Digital Root

The Mind-Language-Physiology Link

When I was a boy, I believed that people with blue eyes could perceive the color blue more acutely than people with brown eyes.

This research doesn't exactly prove that, but it does suggest there's a mind-language-perception link going on.

It was clearly the language system interfering with a task that was presumably purely visual: distinguishing between different colors. Categories in our thinking may go much deeper than we think.

While not conclusive, this research indicates that having a greater definition of a given color enables perception to a more granular degree.  Its conceivable that this language-physiology link permeates our ability to think and perceive well beyond colors.

QR Codes and That Guy Who Uses Them

 

Forrester released a study on QR/Barcode mobile usage.  QR/2D barcode scanning is up to 5% (from 1% a year ago) for smart-phone users in the US. Earlier reports stated a total of 14MM "scanned a QR code ever", but its unknown if this was a superset of smartphone or general mobile users.

Either way, it’s still a small market – 6MM people in the US, primarily in the 25-34 "affluent" demo.  Something to consider when adding QR codes (especially if its intent is solely to redirect to a URL) to print/web materials.  In fact, if you're setting up QR codes to redirect to a simply root URL, you're wasting your time.

Because these reports rarely quantify their numbers, I pulled the latest survey data from Q3'11 US Census, CITA, ITU and Nielsen data.

 

Mobile QR Code Usage in the US

Approx Mobile accounts:                 285.0MM

Approx Smart Phones:                     114.5MM

Approx QR/Barcode Users:                5.7MM

 

Over the next 24 months I'm expecting to see a shift from QR code usage to object/image recognition as redirect markers.  Most smartphones can train and recognize an image or object in the real world as a redirect or augmentive marker, rendering the insano checkerboard that is a QR code less novel.

Agile Is Dead

Agile methods for planning - from software to market research - have evolved to a buzz/hype laden, heavy, and bureaucratic burden. At least in many shops that have too many people bearing spreadsheets and too few creating unit tests.

In practice I've restricted Agile/RUP/XP/etc to the minimum applicable elements necessary to get the job done consistently and humanely. Typically this means a basic sizing/risk estimate for discrete stories (oops my 1990's XP training is showing), pushing the high risk elements into early prototyping, burn/build tracking and iterative assignments. 

Generally it gives our teams what they need - projects that have the features and functions it needs, clear visibility, ability to change scope and assurance of date/budget delivery.

Applied to planning (advertising / product or research) a recent book eschews the mythos and hype around agile and boils it down to simplified experimentation. I read a good review that simplified it even further:

It's about doing the minimum amount of work needed to get to, or inspire, an idea.  It's about cutting down the waste - things like deliberating adjective soup and brand vegetables for 3 months.

It's about generating hypotheses that can be tested, not sitting in an ivory tower with a damp towel on your head waiting for the answer to appear like divine inspiration.

It's about making stuff that can be tested in the real world, not running weeks of focus groups to hear people talk about how they think they think or might behave.

It's about making as few charts as possible to explain a strategy or idea - get to the work

It's about understanding that strategy has to evolve and morph over time (whilst being aware there's a danger that this could become an excuse for superficiality).

It's a bias to doing over thinking.  

It's about doing stuff to learn stuff.

It's more like experimentation than planning as we know it.

 

 

Topography of Infographics

A nice topography of infograhpics.

Credits: http://www.flickr.com/people/smoy/

Gaming YouTube

Youtube is reported to be the second largest search engine in existence, placing well above Yahoo! and Bing in terms of traffic. However, unlike Google, Yahoo! or Bing, there's little you can do to drive your videos to the top placement; short of being wildly popular.  And if that's the case, you're probably not reading this blog post.  YouTube while having the second largest search volume, they have less than .1% of web pages indexed by Google, providing a dramatically less competitive environment to get placed.

Based upon various online discussions, I concluded that searches within YouTube were subject to unintended but effective optimization techniques - just like normal search engines. To prove this, I setup a test with a well known brand (name withheld) that had less than 50 videos and less than 1500 lifetime views.

Cleaning up their tags, links and descriptive text was the first step, to make sure they were consistent, but the break-through was enabling Closed Captioning (as in for the deaf closed captioning) for every one of their videos. Essentially, we created our own closed caption script and uploaded the videos with them vs. using YouTube's automatic option.

The results were impressive – a 508% sustained improvement search results.

