Google’s power lies in knowledge. The knowledge of what people search for.
Google shares some of this knowledge as it updates their tool – Google
Insights for Search .
With a posteriori examination of historical trends – it suggests a basic
model for predicting the next search trends.
We can now improve the understanding of consumer behavior in variety of
markets and geographical locations.

How does that work?

On many search trends queries, conducted on the Google insight for Search
website, you receive along with the historical data, a 12 month forecast. For good example search for  “iPhone“.

What’s behind Google’s forecasting model?

Google has characterized the predictability of a trends’ series based on its
historical performance. To do so they compared the discrepancy of forecasted trends, applied at some point in the past, to the trends’ actual performance. When the discrepancy between the forecasted trends and the actual trends is smaller than a predefined threshold, Google denotes the trends query as predictable.

Google’s research observations:

– Over half of the most popular Google search queries were found predictable in 12 month ahead forecast, with a mean absolute prediction error of approximately 12% on average. That means that nearly half of the most popular queries are not predictable.

– High predictable categories: Health (74%), Food & Drink (67%) and
Travel (65%).

– Low predictable categories: Entertainment (35%) and Social Networks &
Online Communities (27%).

– To earn more predictable indications – best to search for trends of
aggregated queries (“all categories” option): 88% of the aggregated category
is predictable with a mean absolute prediction error of less than 6% on
average. The larger the aggregation set is, the smaller would be the
variability of the aggregated time series

– There is a clear association between the existence of seasonality
patterns and higher predictability, as well as an association between high
levels of outliers and lower predictability.

For the above summary results of the 10 categories, the correlation between
the Predictability and the Seasonality Ratio is r= 0.80 while the Deviation Ratio has a (negative) correlation of r= -0.94 with the Predictability.

While TrendsSpotting is working on a new research on US Recession Trends
(you are welcome to contact us for more information) – we have
observed such trends, where seasonality seems high and predictability can
be inferred. Take the trend search for “gift” as a good example for the seasonality effect:

In many of the consumer industries we have followed, the recession has
definitely distracted seasonality and made forecasting abilities more difficult: Refer to the search for “flight” for example.

Whats missing in Google Insights for Search?

1. As I have pointed out some  years ago – Google Trends Search gives you
relative search volume only (relative to the total number of searches done
on Google over time). It doesn’t represent absolute search volume numbers.
2. While Google does calculate forecast parameters – how about releasing
the data on seasonality and deviation for the specific search?

A Market Research perspective:

I welcome Google for sharing more trends search capabilities.
I personally met a few of the Google Trends team and highly appreciate their work.

Web search trends are very important for the understanding of consumer behavior. In the many years I have worked as a market researcher, I have always preferred direct methods to analyze consumer behavior.
Compare that to surveys – those can tell you mostly about intentions (far future), about previous actions (depending on good memory skills that respondents  don’t have..), and perceptions (but not in a natural manner where one freely chooses to express her/his feelings). Surveys can indeed provide some indications and estimations for behaviors but they lack the means to reflect consumers’ real behavior.
In the last years, web metrics as search trends, blog citations
tools, and social media monitoring tools have captured my full attention. As a Social Psychologist – I find them to be valuable tools in examining people’s natural behavior. TrendsSpotting (company and blog) is all about making sense of it all.

Tweet This Post Plurk This Post Buzz This Post Delicious Digg This Post Ping This Post Reddit Stumble This Post