Privacy vs. Anonymity Online

There was an article in Scientific American, The End of Privacy, by Daniel J. Solove. The article was fine. What was more interesting was the response to article by Sheri S. Tepper, who raised the issue of privacy versus anonymity.

Solove’s article is okay, with some good nuggets, although in parts it's typical of these privacy articles that focus on a couple of case examples and present a ‘sky is falling’ aspect to what is happening to the concept of privacy online.

Although there are certainly real online privacy issues (e.g., indentify theft), most of the privacy issues that the Solove article brings up are, honestly, so last generation.

What do I mean by last generation? Here is a typical scenario that some folks present – be careful about what you post on these social networking sites because job recruiters review them. It is true that job recruiters look at these pages. And, if one is a total dirt bag, it’s probably not good to let the world see that.

However, if you engage in these social networking site in any true way, you *are* going to put something stupid or embarrassing there for the online world to see.

And, you know, it’s okay … because, everyone else is going to do it also.

A few embarrassing things on your social network page(s) will be to this generation what smoking dope was to the baby boomers. Probably not the smartest thing to do, but … it happened. The folks that went through it can place it in the situation, environment, and period that it happened. We see this with many both online and offline situations (e.g., if every college student that ever got drunk was denied a job, we would have an extremely large unemployment problem).

Instead, most folks can place these types of occurrences in context and are smart enough to realize that behavior in one context does not necessary mean similar behavior in another. In other words, you can get a little crazy with your buddies and still be a professional employee at work.

In a response to the article, Sheri S. Tepper, raised the issue of privacy versus anonymity. I found the distinction really interesting and may be why I don’t relate much the online privacy concerns that some raise – privacy is an overload term. We may be using the same term but talking about different things.

Sometimes we mean private, as in controlling the information about yourself and your activities (e.g., the Facebook fiasco), and sometime with we mean anonymous, as in keeping one’s identity unknown relative to certain information or actions (e.g., Google aggregrating search data).

Tepper discusses the privacy versus anonymity using the analogy of cities (where one can have privacy and anonymity) and small towns (where one can have privacy but anonymity is much harder).

In his reply, Solove sort of blows Tepper off by glossing over the interesting privacy – anonymity duality, instead focusing on the small town versus city aspect.

I believe privacy versus anonymity is a really interesting and worthwhile distinction that may help clarify some of the online privacy discussion.

You can read Tepper’s response to the End of Privacy article and Solove’s reply.

comScore Report on Metrics Other than Clicks for Online Ads

Read an interesting article from comScore, How Online Advertising Works: Whither The Click?, by Gian M. Fulgoni and Marie Pauline Mörn.

The white paper is an output from the Empirical Generalizations in Advertising Conference for Industry and Academia. The Empirical Generalizations in Advertising Conference has a blog with comments from the conference. (Note: Wouldn’t bother with the blog; rather dry.)

Fulgoni and Mörn report results from a study of 139 online display ad campaigns conducted across several industry verticals.

The aim of the research was to (a) focus on the lack of clicks in online ads (both content and search) and (b) propose that maybe clicks are not the only metrics that one can use.

The report first sets up the scenario that click through for online ads are declining, referencing several sources. The authors then use the data from the 139 ad campaigns to evaluate metrics other than clicks for measuring the success of these advertising campaigns.

Aside from click through, the report presents several positive impacts of these online ad campaigns, including:

• Increased visitations to the advertiser’s Web site (a lift of approximately least 46% over a four week period)
• Increased likelihood of consumers conducting a search using an advertiser’s branded terms (a lift of about 38% over a four week period)
• Increased likelihood of consumers buying the advertised brand online (a lift of approximately 27% in online sales)
• Increased likelihood of consumers buying at the advertiser’s retail store (a lift of approximately 17%)

These seem to fall in line with the dual role of advertising – direct sales or branding.

ComScore uses panel data, but what I really found interesting in this report was that some of these folks allow comScore to access third party programs (e.g., grocery card programs, frequent consumer programs, etc.). One can really do some interesting research with such a data set. It’s a really amazing resource.

There has been some criticism of the comScore data and results in the past, so it’ll be interesting to see the reaction from this report. I don’t know though. With such a rich dataset that comScore has, I just don’t see how the results could be off (assuming the data analysis methods are sound and unbiased).

You can download the How Online Advertising Works: Whither The Click?" White Paper report.

