Web Sense

Don’t Use a Web Site that Does Not Cookie its Visitors

In his article Internet Says: "Me Want Cookie," WSJ 5/5/2008, L Gordon Crovitz points out that “Web sites do a poor job of explaining how and why this information [cookies] is used…” and that “Web sites and marketers have failed to explain why cookies are harmless…. So people don’t understand how the Web works, so fear they are being spied on and manipulated.” The consequence of which is more and more web users are deleting their cookies, and “Do-Gooders” are lobbying for and politicians are considering legislation to restrict cookies.

Crovitz’s primary justification for cookies is that they make advertising more relevant – visitors get ads for relevant products and services and “in exchange for seeing targeted advertising, we get access to Web sites, usually free.” Relevant advertising and free access are nice benefits from cookies but aren’t their most important benefit. The most important benefit of cookies and the reason I try to avoid Web sites that don’t cookie their visitors is that cookies enable Web sites to distinguish between visitors (not identify who they are just that they are different visitors). This is the foundation of the Web measurement necessary to improve users’ Web experience. Without cookies Web sites don’t have the information they need to improve their site. Sites that don’t cookie their visitors don’t care enough about their visitors to want to give them a better user experience.

Today most major sites use cookies to understand how their visitors use their sites. While companies such as DoubleClick specialize in the use of cookies for marketing purposes, the majority of cookie usage is for site improvement (most Web sites use cookies for both marketing and site improvement). Cookies provide the information necessary to answer questions such as:

Can a site’s visitors find the products they’re looking for? Is the check-out process confusing? How well does online support work? What features are useful and how often are they used? What kind of enhancements would best serve our customers? What types of capacity are needed? How can we improve the customer experience? Did the changes made to the site really improve the user experience?

Comparing the online experience with brick and mortar facilities, the absence of cookies would be like prohibiting companies from looking at and identifying their customers when they came to their facility. Today, banks observe customer behavior and use that information to do things such as have special windows for Merchants who come in with large volumes of cash, to have a single line for retail customers that is serviced by multiple windows, and to position customer service personnel. Retail establishments observe customer behavior to determine where and how to place and displace products, how to design traffic flows, as well as the placement of checkout stations. Grocery stores use membership cards to identify what products customers buy together to better place products making it easier for customers to find what they want. If done properly, it does improve profits, but it also improves the customer experience. No one would even think about prohibiting commercial establishments from observing their customers and using that information to improve their business. In my opinion, Web sites that don’t cookie their visitors don’t care enough about them to want to improve their site and should be avoided.

They only reason to prohibit online merchants from doing so is, as Mr. Crovitz says fear of the unknown – what cookies can and cannot do. Mr. Crovitz’s solution is to have better disclosure statements and uses as an example Walt Mossberg and Kara Swisher’s disclosure statement. Here's what is says:

“Some of the advertisers and Web analytics firms used on this site may place 'tracking cookies' on your computer. We are telling you about them right upfront, and we want you to know how to get rid of these tracking cookies if you like."

Read more »

Tracking cookies are small text files that can tell such companies what you are doing online, even though they usually don't record your name or other personably identifiable information. These cookies are used by these companies to try and match ads to a user's interests. They are used all over the Web, but in most cases, their presence is only disclosed deep inside privacy policies.

We want you to know how to get rid of these tracking cookies if you like. Here are links to pages where you can opt out of the cookies set by our ad-placement contractor and our analytics contractor:

• http://www.doubleclick.com/us/about_doubleclick/privacy/
• http://www.omniture.com/privacy/2o7

"We'd prefer a totally opt-in system, but, as far as we know, the ad industry doesn't have a practical one as of now.”

I enjoy reading Mossberg’s column and respect his judgment. However, in this case I think Mossberg and Swisher’s approach is wrong. It’s great that they tell people upfront about cookies. However, their whole emphasis is on what to do if you don’t like cookies and their explanation of the benefits of cookies is tepid at best. They subtlety pander to the fear of the unknown while pretending that they are objective. They should explain how valuable cookies are and that they enable Web sites to get better. Perhaps they don’t truly appreciate the importance of cookies in Web site improvement. Perhaps they are only aware of cookies as a marketing tool. They should advise their readers to keep their cookies and to avoid sites that don’t cookie their visitors.

