Why Jerry Seinfeld Probably Cost Microsoft a Lot More than $10 Million

By Nitesh Dhanjani
November 10, 2008 | Comments: 9

In a previous article, Hacking the Psyche, I presented the security and privacy implications of capturing feelings of individuals using on-line mechanisms for good use as well as abuse and manipulation. Whenever controls around individual privacy are called into question, there is always, on the other side of the coin, a clear business opportunity.

Corporations often use indirect data such as demographic information and sales statistics to measure the health of their brand because the direct data, i.e how the public and their customers actually feel about their brand, is not available for capture. In this article, I want put forth a case study to demonstrate how capturing feelings on the social web can allow companies to measure the reputation of their brand.

Financial Tools for New Times

The second edition of the O'Reilly Money:Tech Conference will be an even deeper dive into the space where Wall Street meets Web 2.0, using technology as a lens to provide a unique view of the most pressing issues facing the industry now, ranging from securitization and trading velocity to risk measurement and the evolution of research. Happening February 4-6, 2009 in New York City.

In September 2008, Microsoft reportedly paid Jerry Seinfeld $10 Million dollars to star in it's recent TV commercial campaign. In this article I want to provide evidence to facilitate the hypothesis that Microsoft, in addition to paying Seinfeld, suffered the additional cost of damage to its brand from the commercials. On a positive note, the I'm a PC commercial that followed seems to have up for the damage.

Here are the TV advertisements:

September 4, 2008: Shoe Circus [starring Jerry Seinfeld and Bill Gates]

September 11, 2008: New Family [starring Jerry Seinfeld and Bill Gates]

September 18, 2008: I'm a PC [not starring Jerry Seinfeld]

Now, let's turn to Twitter to measure the feelings expressed towards these commercials during the month of September 2008. Using the Emotion Dashboard tool I presented in Hacking the Psyche, I was able to visualize how people on Twitter felt about these commercials. Here's a video of the tool in action:

Here is a screen-shot of the result including some annotations:

  1. Most people disliked the first commercial (Red bar indicating overall negative feelings). The most common word used to express feelings towards the first commercial was "WTF" as indicated by the word cloud and the video demonstration.
  2. Feelings on the Microsoft brand started to pick up to a positive state only to be re-plummet once the second commercial was aired (Red bar).
  3. The third commercial, I'm a PC, devoid of Seinfeld, was generally liked and appreciated, helping feelings towards the Microsoft brand return to a positive state (Yellow bar indicating 'happy' feelings).

There you have it: a powerful method to use feelings expressed in social media to measure a corporation's brand and marketing efforts.

Brand reconnaissance is not the only effort that can be leveraged from feelings on the social web. If you are interested in this topic, I invite you to consider my upcoming talk the O'Reilly Money Tech Conference titled Emotion Dashboard: Harvesting Feelings on the Social Web for Powerful Decisioning.

You might also be interested in:


This is an interesting assessment, but must include caution as to the utility of the metrics involved. The metric is limited by at least two facets.

The first is the percentage of Microsoft-affective customers on Twitter. Given that the percentage of Twitter users is something like 1.5% of the total US Population, and a fraction of that for the world population, the metric is representative only of Twitter users, not of effectiveness of the commercial in the full media space.

Second, it is affected by viewers who've actually seen the television commercials. A small percentage (myself included), but nonetheless important.

I applaud the conceptual framework of the question- can brand reconnaissance be leveraged from feelings. However, the source of the data, the demographics represented, and the value of the data stream should always be considered.

Thank you John - I agree with your statements. I willingly concede that the results presented are limited to a non-representative set of participants - the main point of this blog was to present the conceptual framework and to get people thinking about the idea of sentiment analysis using feelings on the social web.

Which is more important? That the (paid) commercials be enjoyable or sell? These two things are not often combined.

I had never bothered to see the commercials but your great article made it worth watching. It is amazing that such supposedly smart people can spend the equivalent of winning 10 lottery tickets on such wasted effort. WTF was exactly my feeling, so it was even more amazing to see how accurately your tool demo reflects twitter et. al. as gauges of the masses' moods.



Interesting post. I lead the social media team for Windows and thought I might just give my two cents, for what it's worth. John Metta's comments are spot on point. While I am the first to encourage marketers to listen to all channels when evaluating the success and reception of a campaign I would caution marketers to make sure the audience they are measuring was the intended audience.

Twitter is a vibrant, engaged, highly technical community that is very vocal but not a good representation of the consumer audience the campaign was designed to reach so I would argue a statement like "Microsoft suffered the additional cost of damage to its brand from the commercials" inaccurate if you aren't measuring the audience the ads were intended for.

My social media team does measure sentiment for Windows and our marketing campaigns on a daily basis as well as campaign sentiment on sites like YouTube. Bill and Jerry was the #1 viewed video in the world the day the 2nd video aired. The long form video was never shown on TV but still received millions of views on YouTube showing a high level of interest which is quite a different data point from the one above. We combine multiple sources of data including YouTube, Facebook and blogs and forums as part of our all up reporting. IMHO looking at one specific channel that was not the entended audience of the ad will not give a full representation of the data.

Hi Marty

Thanks for your comments. I do want to reiterate that the main goal of this blog was to get people thinking about sentiment analysis, and yes, only using Twitter to do this is probably quite limiting (although I have to say, the Twitter sentiments presented here seem right on with people I have met in real life - so maybe there is something to ponder there instead of discounting it outright).

That said, I am intrigued by your comment on using Youtube. It always seems that the actual user comments on Youtube are seldom useful or coherent - do you mainly look for number of views on Youtube?

In my previous article, Hacking the Psyche, I commented on extending this to use multiple sources of social media, which you alluded to. I think we are on the same page.


You're missing the point. Most of the people watched the video(s) on YouTube because they heard the commercials were so bad. You can't infer that everyone who watched came away w/a positive impression...

The YouTube video was shown at our local User Group meeting shortly after it came out - it was uniformly ridiculed by the 100+ folks in attendance.

Classic MS BS positioning...

Nitesh - the "I'm a PC" ads are only deemed better by most folks because the previous ones were so horrible. Compare these ads to the "Mac vs PC" ads and I think you'll see a big difference in viewer emotiveness.

This is definitely an interesting assessment, but any assessment of advertising is fraught with pitfalls.

For one thing, the Seinfeld/Gates ads weren't really about a product, they were meant to build an emotional connection. It's unfortunate that MS led off the campaign with the rather opaque "Shoe Circus" ad, because I'll admit I found it mystifying.

"New Family", though, was funny, and would have been a great lead-off. Since "Shoe Circus" fell so flat, though, the entire campaign had already been discredited and doomed to ridicule.

Has there been any Twitter conversations/statistics on people's reactions to the innovative Australian bank's TV campaign which launched last week, starring Jerry Seinfeld?
(Aussie TV coverage has been at saturation point & the Greater Building Society was virtually unknown across mainstream Australia a fortnight ago - but is now a "household name" as a result of the advertising coup with Jerry.

Was one of You Tubes' most visited videos last wek in Australia apparently!!

Very funny ads with lots of cut- through in my opinion.
(I'm a year 3 University student studying Business & Communications)

I think this Seinfeld campaign is far superior to the much-criticised Microsoft one - you can tell that the Aussie bank guys put a lot of STRATEGIC THOUGHT behind it & have used Seinfeld in the best way possible - they just let Jerry be Jerry!!

Would be interested in your thoughts??


News Topics

Recommended for You

Got a Question?