All Posts Tagged ‘Data’

Kevin McCarthy

Miami Dominates LeBron Coverage in Final Days

After yesterday’s movie/company buzz article, Austin had a great idea:  Do the same thing but for LeBron James and the teams vying to sign him.  Without further ado, below are the results.

(expand)

Comparing the amount of articles that mention “LeBron” and the city name, you can see that Miami came on way strong in the final two days.  It seems that the rumor mill was abuzz with LeBron staying in Cleveland or going to Chicago consistently up until recently.  According to the data, it looks like the Knicks or the Nets never stood a chance.

Here is the data of articles that mention “LeBron” and “Ego”.  No doubt this graph will be exponential.

(expand)

It’s interesting to see that the media coverage mentioning the other cities dropped so sharply even before the announcement . Is this a sign that the Heat are better at generating buzz, or simply that an early leak led to a fast-moving rumor mill. Either way, LeBron’s ability to generate coverage is impressive. Since the end of the NBA Finals, we picked up over 4,700 articles that mentioned LeBron somewhere in the title. To put that into perspective, we picked up just over 5,000 articles mentioning ‘oil spill’ in the title.  The numbers aren’t a definitive record of all coverage, but they certainly paint an interesting picture.

Kevin McCarthy

Measuring Media Buzz Over Time: Movies & Companies

Last week, I did some research on the amount of times a movie was mentioned in the press compared to the movie’s box office take.   It figured it’d be cool to measure the amount of daily movie mentions in the days leading up to opening night.

Movie Mentions

Here is some interesting notes about the high peaks:

  • Peak A: Iron Man 2 release date
  • Peak B: Shrek 4 release date
  • Peak C: McDonald’s recalls Shrek 4 glasses because of cadmium used in the paint.
  • Peak D: Toy Story 3 release
  • Peak E: Twilight 3 release

I thought that checking major tech company mentions over Q2 would be interesting as well

Company Mentions

Here are some of the major stories related to these peaks:

  • Peak A:  Google buys Pink, Google rewrites Docs, Italian judge says profit behind Google verdict
  • Peak B: iPhone 4 leaked, Apple 2Q earnings revealed
  • Peak C: German prosecutors investigating Google, Google App Engine
  • Peak D: iPhone 4 release

I think this method is a great way to determine highs and lows in buzz over long periods of time.

Kevin McCarthy

VentureBeat Authors and VentureBeat Articles: The Statistics

Yesterday, I read an informative VentureBeat article by Kim-Mai Cutler via the Yadda.  When I finished laughing-out-loud at the title, I thought a cool exercise would be to dive in and examine VentureBeat’s publication by their authors and their source code.  So without further ado….

Venture Beat Authors

Over the week between Tuesday, June 22nd to Tuesday, June 29th, VentureBeat (VB) published a total of 161 articles.  These articles were written by 25 authors and 12 of those authors contributed more than 1 piece.  The graph below shows the breakdown of author contributions.

(click to enlarge)

Dean Takahashi was the most prolific writer with 43 articles while Camille Ricketts and Kim-Mai Culter each had 22.  Anthony Ha rounded out the top four with 19 submissions.

Venture Beat Articles’ source code

148 of the 161 articles were marked with at least one tag.  Since VB uses a tagging system built on top of WordPress, no VB article has a meta-keyword field in their source code.

The meta-description field appears to be the first sentence of so of the article.  Since the first sentence of any article generally includes pronouns revolving around the article’s topic,  the meta-description field appears to be optimized for SEO.  The article’s title appears in two key places: in the <title> tag and the <h1>.  The <h1> tag is used to ensure that the article at least has chance to have a high page rank.

There you have it.  Is there another publication anyone would like seen broken down in this way?  Let me know in the comments section.

Kevin McCarthy

The Implications of News as a Commodity

Austin’s latest piece described news as a commodity.  Does it matter that you got breaking news from a Snooki tweet instead of from the New York Times?  The answer is no. (A better question: are you seriously following Snooki?)

