Insights Gone Wrong

------------------------------------------------------------------------------
If you have not read the previous post "Do online trends = offline trends" Read it first.
------------------------------------------------------------------------------
This is an example of how online trend spotting can go wrong. Unfortunately this example involves me and my daily analysis of Google Insights.

Because one of our biggest clients is a large salon, I decided to see if there was a search trend for the term "Shampoo" and "Conditioner." This would help our client order large amounts of shampoo and conditioner at optimal times, and also help MarketR optimize the client's website.

Sure enough there was a trend. See image below.
The red line reflects "Conditioner" and the blue line reflects "Shampoo." According to Google Insights, people in Canada search for "Conditioner" the most during the months of May, June, July. The search term peaks in June.

This was an exciting finding!

Luckily I have learned the hard way not to jump to conclusions. I called the client up to to notify them of my potential finding. I asked why they thought this might be true? A few ideas came up, people travel and go swimming and they don't want their hair to be damaged, or that the heat can cause dryness so they need a good moisturizing conditioner.

These ideas sounded pretty good to me, but I wanted to be completely sure that this was a trend. I decided to look at news stories during the time of May, June, July. (The news can spur trends and also reflects trends.) Sure enough I found my answer. I only needed to see one news headline for the answer to hit me. "Air Conditioner sales are up."

Google Insights tracked all search terms including the word "Conditioner." Such as dog hair conditioner, pond conditioner and air conditioner. The results were completely skewed, and now it made complete sense why the search term "Conditioner," increases in the summer.

It wasn't all bad news though. We were able to refine our search terms and found an interesting trend for Shampoos and Conditioners. I advised our client to stock up on a certain type of shampoo and so far it has made them a lot of money.

Try This. Remember online insights, and all insights for that matter, have to be investigated. See if you can figure this one out. An application on Facebook called Lexicon allows you to analyze keywords Facebook users write about on their wall. The application is not nearly as sophisticated as Google Insights, but we believe once refined it will be incredibly valuable. When web gurus analyzed the keyword "sex" they saw a spike. See below
In May there was a huge spike in the keyword "Sex" on peoples walls. Were people having more sex during May? What was the cause of this spike? This left some of the smartest web analysts perplexed. Can you figure it out?