Are You Looking for Statistics about Social Networking Sites (e.g., Facebook, Twitter, LinkedIn, Myspace, etc)?

Here is a nice collection of usage statistics on social networking sites, including Facebook, Twitter, LinkedIn, and Myspace.

The social networking applications are really interesting from an information seeking and from an advertising perspective.

I am presenting research on using social networks and technology (specifically micro-blogging on Twitter) at the CHI 2009 Workshop Social Mediating Technologies, along with my co-authors for the research paper, Abdur Chowdhury, Mimi Zhang, and Kate Sobel.

Neural Networks for Predicting Click Through

One of the key metrics for Web searching is click through, as it is a key component for evaluating both search engines and keyword advertising campaigns.

There has been some interesting descriptive analysis of click through, including some research that I’ve done on click through rates. However, there has been limited research into building predictive models of click through, with work by Brooks on the effect of rankng on click through being one of the few exceptions.

Working on building a predictive model of click through, I was co-author on a paper that reported results of our research to identify factors that significantly affect the click through rate during Web searching.

The underlying goal of the research was to determine more efficient methods to optimize click through rate. That is, if one can predictive, then one can influence.

For the research, we devised a click through metric for measuring customer satisfaction of search engine results using the number of links visited, number of queries a user submits, and rank of clicked links. Since we focused at the session level, we believed that need a more nuanced metric than just links clicked per query.

To achieve this research goal, we use a neural network to detect the significant influence of searching characteristics on future user click through.

Our results showed that high occurrences of query reformulation, lengthy searching duration, longer query length, and the higher ranking of prior clicked links correlated positively with future click through.

Read the entire study of using neural networks to predict future click through rates.

Personalization – Makes Sense(s) But It’s Really Difficult to Do

Designing algorithms to provide personalized information is really hard, actually much harder than one would assume. My dissertation was on personalization agents for information retrieval, and the field is still pretty much at the same stage.

One of the most interesting academic challenges in the personalization of information area is the Netflix Challenge. Netflix offered $1 million to anyone who can improve the accuracy of its DVD recommendation engine by at least 10 percent. As of this posting, the prize remains unclaimed.

Netflix, like Amazon and other major online retailers, has applications that learn customers' preferences by the ratings customers assign to movies when they log into their accounts. These applications then recommend movies (or products) to customers. They do evaluations to gauge the accuracy of the recommender systems. The Netflix Challenge is to improve the recommendation accuracy by at least 10%.

So far, the biggest improvement has been approximately 6% by the Simon Funk team, using a singular value decomposition, which basically results in a clustering approach.

So, there is a long way to go to reach the 10% improvement mark. Naturally, the easier stuff has already been done and the remaining portions are running into the 'hard to define' movies and the changing taste of customers over time.

I also believe part of the issue is that ‘like’ and ‘dislike’ are really irrational processes and using a rational algorithm is always going to have limited application.

Maybe the answer is to use something that is ‘irrational’ like simulated annealing or even more crazy, a semi-random algorithm.

Read the full NY Times article, If You Liked This, You’re Sure to Love That, by Clive Thompson

Google Principles, Tenets, and Motto

Been doing quite a bit of research investigating branding in the search engine market, and the effect of the Google brand and business model are, of course, really intriguing. Generally, the Google brand adds 10 to 15 percent in the searching area.

I’ve been reading various Google documents. They paint an interesting picture of a company, or at least an interesting picture of what a company says about itself.

First, the Google mission, “is to organize the world's information and make it universally accessible and useful” – one can trace this back directly to Bush’s vision of a technology to make the sum total of human knowledge available at a person’s fingertips. The emphasis is on technology as the way to make this information available.

The Google philosophy: Never settle for the best.


The Google Management Tenets (a.k.a., the 10 things Google has found to be true), which are:
1. Focus on the user and all else will follow
2. It's best to do one thing really, really well.
3. Fast is better than slow.
4. Democracy on the web works.
5. You don't need to be at your desk to need an answer.
6. You can make money without doing evil.
7. There's always more information out there.
8. The need for information crosses all borders.
9. You can be serious without a suit.
10. Great just isn't good enough.

Attended a briefing at Google, where the presenter stated that the 3 drivers of change on the Web are access, production, and storage. Nice and succinct. Certainly, the macro-dirvers of change on the Web.

