10.7 percent growth for the same period. In some cases, one company will report growth in an area when the other reports a decline.

Why the discrepancies? For starters, panels are extremely problematic. On top of that, the companies have widely differing recruitment methodologies and weight their data differently. Moreover, online measurement is nowhere near as clear cut as it initially appears.

Yet advertisers and their agencies pay a pretty penny for NNR and comScore data and stake even more on their numbers, both in terms of spend and return on those investments. The IAB’s calling for transparency, accreditation, audits, and standards and for methodologies that take the technologies of both the present and the future into consideration. As the IAB SVP and GM Sheryl Draizen pointed out at ClickZ’s Web Analytics conference this week, “A rise in the use of AJAX programming has publishers scrambling to find a metric to replace one that could be made obsolete: the pageview.”

The goal, says Draizen, is to provide real accountability and to spur the growth of online advertising and marketing.

The IAB’s method of doing this, publishing an open letter, seems hard-ass to some, not the least of all to the measurement companies themselves. I’ve spoken to people at both firms who claim they were working along these lines all along, a fact the IAB was aware of. Draizen says the open letter became open as the result of a leak, prompting the decision to publish the document.

Either way, the IAB turned up the heat, and the tactic seems to be getting the desired results. ComScore and NNR will sit down with the leading industry trade organizations on May 16.

This can’t be a bad thing. Better numbers and greater transparency mean better online advertising. That’s not just a good thing for advertisers, publishers, and the trade organizations that represent them; it’s ultimately good for the measurement companies that count this constituency as the lion’s share of their client base, too.


Who is counting clicks and how

There are lots of ways to measure Web traffic. Here’s a glance at some of the more common approaches

By Catherine Holahan (Business Week)

A host of companies are doing their level best to keep tabs on how many people are visiting a given Web site, and what those users do while they are there. Some use tags installed on a user’s Web browser, while others extrapolate surfing habits from a sample audience. A handful use a combination of several methods. Here’s a glance at a range of approaches, with a look at what each has to say about one popular online video site, Metacafe.

Alexa Internet, owned by Amazon (AMZN)

What: Provides Web-site traffic rankings based on page views and users, and data on site “reach,” the percentage of total Internet users who visit the site
How: Monitors sites visited by Web surfers who have downloaded its toolbar, which provides safety data and other information for Web sites

Strengths: Includes country-specific numbers and data on trends; information is free and publicly available

Weaknesses: Panel may be weighted in favor of those who visit affiliated sites such as Amazon; tends to favor sites popular in countries where its toolbar has taken off, such as China; excludes users who surf using Opera or AOL/Netscape (TWX); count can be skewed by technologies such as Ajax that let parts of a page change without reloading

What It Says About Metacafe: Three-month average traffic rank is 148; average reach is 0.37%; site fell 15 notches in rankings over three months, and reach declined 5%.

Compete Inc.

What: Monitors surfing behavior of a panel of 2 million consumers

How: Culls panel from people who have downloaded its toolbar in exchange for information on site safety, traffic, and financial deals; randomly surveys panelists on Internet use; buys data from Internet service providers; devising an approach that uses tracking tags, which sit on a site and monitor visits from IP addresses

Strengths: Large panel; supplements toolbar data with survey data, making it less vulnerable to bias; statistically balances panel to better reflect U.S. population; snapshots free and open to the public

Weaknesses: Traffic reported for smaller sites may differ from those sites’ own measurements if Compete’s panel fails to include enough of the site’s audience

What It Says About Metacafe: 2.5 million unique visitors in March, up 35.6% from the prior month; rank is 602, up 263 notches


What: Tracks the behavior of more than 2 million people

How: Panel made up of people who have downloaded its monitoring software in exchange for security software and the chance to win prizes

Strengths: Lage panel; statistically balances panel to reflect U.S. population; can provide detailed demographic analysis on user behavior

Weaknesses: Traffic reported on smaller sites may suffer if the panel does not reflect enough of that site’s members

What It Says About Metacafe: 5.6 million unique visitors March, 2007, up 40% over the previous month. Site rank by unique visitors: 171.

