Super Crunching the Ontario Election

In the Ontario Provincial Election tonight (in Canada), TV network Global announced the minute the poles closed that the Liberal Party had won a majority government with at least 54 seats in the assembly.

No ballots had been counted, no ridings had reported.

Immediately after this prediction, Global went on to interview high ranking members of the Liberal party to congratulate them on the win and the majority.

How did they do it? Global conducted an exit poll across Ontario with approximately 6500 respondents. They asked respondents to disclose which party they had voted for and some of the factors as to why. The analysts were confident that their poll was accurate 19 times out of 20, with an error margin of 1.2%.

It was strange to say at the least, but it highlights a trend that is going to become more common - using statistical regression models (tracking the history of an event) to help predict the future.


This concept is wonderfully described Ian Ayres' new novel "Super Crunchers." It's scary to think that we can predict the future based on previous trends / polls.

For advertisers, having terabytes of consumer data to crunch provides us the ability to predict future actions - developing highly effective executions. Think it's far fetched?

Think about all of the information that you currently give to companies. Credit Companies, Teleco's, Amazon and Facebook - all storing your purchase history, likes and dislikes, goals and dreams. This information can be used for advertisers to effectively target the exact group they want.

If anything, it's interesting to see Super Crunching hit the political mainstream.

My Favorite Reads of 2008 [Another List]

In 2008, I made a resolution to read more. A lot more. 12 months, 30 novels and about 1,000 articles later, I feel more up to speed and interested (in everything) than I have before. Here is a short list of my favorites (warning: some are pretty random):

10. Click by Bill Tancer - interested in the world of Search? A must read (and a quick one) about how to use search data to influence your future decisions. Some cool stats and a great guide to understanding the digital world.

9. Whose Your City by Richard Florida - what's the most important decision you will ever have to make in your life? According to Florida, it's where you live. Want to understand the future of cities and how they directly impact your life? Check it out.

8. Tribes by Seth Godin - I like Godin. Like his blog and his stuff and this was a decent read (and if you have read anything before you can finish it in about 2 hours).

7. Hey Whipple, Squeeze This by Luke Sullivan - want to read the book that is mandatory for CP+B creatives? If you want to understand the creative mind, process and struggle (and laugh while doing it), pick it up.

6. The Looming Tower by Lawrence Wright - Not ad related but important. Wright interviews thousands of people to understand the history of Al Qaeda and puts together a game-changing story about how 9/11 came to be. Mind blowing and the winner of the Pulitzer Prize.

5. Groundswell by Charlene Li and Josh Bernoff - I had the pleasure of hearing Bernoff speak in Toronto a few months ago and although a lot of the examples in Groundswell are very well known, the book does a good job at classifying digital behaviour in order to help ground strategy. Plus, almost every client has read it (well...any client that thinks digital is important)

4. Super Crunchers by Ian Ayres - I truly believe that data will drive the majority of our decisions in the years to come. This novel is about using data to be more accurate in your predictions and decisions (and has some mind blowing examples).

3. The Futurist by James P. Othemer - I don't read a lot of fiction but when I was given this as a gift (thanks Crowe) and found out that Othemer is a copywriter I thought I'd give it a try. 3 days of obsessive reading later and I was finished (and super satisfied).

2. Hot, Flat and Crowded by Thomas Friedman - the future of the world (if we have one) is green. This is the definitive plan on how we get there.

1. Wired - I'm a huge fan of Wired (and just got a subscription...finally). I think I'm most excited when I find the issue on my desk or grab one from my local store. If you love digital, if you are passionate about the space and want to learn more, read this magazine every month. Cover to cover.



Read every article, even if it seems - at first - like it's a topic that doesn't interest you (who knew that I would love reading about genetic sequences?). I'm a firm believer that if you want to be good at digital, you should learn from the best. And the writers, staff and stories found in this publication are all at the top of their game.

What about you? What were your favorite reads of 2008?

eHarmony Ideas [Seriously]

Don't ask me why, but I've been thinking a bit about eHarmony lately. It could be the relentless TV campaign of people who have finally found one another or a shout out to Super Crunchers but the system that eHarmony uses seems like it could have benefits across multiple industries.

Consider it for a moment. The founder of the site, Dr. Neil Warren, studied couples who had been married for years and were happy with their partner. He developed a system of key relationship attributes (individuals backgrounds, interests, location, family, religion, personality traits, etc) and created a system to link people based on deep, similar traits. His method might not be proven, but based on the stats from eHarmony, he's matched a ton of couples who are happy and now do a ton of testimonials for the site.

For all accounts, the idea that successful relationships can be mapped against an algorithm is an interesting one. That we can just figure out what the formula for a successful relationship is and apply it to two people who have never met. Two minutes and we'll tell you the rest of your life.

OK. Maybe their are some flaws. I'm sure the process doesn't work every time and that people on eHarmony don't always meet their soul mate after they've paid to contact their "matches". But still - a lot of people believe that this works. And for some it has.

My question is, why don't algorithms like eHarmony's exist for other categories?

Take job finding sites like Monster, Career Builder or TalentEgg (run by a friend of mine). Right now, the big ones allow you to post your resumes and search a deep data base of all the jobs that have been posted by corporations. Once you find a few that you're interested in, you just apply right there and hope for an interview.

Couldn't we do a bit better than that? When Seth Godin is writing about how important your job is, someone has to listen, right? Why don't we look at people who are happy with their companies and find out the key attributes that make for a successful employee-company relationship? That way, when I search for a job I'm interested in on one of these sites, I not only know the basics, I know whether I'll be a good fit or not - for the long term.

What about if we translated this into real life and created bars that guaranteed a similar demographic of person? For example, what if you knew that when you went out you were going to a place where everyone shared the same interests, had the same tastes and was compatible? It would make talking to strangers a hell of a lot easier.

That being said, it might get a bit bland as well.

All in all, the idea of an system to guarantee at match (whether it be relationship, job, charity, etc) by comparing key insights and information is interesting. Where else do you think something like this could work?

Google Canada - What Do We Search?

A collegue of mine (who is a Doctor...not like a real one but a P.H.D) sent me this article about Google Searches in Canada. For the past few years, Google has released its Zeitgeist report (German for "Spirit of the Times") to help show current web trends and provide insights into what people are searching.

This is the first time they've included Canada and it's interesting to see what Canadians have been searching for the past year. The top search? "Facebook". At number 2 and 3 are "YouTube" and "Lyrics" respectively.

The article also provide a list of the most searched political parties in Canada (Green Party, Liberal, Conservative, NDP then Bloc) and the top celeberities (Britney Spears, Jessica Alba and Heath Ledger coming in the top spots).

Search is a critical component to developing effective communications plans and I think many marketers simply think of it as an afterthought. ("Great, now that we've got the TV spots nailed, let's buy some search words!") Some of the best books I've read this year (Click and Super Crunchers) are solely about using data (and search data) to create campaigns that resonate and stick.

Tools like Google Trends for Search are immensly helpful in understanding seasonal trends in search as well as understading how your target searching for something. There are fantastic, free tools out there that people just don't use when planning campaigns or thinking about their target.