Posts filed under 'Web Analytics'
There are reports
on how many videos Hulu streams but how come they don’t tell us how many views individual shows are getting? Usually TV shows are quick to point out
which ones are getting the highest ratings on TVs even though the way ratings are derived
is anything but exact:
Nielsen is making an assumption using a sampling statistic based on 5,000 homes what the approximately 113 million U.S. television-viewing homes are watching.
Yet online, exact amounts of viewership is much more possible. Hulu knows exactly (almost exactly depending on the constraints of their web analytic providers) how many people are watching which shows, how many people drop out and watch shows only half way and they also know the mix of shows people are watching. For example, they would know that a high percentage of people who watch the Simpsons also watch Family Guy, etc.. Sure, Hulu has their “most popular videos” category but they don’t show how many views to substantiate their claim of what is most popular.
You would think they would advertise things like, “Come see the most viewed show on Hulu!” but they don’t, why not? They are hiding something. I bet there is some conflicting data between what the Nielson ratings show and what online shows and they don’t want their advertisers to know about it. And their “most popular videos” category is probably anything but the most popular. I think they cherry pick which clips they want people to watch more of based on which shows demand the highest costing CPMs.
What if Arrested Development is the most popular? But since that show is not airing on TV they don’t want people to like it more, they want people to like The Office more so they can get those people to tune in on Thursdays to sell more advertising. Is their new show Community, which is on the top row for most popular, among the most viewed? Doubtful, I bet they want more people to be exposed to the show since they have a lot riding on it becoming a success. Does Hulu take stocking fees like in supermarkets where networks pay them to put their show on the homepage? Maybe.
For sure they have some good reasons why they don’t reveal which shows get the most views.
October 30th, 2009
I finished reading the book Super Crunchers
by Ian Ayres. I thought it was good. I liked his explanation of randomized a/b or multi variant testing done online and off.
He explains that randomly dividing prospects into two groups and seeing which approach has the highest rate is one of the most powerful super crunching techniques ever devised.
When you rely on historical data, it is much harder to tease out causation. The sample size is key. If we get a large enough sample, we can be pretty sure that the group coming up heads will be statistically identical to the group coming up tails. If we then intervene to treat the heads differently, we can measure the pure effect of the intervention…after randomization makes the two groups identical on every other dimension, we can be confident that any change in the two groups outcome was caused by their different treatment.
Of course, randomization doesn’t mean that those who were treated differently are exactly the same as those who were not treated differently. If we looked at the heights of people in one group, we would see a bell curve of heights. The point is that we would see the same bell curve of heights for those for those in the other group. Since the distribution of both groups becomes increasingly identical as the sample size increases, then we can attribute any differences in the average group response to the difference in treatment.
In lab experiments, researches create data by carefully controlling for everything to create matched pairs that are identical except for the thing being tested. Outside of the lab, it’s sometimes simply impossible to create pairs that are the same on all peripheral dimensions. Randomization is how businesses can create data without creating perfectly matched distributions.
The power behind randomized testing is undeniable. So should we just have computers make all our decisions for us? With that question in mind is were he goes throughout the majority of the book.
Randomized trials require firms to hypothesize in advance before the test starts. Historical data lets the researcher sit back and decide what to test after the fact. Randomizers need to take more initiative than people who run after the fact regressions.
The most important thing that is left to humans is to use our minds and our intuition to guess at what veriables should and should not be included in the statistical analyisis. The regressions can test whether there is a casual effect and estimate the size of the causal impact, but somebody (some body, some human) needs to specify the test itself.
So then the question becomes what do we test, and after we test the question becomes, what are the results telling us?
June 12th, 2009
You cannot manage what you cannot measure…And what gets measured gets done.” Bill Hewlett, co-founder of Hewlett Packard
This quote entails the essence of data driven Internet marketing. Do you know where your customers come from, how much the average customer spends or how often your customers come back? Powerful decisions can be made from looking at the answers to these few questions alone. You could target your marketing efforts to the places where most of your customers come from. You could try up-selling techniques to improve your average profit per sale. You could give your most loyal customers tools to spread your message via word of mouth to their friends.
Wal-Mart keeps track of the number of items per hour each of its checkout clerks scans at every cash register, at every store, for every shift as a means of measuring their productivity. These obsessive data gathering habits are at the heart of Wal-Mart’s strategy. A small business cannot afford to ignore the importance of marketing accountability and measuring success.
March 16th, 2009
An excerpt from Web Analytics:An Hour A Day
by Avinash Kaushik:
Imagine walking into and out of a supermarket. If you did not purchase anything, the supermarket managers probably didn’t even know you were there. If you purchased something, the supermarket knows something was sold but that’s about it.
Visiting a website is a radically different proposition if you look from the lens of data collection. During the visit to a website, you leave behind a significant amount of data, weather you buy something or not.
The website knows every “aisle” you walked down, everything you touched, how long you stayed reading each “label,” everything you put into your cart and then discarded, and lots lots more. If you do end up buying, the site manager knows where you live, where you came to the website from, which promotion you are responding to, how many times you have bought before, and so on. If you simply visited and left the website, it still knows everything you did and in the exact order did it.
With this kind of information, imagine the kind of improvements you could make over time to help your website grow.
March 16th, 2009
It’s easy to be overwhelmed by the data the Google Analytics provides. Here’s a look at three units to measure – hits, pages views and visitors.
A “hit” DOES NOT actually refer to the number of times a user visits and/or clicks on a Web page. A “hit” refers to the user request for a Web Page “hitting” the web site’s server. Thus, you could have multiple “hits” to the server but only one view of the Web page. For example, if you have a page with 10 pictures, then a request to a server to view that page generates 11 hits (10 for the pictures, and one for the html file). A page view can contain hundreds of hits.
A page view is each time a visitor views a webpage on your site, irrespective of how many hits are generated.
A visitor counted only once in a specific time frame. So if someone visits the site today and tomorrow, they’re are counted as 1 unique visitor and 2 page views.
Google Analytics Blog does a good job of describing how to measure visitors accurately on Google Analytics.
The ultimate goal is to measure quality. One way to measure the quality of a site is a low bounce-rate or the visitors who move onto another site immediately after visiting your site. What does a high bounce rate tell you? Avinash Kaushik defines it as, “I came, I puked, I left.” So in other words a high bounce rate isn’t good.
March 12th, 2009
The key to a successful PPC campaign is determining the keywords/phrases that your target audience will search for to find you.
The first step is creating a “keyword universe”
- Think about what words your customers use when referring to your product/service.
- Use a keyword tool to get a list using those initial keyword ideas. Google’s keyword tool and the SEObook keyword tool work great.
- You can also have Google go through your site and come up with more ideas.
- With that list expand it with common misspellings, plurals and abbreviations.
Now, all of these different keywords can be used by customers at different stages of their buying cycle. With some analysis you can understand to a degree what the customer’s motivation may be.
Learning Stage: the customer is gathering information. They use broad keywords like TV.
Shopping Stage: the customer is comparing products, brands and features.They use a little bit more refined keywords like Plasma TV or High Definition TV.
Buying Stage: the customer is ready to buy. They will use exact keywords of model numbers like Sony BRAVIA 46″ 1080p HDTV.
Is this strategy fool proof? No. But utilizing your web analytics to measure the success of certain keywords will allow you to see those keywords that are catching people too early in the buying process. If a lot of people are bouncing quickly, they may be too early in the buying process.
March 12th, 2009