Click: What Millions Are Doing Online and Why it Matters to You as an e-Learning Professional. – Bill Tancer, General Manager, Hitwise
Keynote – Thursday, March 12, 2009
I found this keynote so interesting I took a lot of notes. I guess I love data too!
Saturday, March 13, 2009 is Pi day.
“To become aware of the possibility of the search is to onto something.” – Walker Percy
His blog – ilovedata.com
Looking through junk emails he saw an email requesting that he write an article. Was actually the editor of Time.com. Requested 3 submissions by the following week. First thing he googled was how to right a magazine column. Then from there the column became more and more popular.
Where does the data come from?
- 25 Million internet users world wide
- 10 Million from the US
- 1 Million websites
- 172 Industry Categories
- Millions of Search terms per month
- Monthly, Weekly, & Daily
All from direct internet providers
One of the first charts that interested him was what was being searched for on one specific term. The interest is that looking at one term’s search and how it is reflective over the full year. Ie: diets are searched for huge on Jan 1 and then end Jan 5. Thanksgiving is the lowest day for the search to be term. This repeats year after year.
Prom Dress Obsession:
Gave an assistant an assignment – take 4 million search term set, using all the retail searches what is the number one thing they are searching the most. Searching to buy. After a few weeks – she said something wrong with data, number one thing was prom dresses. They thought perhaps something had corrupted the data. This was not the case, not effected by bots or anything else. First week of January there is a spike as well on prom dresses. Called their clients, retailers, buyers asking when they did marketing, for dresses and this occurred in March.
Finally at a publisher meeting a gentleman there was the GM of a major magazine and the teen division. He was specifically at the meeting for the prom dress phenomenon. Most revenue also comes through the magazine in March. They decided to try to move the consumer behaviour to begin to start marketing a pre-issue in January.
We can now look at aggregate user behaviour and make very informed choices.
Engagement Ring searches are similar. Spike was the week before Thanksgiving and largest buying time is after Thanksgiving. Consumers are tending to do research about a week before they go to the store.
Seasonality of the data trends can be quite different than we expect. Example: Dancing with the stars seems to be more of a popularity contest than anything else. Looking at who is eliminated and who is searched there is a direct correlation. Charting the season with Stacy Keibler, Drew Lachey and Jerry Rice. So, he predicted that she would win. In fact she came in second. Why the mistake in prediction?
Instead they dug into the intent of the searches regarding Stacey Keibler. The increased searches were not because of the ball room fans but those looking just for “hot pictures, pictures of, etc.” The SKCC – Stacy Keibler Correction Coefficient. This means you need to dig into the reason for the search not just the search term.
Another example looking at how people sell homes, could now predict what would happen in home sales. He started making predictions on CNN. He predicted at one point they would go up, all economists said they would go down 2%. He agreed to come back when the prediction could be validated or not. The numbers came out and they actually had gone down. SKCC he had not thought of was the intent for searches on existing home sales. People used to look at homes for sales and look to sell homes, now instead they were actually looking to see how much their home was valued at in the current market with no intent of selling.
Cognitive Dissonance & the Adult Entertainment Industry
– when you ask someone a question they usually will answer it in the best positive light. But observed behaviour is often different.
Ie: how many people in the audience regularly go to adult entertainment sites? No one answers yes. Lead us to believe that the internet is porn free. LOL.
Someone in the gov’t decided to see if the internet is then 99.95% porn free.
HoweverIf you have a theory and then go and look at the data, you can find or only look at the data that matches your theory and then believe your theory to be true.
When times get tough lipstick sales go up. Things that aren’t very expensive in times that are tough people will be the cheaper luxury items rather than the expensive items (ie a fancy dress). Looking at data about lipstick searches this to timed perfectly with the bank info drops, DOW drops, and this theory seemed to make sense. However, he went back to the data and looking deeper at the searches there were “lipstick on a pig”, “lipstick on a bull”. The peak in searches was actually related to the presidential debates of the time.
Looking at all the online learning facilities of the world he charted how well elearning is doing in this economy. Appears that more and more people are taking a look at their careers and rethinking relooking at updating and improving their education. Good news for us elearning creators.
Income brackets of people that are looking at online learning more than anyone are the upper income groups. Over $50,000 went from 0.7% to 12.9%.
Clickstream
This allows him to see where they went just previous and where they go when they leave a specific site (or search term). Driver of portals is huge to online universities. Social Networking is at about 3% right now, but it is steadily increasing as a drive to online universities.
Search.twitter.com lets you search all of Twitter for specific terms. His big ah-ha was when he searched his own name. Saw that someone twittered and asked if his book was any good? He responded that is was good. LOL
Google Trends for your own sites/terms is a great search. Currently 70+% of all searches are done through Google.
Looked at the pattern of trends looking through elearning platforms. There is a bit of a trend following some of the school calendar, but not exactly. No decline seems to be happening as with other industries currently, but also not quite as dynamic as online universities.
Looking at the age demographics though is a doubling of millennials. This is the only group that is increasing in searching online universities.
Looking at platform searches, they next place they go are technology industry, education, then retailers and the entertainment industry.
Conclusion:
Observed behaviour is one of the key components. Look deeper at what they are truly looking at and why.
eLearning and Institutes are moving in the opposite direction of the economy.
Looking at the data is going to be key to keeping competitive.
Twitter.com/hitwise_US
Click column