Trend data can be useful, and there is no cooler tool on the market than Google Insights for Search. Understanding the Data It's important to understand what the data does and does not show. The graphs show the 'relative popularity' of searches, not the overall number of searches. In other words the graphs represent the fraction of overall searches that the given query represented over a period of time. A graph showing a flat line, neither rising nor falling, does not indicate that there is no fluctuation or trend in the overall volume of search. There could be more and more search on your brand every year, but the trend line could be flat or down if the overall search volume has kept pace or exceeded the pace of growth on your brand. Presenting the data as relative to the overall search volume is useful in the sense that it smooths out day of week and seasonal fluctuations in search volume. Clearly, my brand gets more search volume during my peak season, but the question this answers is am I capturing a larger share of search volume as well. But, given that we can't jump to conclusions about the exact volume of search based on this data, are these data useful at all? Absolutely, if you can tolerate some squishiness. 3 Not So Helpful Uses of Search Insight Data In general, the most insight seems to come from looking at comparative trends across several keywords.
- Who's hot and who's not If you're handicapping the Grammy Awards, this could be useful. However, it's important to note that we don't know exactly why people are searching for a particular person. Are they popular, did they release a movie, win an award...or do something really nutty?
- Investment Decisions Which stock should you buy? Sidebar: Isn't it a little weird that people are searching for 'Google'...when they're ON Google? Okay, okay, it's mostly broad matching folks searching for Google maps, Google Chrome, etc, but still.
- Election handicapping? The geo-targeting capabilities present some interesting possibilities for election data: One might look at this data and say: "Wow, we could have seen Herman Cain's victory in the Florida straw poll coming"...except that the Florida straw poll was taken 9/21...before the spike in interest in Herman Cain.
- Understanding YoY Performance Changes In the introductory disclaimer I pointed out that we can't really tell whether the absolute volume is up or down when we see a trend line going up or down, all we know is the keyword(s) are increasing or decreasing as a percentage of the total. However, it isn't unreasonable to think that absolute volume of search is increasing year over year -- indeed we can probably find benchmark data that suggests how much -- so it's reasonably safe to assume that a flat line from one year through the next means that the absolute volume of search for that keyword(s) is increasing enough to keep pace with the overall search volume. An upward trend probably indicates substantially growth. A downward trend could mean anything depending on how steep it is. So if we look at a graph like this: Pretty clearly something good has happened in 2011. The trend has been generally downward as a percentage of overall search if not in absolute volume, but in 2011 that trend reversed. As we look at explanations for improved search results, good management is one explanation, but increased volume of search in the category might be the dominant explanation.
- Competitive Insights How are we doing compared to our competition?
- Offline Media Tests If you're a national advertiser, you could do a big media blitz in one or two cities, but not in others. Using the geo-targeting options you could see if the trends look different in cities where the blitz hit from those where it did not. There's always talk in our industry about measuring the impact of online media on offline results, but clearly it's a two way street.
- Lift Assessment Similar to the above, how much do email blasts increase search on your brand? Can you see it in the data? How about a Facebook advertising blitz? It may be impossible to separate the impacts of these online marketing pushes from day-of-week effects or other offline marketing efforts, particularly if the "blasts" are small relative to the size of the business. And, clearly, this data doesn't give 'the answer' the way a well designed test with control groups does, but it's free, and it may give you a scent of the impact without incurring the cost of a test.
Join the Discussion