Here are the 7 Principles of Marketing used at Google:
Let others speak for you
Data. Not hype
Results must be trackable
Promote trial
You’re smart. And your time matters
We’re serious. Except when we’re not
Big ideas move us

The Google user experience tenets:
1. Focus on people – their lives, their work, their dreams.
2. Every millisecond counts.
3. Simplicity is powerful.
4. Engage beginners and attract experts.
5. Dare to innovate.
6. Design for the world.
7. Plan for today's and tomorrow's business.
8. Delight the eye without distracting the mind.
9. Be worthy of people's trust.
10. Add a human touch.

And, of course, the unofficial Google motto, Don't be evil.

Site Visit to Google Ad Center and In-sourcing

Did a site visit to the Google Ad Center in Ann Arbor, Michigan, along with a colleague, Clay Voorhees. Ad Center in Ann Arbor, you ask? Larry Page has roots in Ann Arbor.

Tthe visit was really informative concerning the levels of advertising clients and the laying of support available directly from Google.

At the time of my visit, the Center had approximately 250 employees focused on the range from small to medium size enterprises (SMEs) to large firms. Google has done a good job of attempting to recreate the ‘Google feel’ in the Center with snack and coffee room, employee art on the walls, feeds of search terms and traffic from major Internet hubs projected on the wall, and a nice cafeteria.

From a business perspective, it makes sense to locate ad clients (and probably many other knowledge worker functions) outside of Mountain View. With clients throughout the United States, there is no reason for them to be in high cost of living Silicon Valley area. Plus, it has got to be good for other reasons, including risk diversification and public relations.

I often wondered why US corporation are so big on off-shoring customer relationship and some other knowledge worker functions, when there are numerous areas within the US (e.g., MidWest, Up north in New England, Plain States, South, etc. – basically, most areas outside of LA, NYC, DC, Seattle, etc.) where there are great universities, educated English speaking workers, great transportation hubs, etc. where these knowledge work functions could be in-source right here in the US. Has a lot of benefits and the cost in terms of managing has to be less.

Yet, another “The Search Engines Are Watching You” Report

The NY Times had an article concerning the data that they (search engines) record on searchers.

Honestly, I find this whole privacy thing concerning search engines WAY overblown. I am more concerned about the tracking with my EZPASS, ATM machine, credit card, the Internet Service Providers, employer, or cell phone than I am with the tracking that the search engines do.

Personally, identifying a person based on the information search engines collect is really hard to do. Now, over an extended period of time loggin interactions, I agree, it becomes much easier and can be an issue.

In reality, though, search engine companies don’t like to keep data around that can be traced to an individual (due to court actions and other issues), so they either get rid of it after a time or they in some way anonymous or aggregate it.

Distractions like those reported in this newspaper article hind progross on the issue of privacy on the Web (and there is an issue for sure) . A lot of the concerns listed in the above article are just way overstated.

Registrations close 23 January for Google Online Marketing Challenge

Hey, Academic Professors!

The Google Online Marketing Challenge 2009 registration is well underway and will close 23 January 2009. If you have not registered, please do so ASAP.

The Google Online Marketing Challenge (GOMC) is just a wonderful educational experience, where students get real money, work with real clients, and compete again teams from all over the world. Taught a course using the GOMC in 2008; it was just a great.

Here are the Google Online Marketing Challenge 2009 winners. Here are news story about the Penn State GOMC winning team and the James Madison GOMC winning team from 2008.

The Challenge uses Adwords, which has a multi-dimensional paraidgm of searching, marketing, technology, and advertising.

Social Networking Applications’ ‘Friends You May Know’ Feature

Many social networking sites have some type of feature that suggest people that you may know for you to add to your social network. A couple of really interesting aspects of these recommender systems are their accuracy and their inaccuracy. Both are extremely useful.

Certainly, I am often amazed at how accurate many of these recommender systems are. LinkedIn has an especially effective contact recommender application that has sometimes been so freaky accurate in recommending people and the timing of recommendations that it is scary (in a good sort of way).

What I really find useful though is when the recommendations are ‘wrong’; when they recommend someone that I don’t really know. Many times I find myself looking at these possible contacts/friends/whatever your social networking app calls them and thinking “Hey, this is a person that I would really like to meet!”.

These incorrect recommendations have been an impetus for me to reach out to people (in real life – not virtual world) and establish a relationship.

This is one of those unintended consequences that many times occur with technology features. Wonder what other unintended consequences are going to happen?