It appears that Youtube/Google uses the text from the closed-caption to index your video. When a user searches for a topic the words placed in your closed captioning text will be matched and placed higher in the search results.  In cases where your keywords are clouded by unrelated videos and hash tags, closed captioning may enable you to consistently place above other videos.

Optimizing search for images, videos and other rich content is untapped territory with a big payoff for relatively little effort. The simple technique of adding closed caption text to every video out performed paid sponsorships, link building and generating content.

The Phone Game

This is what happens when the simplest of tasks is repeated from one person to another. Imagine how highly fragmented and complex ideas become when there isn't a common language or protocol for describing the concept. 

Its also why I have the "3-email rule".  If something takes more than 3 emails to convey, its probably too complex or political and should be discussed in person.

Note: The exception to this case is when a team has a highly concise protocol for conveying information and a means to uniformly test for parity.

 

A Sequence of Lines Traced by Five Hundred Individuals from clement valla on Vimeo.

Showing Numbers In Context

The Boston Globe demonstrates the value of showing numbers in context when building charts.  Often times, charts will casually depict numbers on different scales (e.g., Millions and Billions side by side) which obscures interpretation.

When creating visualizations, try to keep values on a relative scale whenever possible. If not, try to create a common measurement for context (as Shaq is below).  And when all else fails, try using a natural log version to illustrate the relative shape of the data.

 

When Good Design Happens To Bad Data

This infographic has been getting all kinds of love on Mashable. While it was created by a professionals I know and respect, it frosts me when wishes are presented as fact.

A portion of the underlying data originated from the ecumenical Addthis  of Clearspring Ad-network fame, but the comparative set is fit into models that are dubious at best.  Taking unlike metrics (e.g., Facebook's Monthly Active User, vs number of Gmail accounts) and using them for corollary or comparative models is risky if not deceptive form.

In the case of this infographic, the objective is to transfer the merits of well-known data (addthis tracking data) to unknown (pretty much everything else) and obscure the equivocation with slick design. 

It's doing exactly what most people expect advertising to do; Lie.

The Joy of Stats

For all of you who wished there was a movie version of your stats class.

Hans Rosling says there’s nothing boring about stats, and then goes on to prove it. Only with statistics can we make sense of the world and harness the data deluge to serve us rather than drown in its confusion.

A one-hour long documentary produced by Wingspan Productions and broadcast by BBC, 2010.

Enjoy!

But Her Chi-Square Is Really Cute

Repost from Andrew Gelman (Professor of Statistics @ Columbia University) :

Brian Mulford writes:

I [Mulford] ran across this blog post and found myself questioning the relevance of the test used.

I'd think Chi-Square would be inappropriate for trying to measure significance of choice in the manner presented here; irrespective of the cute hamster. Since this is a common test for marketers and website developers - I'd be interested in which techniques you might suggest?

For tests of this nature, I typically measure a variety of variables (image placement, size, type, page speed, "page feel" as expressed in a factor, etc) and use LOGIT, Cluster and possibly a simple Bayesian model to determine which variables were most significant (chosen). Pearson Chi-squared may be used to express relationships between variables and outcome but I've typically not used it to simply judge a 0/1 choice as statistically significant or not.

My reply:

I like the decision-theoretic way that the blogger (Jason Cohen, according to the webpage) starts:

If you wait too long between tests, you're wasting time. If you don't wait long enough for statistically conclusive results, you might think a variant is better and use that false assumption to create a new variant, and so forth, all on a wild goose chase! That's not just a waste of time, it also prevents you from doing the correct thing, which is to come up with completely new text to test against.

But I agree with Mulford that chi-square is not the way to go. I'd prefer a direct inference on the difference in proportions. Take that inference--the point estimate and its uncertainty, estimated using the usual (y+1)/(n+2) formulas--and then carry that uncertainty into your decision making. Balance costs and benefits, and all that.

Moving forward, you're probably making lots and lots of this sort of comparison, so put it into a hierarchical model and you'll get inferences that are more reasonable and more precise.

But . . . who knows? Maybe Cohen's advice is a net plus. Ignoring the chi-square stuff, the key message I take away from the above-linked blog is that, with small samples, randomness can be huge. And that's an important lesson--really, one of the key concepts in statistics. Don't overreact to small samples. If the silly old chi-square test is your way of coming to this conclusion, that's not so bad.