Let’s assume that the news is a commodity and the atomic unit of the news is an article.  An article therefore has the following two characteristics:

  1. Each article has the same value.
  2. Their values are additive.

So, if there are 15 articles about the new Karate Kid movie and each has a value of X, the value of the Karate Kid story is 15X.

But what is the value of X?

Comparing opening weekend box office gross to the number of article mentions leading up to the film, you could ostensibly calculate the value 1 article mention has on box office reveneue.

For fun, I thought I’d put some of this data together to see the correlation between article mentions and box office numbers (see the following chart here).  Although this test is riddled with inaccuracies* and thus the data holds no weight, I hope you can see the value of this methodology if done correctly (which I plan on doing for a particular industry when the time is right).

Imagine you are an exec at Hewlett Packard in charge of marketing the Tablet and you know the Tablet’s value of X.  To meet your bottom line, you can assume that you’ll need 500 article write-ups in the next 3 weeks to meet your bottom line.  With this knowledge, you could allocate resources for marketing more intelligently (and, more likely, start getting on the horn with every Joe Blogger out there).

There has got to be some industries where something like this would be useful.  Any ideas out there?

* Inaccuracies include but are not limited to:
1) assumption that a high-powered critic’s opinion is the same as a no-namer
2) lack of consideration how long the movie has been in theaters

Kevin McCarthy

Where Should You Get Your SEO Info?

Earlier this month, I explored the Inbound Marketing feed on Pinyadda.  In testament to my loyalty (albeit blind) to social media, I was surprised to find out that SEO was still a major discussion point for new marketers.  As such, I figured I’d resolve my SEO ignorance and take to the SEO topic on Pinyadda.

From May 21st to June 21st, 15% of Pinyadda’s SEO feed revolved around Google, 1.2 % around  Twitter and 1.6% around Facebook.  Could the relatively low percentages of the two social giants have something to do with the fact that social media links are not highly touted by SEO professionals?  What value (if any) does a social media link have in the eyes of search engines like Google and Bing with regards to page placement?

In efforts to learn more about SEO, I next found out which sites were publishing the most SEO content.  With this in mind, the graph below was created.

clink here for enlarged view

These ten sites shown (Media Post Publications, Marketing Vox, Hubspot Blog, Search Engine Land, Brafton Custom News, Online Marketing Blog, Econsultancy, Clickz, SEOmoz Daily SEO Blog and Search Engine Journal) all had more than 3 SEO items published over the last month.  These ten sites also account for about 76% of SEO content on Pinyadda, so might as well follow all of them.

Kevin McCarthy

Comparing media coverage of BP, Facebook, World Cup, and LOST

This week, four big stories have been BP’s oil spill in the Gulf of Mexico, Facebook’s revamped privacy policy, the 2010 World Cup in South Africa and the series finale of LOST.  To compare each story’s online coverage, I went to Pinyadda’s index and began comparing title strings.

For the oil spill, articles with ‘BP’ and ‘oil’ were matched in titles.  For Facebook’s privacy woes, ‘Facebook’ and ‘privacy’ were matched in titles.  For the World Cup, any article with ‘World Cup’ in the title was counted and for LOST, any article with ‘LOST’ in the title and from the Television category was counted.  Below is a graph comparing my findings on each story since April 20th (weekends not included**).  Click on the graph to see an expected view.

Graph 1- Comparing weekday date to relative total of items.

Some interesting notes on the trends in this graph:

  1. As expected, the number of World Cup articles are steadily increasing as the June 12th kick-off slowly approaches.
  2. The number of LOST articles sky-rocketed after the series finale.
  3. On 5/10/2010, the number of BP oil spill articles had a significant jump.  I can’t figure this out, but hopefully someone (cough Phil Arscott cough) can.

**side note:  weekend dates were not included because the volume of published items from publications is much less compared to weekday dates.  another site note:  have you ever noticed how people choose to spend their weekends?  I was walking through the Common last Saturday and saw a otherwise-together 50 year old man hula hooping by himself.)

Data posts from Pinyadda are created using Pinyadda’s index- a database of article links, titles and meta-data from thousands of websites.  The titles and meta-data are analyzed to determine accurately what the article is about.