It’s always fun to look at the site after awhile and review the analytics to see what posts were popular. This blog is really just a collection of random thoughts and things I collect so there isn’t much in the way of a cohesive theme. This randomness is detrimental to the popularity of the blog, but good for my personal well being so whatever. Anyway the top posts are sort of random as well.
This picture is called “Sea of Nudity” by Spencer Tunick. I had corrected an issue the site had with getting a proper Google index page and once Google had indexed it the traffic shot up.
The second most popular post was on the Butcher’s scandal. I titled the post, “The butchers scams Toronto?” and that combination of keywords generated a lot of interest. It is also the top commented story of the year as well with 15 comments. Luckily I turned out to be okay, I got my money back; but from the sounds of it many people did not which is unfortunate.
The third and fourth most popular posts were also about the butchers and were follow up stories.
Over all the blog saw a 5 fold increase in page views, and a reduction in the bounce rate of 10 percent, although the time on page dropped by 26 seconds.
As a Web Analyst the supposed holy grail of data usage and optimization is personalization. Disregarding the whole PI issue for the time being, the notion of a website that uses data to optimize itself to you; that tailors itself to your interests is awesome. Think of all the sites that do it, Netflicks recommends films based on your personal tastes. Google shapes its search to present information that is more relevant to your needs. Facebook changes your news feed, and there are more and more sites that do this all the time. The technologies behind it (predictive modelling, data mining) are interesting but not nearly so as the ethic dilemma’s it produces.
Personalization at first blush may seem like a good idea, but there is an interesting TED talk by Eli Pariser which presents another side to it.
Eli brings up an interesting notion of the “gatekeepers”.
I work for the CBC, which is (for Canadians anyway), a gatekeeper of information. The content that the CBC chooses to air is decided on by a relatively small group of people who follow a journalistic code of ethics. These individuals are making decisions based on data, Comscore ratings, Neilsen ratings, PPM, that sort of thing. They are making the best choices they can given the data they have and the mandate from the Canadian government. These people, who are smart, hard-working, and talented (if they weren’t they wouldn’t be directors and c-levels) also choose things they think Canadians should be exposed to.
The government is at arms length so it cannot dictate to the CBC what news to report or what shows to air but it is still filtered to some extent because it is someone making the decisions about what to report and what to ignore. It is a choice. This lead me to thinking about Chris Berry’s recent post on what you choose to ignore. Personalization it would seem, is as much about what we choose to ignore as it is about what we choose to focus on. The problem is it is effortless on the part of the user.
I think this quote from Mark Zuckerberg, Founder of Facebook is very telling,
“A squirrel dying in front of your house may be more relevant to your interests right now than people dying in Africa.”
How sad is that. This thing, which was created to connect people all around the world, has changed to allow people to bury their heads in the sand and only see those point of views which confirms their current beliefs. Never before has a single ordinary person been able to reach out to millions, but now that is tempered by data and mathematics. And it might not be that the squirrel is more relevant to your interests, it might be that based on your past selections you are not even exposed to this new information about people in Africa.
For me, one of the greatest joys is stumbling upon something at random you might otherwise never discover. Whether it is a point of view on the sex trade you hadn’t considered before, a new movie you might never choose to go to or this incredible photo collection you might never see because the algorithms didn’t select to show them to you.
I think Eli makes a very persuasive argument that personalization, like all technology must be used in moderation lest we all become cocooned in self-confirming information.
I love learning and I just learned something new today I thought I would share. It’s called Benford’s law and it has to do with mathematics. Don’t worry, it’s not hard.
Isn’t that awesome?!!
You’re probably thinking, “nice going dork, that is totally complicated.” You’re right it is. I can’t read that equation I just think it looks cool. But let me put it into plain English and then tell you why it is so cool.
If you took the lengths of all the rivers in Canada and and wrote out the first digit of every length, that is if the river is 356 km long you take “3″ If it is 87 km long you take “8″ and so on these digits would distribute according to universal probability.
Essentially it means that the number 1 will occur more frequently than the number 2. The number 2 more frequently than the number 3. The number 3 will occur more frequently than the number 4 and so on…
What was interesting to us was whether the expensive item, the Big Ben Brownie should be at the top or not. Both Cindy and I noted that the expensive items should be at the bottom, the idea being that by the time you get to the bottom, you’d probably picked an item and that by changing the price, the brownie is differentiated and may help change your mind for the more expensive item.
Hans suggested that while that might sell more brownies, by having it first, and sounding absolutely deciden,t it may encourage people to buy the other “natural” desserts settling instead for the Apple Berry Crisp which is also featured with a picture and highlighted in red.
I was intrigued so I asked how they both would react if the brownie was $10. Both suggested that because the menu has on it “We encourage sharing” the brownie would become a more viable option since the doubling in price would make it seem as though you should get twice as much dessert.
In that case the doubling in price implies a doubling in amount. However if the brownie was $9, it now differentiates itself in price, close to doubling the other desserts, but still becomes the more economical choice if sharing. If it was $12, then it would become super-decadent and should only be chosen when shared.
This wasn’t exactly a scientific study on our part but I think what Chris is getting at is that there are a number of ways to play the menu game, and that what might appear to be little choices in fact can have a significant impact to your bottom line.
Not that it counts for much I choose the Highland Bread Pudding for nostalgic reasons. I guess that would be an unforseen variable.
Chris Berry, a friend and colleague wrote a very interesting post on the analytics of restaurant menus. Chris took a restaurant’s online menu and analyzed the prices of items, numbers of items, average prices and much more.
This is Chris’ first stab at it but the mind leaps at the possibilities. For example, how often should the words “encrusted” or “fresh” appear within the menu? Do those words perform better with different items, or at different price points? Does an items rank in the list effect it’s sales, most definitely. But what about the verbiage and rank? The mind boggles at the possibilities for analysis.
I applaud Chris’ effort and his willingness to share this valuable information to anyone interest to look. In an industry as cutthroat and unforgiving as the restaurant business, everyone could use a little edge over the competition.