I strongly agree with Mr. Corvitz’s position that there is quite a bit of misunderstanding and ignorance about cookies and their value. I propose that those of us who understand cookies and their benefits do something about it to combat the “real risk that someday soon we’ll find the untested hands of regulators in the cookie jar.” Web sites should have an information page explaining the benefits of cookies in access through a “cookie link” that only talks about cookies and their benefits in clear, easily understandable language (no legalize). Measurement companies like Omniture, digital Agencies like Digitas, and associations like the Web Analytics Association should take out ads in major publications explaining the value of cookies (if they don’t they may be out of business). Web professionals of all kinds should write their elected representatives explaining the value of cookies and asking them to refrain from misguided restrictions of their use. Let’s not be intimidated by fear-mongers and the ignorant.

Let’s standup for what we know to be true.

Data Integrators at X Change 2008

Data Integrators, are you ready for X Change 2008?

I'm looking forward to X Change 2008 for two reasons: Food & Conversations.

Conversations.

Last year we were just starting to integrate online and offline data. To be perfectly honest, all I could talk about were the "challenges" involved in data integration. Much has happened since then. We've actually done it. For me, the most interesting integration is that of Online Support and Call Center data.

To get a good picture of what's happening, we use a scatter plot with Online Support on one axis and Call Center usage on the other. We then divide the graph into four quadrants and look at different user segments such as online tenure and customer value. Wow do the patterns differ depending on the segment. Differ is interesting, but they differ in ways that are actionable for both the Call Center and Online Support.

That is valuable because we can use those differences to improve service and reduce costs. For example in many cases, in the first month or so after enrollment, clients are light users of both the Call Center and Online Support. Very quickly, however, there is a group that uses the Call Center over and over with very little use of Online Support. After about a year, the majority of this group is in the low callers and high online user group. By comparing their call types and online behavior, we are developing training and support programs that promise to reduce Call Center volume -- eliminate or at least substantially reduce the large number of calls made by customers with two to twelve months of experience. Because we can actually monitor the customer behavior based on the training and support they have received, we will be able to tell quite accurately what works and what doesn't.

We also look at repeat Call Center usage. Pilot programs have been started in which special training is given to a sample of Call Center Callers and their subsequent behavior monitored both online and offline. Without a doubt some things work (dramatically reduce Call Center volume) and some don't. Some work with certain segments but not others. What's great is that we have facts instead of guesses to support our conclusions based on our actions.

So what has this got to do with X Change 2008?

Well last year there was a group of hardy souls who were just starting data integration projects. In the X Change 2008 sessions, I'd like to compare notes and share insights with these pioneers. Then in the evening (we were all staying at the same hotel and didn't have to drive anywhere - at least if you reserve early enough) over drinks complain, vent, and get the relief that is only possible by sharing your pain with someone who has had the same experience.

So all you data integrators who have suffered the pain of data integration (How many user ids don't match? What do you do when you get records that show a single caller called 40,000 times in a single month?) come to X Change 2008. We'll share insights, swap stories, and get revitalized.

The Food.

I've met the chef at the Ritz and tasted his food. Each Ritz Carlton has a specialty and the San Francisco Ritz's specialty is food. It really is special. I've also seen all the rooms (dinning, conference, and bedrooms). Very, very nice.

Web Analytics -- An Employment Guide -- Where do you excel - Politics or Analytics?

To meet the growing demand for our web consulting services, we are recruiting newly graduated college students as well experienced analysts. As the use of the web becomes a requirement rather than an option, web analytics is becoming an increasingly popular high-paid career. (In the early days of electricity, there were electric companies - companies that used electricity - and non-electric companies. Today every company uses electricity and the distinction between an electric and non-electric company is meaningless. Similarly the distinction between a web company and a non-web company is rapidly disappearing.) In this environment, we've met some very good candidates. Many of whom have a choice between us (while we're the largest independent web analytics consultancy, we're still less than 20 people) and a large agency or general purpose consulting firm. Even though our starting salaries are equal to or greater than the large firm's offer, the candidates often choose the larger organization.