Presentation of Google Business Model

Here is a presentation from faberNovel, a consulting firm. The presentation presents a consultants’ take on the current Google business model, including strengths and weaknesses of the various Google apps and services and how these apps/services fit together.

All about Google

View SlideShare presentation or Upload your own. (tags: google strategy)

The strength of the presentation lies in the framework that puts everything together in some though provoking ways. Some of the stuff is off the mark (like the PageRank section – that it is no longer the main component is commonly accepted by most tech folks and has been for sometime; however, wouldn’t necessary call it an outdated technology).

Interestingly, the presentation didn’t mention the 3 drivers of change on the Web (access, production, and storage) or the Google management tenets or Google user experience tenets or Google philosophy.

Twitter is Making Me Smarter

One thing that I have notice since using Twitter is that it is making me smarter (or at least more informed).

Via Twitter, I have been introduced to some really insightful comments, blog postings, and articles related to my research interests that I would not have seen if it had not been for Twitter. Folks that I follow are posting links (along with comment!) on a wide array of really interesting and thought provoking articles that I can then leverage for my own research, that cause me to think differently, or inform me of something that I did not know of previously. It is really insightful! Some of these tweets come from folks that I see occasionally in person, and others are from folks that I have never personally met but have read their blogs or published works.

Listservs, blogs, Webpages, email, etc. offer some similar benefits. However, these other forms of communication have conceptual differences that make them less effective for making me smarter.

What makes Twitter so much better than these other mediums? I really believe that it’s the combination of two factors – short message and selectivity of channels. The 140 character limit is a near perfect snippet for information processing. Notice that a standard tweet (or micro-blog) is close to the average length of an English sentence or the average length of a newspaper title and by line? The short message is key to information processing.

Also, since I can select to follow or remove folks, I have a high-level and simple method for filtering incoming information. I can select the channels that I find valuable and remove the ones that are more costly or of little value. The short message and the selectivity of channels are classic factors for an attention economy.

Now, is Twitter making me smarter or just more informed? We can split hairs here, but I believe that it would just be an academic exercise, as the outcome and implication are the same. I have the opportunity to accomplish more of what is important to me. There is the concept that the influx of new thoughts, opinions, and information increases the influence of individuals.

Stretching Your SEM Investment

Attended an on-demand Webinar, Stretching Your SEM Investment, by Omniture, which was really interesting and loaded with some good facts and insights.

For instance, search engine marketing (SEM) is now the 3rd largest advertising channel, behind magazines and TV, and it is ahead of radio. Search or keyword advertising is the largest in the online advertising area (Jupiter Research, Paid Search Forecasts, 2008 - 2013).

However, since the demand for keyword advertising has increased, so has the cost of clicks. Therefore, the cost of keyword advertising is increasing, which puts demands on search engine marketers to hold down costs.

The Webinar was well done, with some overall fact and figures to set the stage at the beginning, then several specifics and case study examples later to maximize keyword returns. The presentation really got down into the weeds in the case studies.

Naturally, there a little marketing for Omniture service and their tool, but it was a generally soft sell.

There was an interesting slide on time management, reporting that typical search engine marketers spend 80% of their time on reporting and getting data, making them ignore other key activities such as exploring strategic options, keyword and copy testing, and landing pages improvements.

The Webinar was worth the time to review.

Search and Branding

Attended a good Webinar today, The Brand Lift of Organic and Sponsored Search, with Gord Hotchkiss, President & CEO, Enquiro, and Mark McMaster, Google. Moderated by Bill Barnes of Enquiro.

The Webinar addressed the key question of: If your company is in a top organic position on the Search Engine Results Page (SERP), why spend money bidding on the same keyword?

The issue is cannibalization of clicks from the organic listings (basically paying for what you would have gotten for free) and the perceived lack of coordination between search engine optimization (SEO) and sponsored search marketing (SEM) teams. The Webinar was focused on marketing leaders understanding how SEO and SEM work together to maximize returns on the SERP. --- a lot of ‘S’ acronyms I know! :-)

Using a case study approach with a focus on consumers early in the purchase process, the key take aways of dual placement (i.e., your result in both the sponsored and non-sponsored listings) were:

• a significant lift in brand affinity
• a significant lift in brand recall
• a significant lift in purchase intent

Enquio has a complete white paper on the brand lift of appearing in both listings.

I have done some branding research, include the effect of search engine brand on evaluation of results and the effect of sponsored and non-sponsored listings on link evaluation.