Google Analytics (GOOG)

What: Measures traffic of participating Web sites

How: Uses JavaScript tracking tags

Strengths: Tracks number of hits, unique visitors; can tell where users are and whether they performed certain tasks, such as clicking on an ad; free to site publishers, though data not publicly available

Weaknesses: It does not provide detailed demographic data such as gender, age, or income; unique visitor numbers can be inflated if a large number of people delete tracking tags on their machines by erasing cookies

What It Says About Metacafe: Not available


What: Measures Web traffic for general public

How: Uses a census-based approach, using data culled from 25 million Web users worldwide, including 10 million in the U.S.; obtains data by purchasing anonymous information on IP addresses from a variety of ISPs

Strengths: Very large panel, closer to a census

Weaknesses: Doesn’t release source of IP addresses; critics say users of certain ISPs may be biased towards or against certain sites

What It Says About Metacafe: Not available

Omniture (OMTR)

What: Monitors Web traffic and site usage for paid subscribers

How: Relies on cookies and JavaScript tracking tags embedded on Web pages to monitor IP addresses that visit Web sites

Strengths: Like Google Analytics, deduces geographic location of users, what sites they are coming from, and what users do once on a page

Weaknesses: Cookies don’t provide the detailed demographic analysis that panels produce

What It Says About Metacafe: Between 10 million and 15 million visitors per month, according to Allyson Campa, Metacafe’s vice-president for marketing

Nielsen//NetRatings (NTRT)

What: Tracks computer behavior of a panel numbering roughly 550,000 people, about 400,000 of whom are in the U.S.

How: Panelists selected via random telephone surveys

Strengths: Software sits on panel members’ desktops, tracking everything they do when on the computer, including the sites they visit, how much time they spend on particular pages, and what they are doing while on those pages; NetRatings collects detailed demographic data on all its panelists; plans to include data from tracking tags and cookies into measurements

Weaknesses: Panel is small, compared to some others; audience of smaller sites may not be reflected in its composition

What It Says About Metacafe: Roughly 3 million unique visitors in June, the last month for which the information was publicly released


What: Tracks Web surfing habits for general public

How: Uses cookies and tags to track IP addresses that visit unique sites; buys information from ISPs; has a panel of registered users, though it is unclear about its size; rates traffic and provides demographic data for more than 20 million sites

Strengths: Relies on variety of methods; provides detailed demographics including income, gender, age, and ethnicity; lets Web sites install tracking tags enabling it to better record the traffic to smaller, less statistically significant sites

Weaknesses: Data on U.S. Internet users only; launched in September, site is relatively new

What It Says About Metacafe: 2 million unique visitors a month

Internal Server Logs

What: Used by individual companies to track traffic to own sites

How: Registers IP addresses that visit the site, how long each computer spent, and whether any actions were performed on the site, such as clicking on an ad

Strengths: Server logs can elucidate where people are coming from when they visit the site, where they go when they leave, who is linking to the site, who has bookmarked the site, and which search engines have found the site. They also will mark each and every IP visit

Weaknesses: Don’t provide demographic data about visitors; numbers are often not trusted because they come from the companies themselves

What It Says About Metacafe: More than 10 million unique visitors per month, according to Metacafe’s own logs


Web analytics vendor comparison

Today CMS Watch released the Web Analytics Report, which evaluates 13 major Web Analytics suppliers based on extensive vendor research, interviews with customers across a range of industry sectors, and “hands on” testing of solutions. Purchase the report at: http://www.cmswatch.com/

The Web Analytics report is divided up into 6 main parts:

Part 1 – How to Use this Report
Provides an overview of the various sections and depending on your role in the organization suggests the best path to follow in order to get the most out of the Web Analytics report.

Part 2 – What is Web Analytics
An overview of web analytics including history and definitions of some commonly used phrases.

Part 3 – Business Case for Web Analytics

Talks about the reasons why Web Analytics is so important in an online solution by addressing business goals such as Improving Customer Service and Increasing Revenue/Audience/Operational Efficiency. It’s followed by the report addressing the various costs that come with web analytics above and beyond licensing. This section concludes by addressing why web analytics is such a worthwhile investment.

Part 4 – Web Analytics Technology and Features

This section goes into detail about the various features that were used to compare the various web analytics vendors:

– Technology (Architecture, Availability)

– Data Collection and Processing (WebServer, JavaScript tagging, Complexity)

– Data Accuracy (Verification, Sampling, User Id)

– Data Services (Importing, Exporting, Integration)

– User Administration (Permissions, Distribution)

– Reporting (Standard Reports, Usability, Advanced Features, Dashboards, Help Features)

– Vendor Intangibles (Implementation, Support, Licensing Costs, Future Development, Stability)

Part 5 – The Web Analytics Initiative

Part 5 talks about purchasing and implementation starting from the requirements all the way to implementation and best practices.

Part 6 – The Web Analytics Vendors

This is where all the major Vendors are compared using the items addressed in part 5. The following web analytics vendors are compared in this report:

Google Analytics
Visual Sciences
The Report can