Is this a good decision or not? For those who know me, you know that one of my favorite saying is: "Know yourself and know others: in one hundred battles you will win one hundred times." Well what does that have to do with choosing an employer? Very simply, before you make the decision you need to know what you like, what your strengths are, and what the opportunities and success requirements of your future employer are. If you really like analysis and are good at it, you will do best at Semphonic. On the other hand if analysis is something you don't mind doing, and you have very good political skills you will do best at a large firm (agency or consulting).

With only one basic kind of service, web analytics, Semphonic must provide outstanding, cost effective consulting with every engagement. A large firm provides a wide range of services. Services that can range from web design, IT operations, media buying, to consulting not even associated with the web. Since their relationship is based on a wide range of services, each one does not have to be excellent, just good enough. Consequently, after a certain point, increased analytic capability is irrelevant. When that point is reached, politics becomes the critical success factor in one's career. Being a powerful, skillful member of the compensation committee, being able to smooze with other senior partners as well as client executives is more important than insightful analysis. The super stars of a large firm are those with exception political skills, and they make the really big bucks, more than a successful analyst at Semphonic. However, as with super stardom in any profession there aren't very many super stars and the competition is intense. The high salaries of these super stars requires that non-star employees (everyone else) work long and hard as they have to generate the billable hours that generate the revenue needed to pay the super stars super salaries.

Semphonic on the other hand has very little politics, a flat organization, and low overhead so that a much greater proportion of its revenue can go directly to the analysts doing the work. Its critical success factor is excellence of service. "Good enough" is not good enough. Consequently, increased analytic capability is very relevant. Semphonic gives its analysts more responsibility, more opportunity for different types of analysis, less stress, and more money than any large firm.

So what's the bottom line? If you're good at politics and like the trappings of a big firm and want a chance for super stardom (remember most people don't make star status) go for a big firm. If you like web analytics go for Semphonic. Choosing an employer is an important decision. You spend a large portion of you waking hours working. Make sure it's something that fits who you are.

If Semphonic sounds like a place for you, e-mail career@semphonic.com with your resume and a letter telling us why you like web analytics.

Why Ask Semphonic?; Web Analytics -- The T'ai Chi Solution

Why "Ask Semphonic"? Because the questions are often more valuable than the answers.

In T'ai Chi (a "soft" marshal art that uses the soft to overcome the hard and the mind to overcome physical strength), there is a saying that I think applies very well to web analytics. "Know yourself and know others: in one hundred battles you will win one hundred times." The question is how best to know yourself and others from a web analytics perspective.

Just when you think you've gotten your web analytics under control, things change. Changes can be from any of three major perspectives: business, web site technology, or analytic tools. A marketing channel is no longer as effective as it once was. Your customer base is changing. You have a new product. Competition changes. Then there is always web site technology. You may redo your site in Flash or some other rich media with capabilities you never had before. Clients may be accessing your site through their cell phones, or RSS feeds may all of a sudden take off. Web 2.0 is here and Web 3.0 is on the way. Finally, there is your analytic tool. Pricing, capabilities, or even availability may change. What do you do?

So how do you know your web-analytic-self as well as web-analytic-others so that you can "in one hundred battles win one hundred times" with all this change going on? Start by asking questions. Start with some overview questions such as does your web analytics support your business objectives and how does it do it? Then get into the details of how. What site KPI's do you measure? How? Do you use segmentation? How? Do you track online campaigns? How? Do you track links? How? Do you measure any online processes? How? Do you do longitudinal analysis? How? What kind of web reporting do you do? How do you do it? Keep on going.

Put together a comprehensive list of questions that covers your web site from all three perspectives: business, web site technology, and web analytic tool. As your experience and expertise increase, keep adding to your list of questions and periodically re-ask them. You'll find that the questions are more important than the answers. Why, because the answers are continually changing while the questions don't change very much. If you have a good set of questions, any time there is a change in your environment all you have to do go to your question list and ask. On the other hand, if you just had a list of answers, how would you handle change?

If questions are the answer to how "in one hundred battles you will win one hundred times," what's the best way to take advantage of Ask Semphonic? Prepare your questions. To make sure you're not missing any, categorize them by the three perspectives, business, web site technology, web analytic tools. Select a few you think are the most important and submit them. Then during Ask Semphonic listen to the questions as much as you do to the answers.

When listening to the questions, remember that there are four different types of questions:

1. Questions you know the answers to.
2. Questions you understand but don't know the answers.
3. Questions you know but don't understand.
4. Questions you don't even know.

All four are important, but the last two, questions you don't understand and questions you don't even know about can be the most important. You may not need them today, but write them down. They could very well be the questions that give you the winning edge sometime in the future.

What I like about Ask Semphonic is that it gives you the best of both questions and answers. In short run the answers may be most valuable, but in the long run it's the questions. Semphonic consultants are a great resource for answers. Participants, and at this writing there are over 50, are a fantastic resource for questions.

The first Ask Semphonic takes place on December 11 at 11:00AM PST, 2:00PM EST (check out http://www.semphonic.com/analytics/asksem.asp) and will be about the Omniture acquisition of Visual Sciences. This will not only affect Visual Sciences and Omniture customers, but also has the potential of affecting others in so far as it changes the market dynamics of web analytic tools. The answers given today, may or may not be relevant a year from now. However, I'm sure most of the questions will be relevant. If you're forearmed with the right questions, you will be able to "Know yourself and know others: in one hundred battles you will win one hundred times."

Will Omniture Learn From Google?

What's Google doing that Omniture could benefit from? As Jessica Vascellaro writes in the Wall Street Journal, November 5, 2007, "[Google] hopes to entice others to develop features to exploit the mobile web.... If Google succeeds at rallying developers ... it could open the way for consumers to start doing more easily on their web phones what they can already do on the Web." Why is Google doing this? It's not because they are a philanthropic organization. It's because they want to own the market and they realize that they cannot provide all the features that consumers want as fast as they want them. They can, however, own the platform.

Omniture is in a similar position. With Web 2.0, technology is changing too fast for any one company to provide all the functionality that customers of web analytics software will need. The scope and complexities of rich media are such that Omniture will always be in a catch-up mode if it tries to do everything itself. Take Flash for example, Omniture is finally getting it's act together, but it's taken awhile.

During this catch-up period, what did customers do? Some suffered in silence. However, others wrote their own code. Since Omniture did not provide a robust API, these customers wrote code and placed it below the line line in the Omniture js file that says "Do not modify below this line". It usually worked until a new version of the js file came out. Then the customer had the choice of staying with their modified js file or figuring out how to modify the new js file so that they could continue using the capabilities they had developed. Let me tell you it wasn't always a very prettry picture. Another popular capability to add "below the line" is enhanced link tracking. The list goes on.

There are also features that customers would like, but even going "below the line" won't help. In particular, enhancements to the Excel interface. The existing Omniture Excel interface has a very nice GUI and is fine for new users or users with limited Excel expertise. However, there is a large group of web analysts who are not only experienced Omniture users but also very proficient at Excel. These users would love a robust Excel API. With a robust Excel API, Omniture could take credit for the power of Excel in reporting and analyzing the web data it collects and make lots of analyst very happy.

With Omniture's purchase of Visual Sciences, does it have any incentive to become a platform for web analytics and provide robust APIs so that others can develop Omniture features? In the short run probably not. However, who knows how short the short run will be. As Web 2.0 continues to take over, a smart competitor may take the software platform route and entice developers to develop enhanced features for their platform and eventually take over the market. Look at Apple. It had a better user interface, but didn't open it up and didn't capture the market. If they had, Windows may have been an also ran operating system.

In the meantime, what's a web analyst to do. You don't want to go below the line, but yet you need features that Omniture does not easily provide. Well, we can help. Our Omniture implementation guide (http://www.semphonic.com/analytics/impguides.asp) provides all sorts of tips on how to get the most out of Omniture without going below the line. We also provide an implementation review that provides solutions that don't require going below the line. In the longer run, let's hope that Omniture will learn from Google and start thinking of itself as a web analytics platform and not as a provider of all things to all people -- an impossible task.

Any features you'd like to see in Omniture that a 3rd party developer could add if there was a robust API. Do you have any horror stories related to modifications to the js file?