Yahoo Clues (Beta) is like Google Insights on steroids … but with one semi-serious limitation: data doesn’t span more than the past 30 days.
When you arrive at Yahoo Clues you’ll shortly see an empty screen but Yahoo is quick to auto-populate it with the term fantasy football.
I’m going with the term “christmas songs”, with the timeline set at the default past 30 days (other options are past 7 days and today).
This is where it gets interesting.
1. SEARCHES OVER TIME
By itself the least interesting information on the page.
As with Google Insights, search volume is described on a relative scale from 0 to 100. The relation of the scale is to time; in this case the past 30 days.
To make this kind of information useful you need to have some absolute numbers in mind to compare these relative numbers with. If you have a pretty good idea that the search “blue widgets” is done about 100 thousand times per month then you can compare both the term to be researched and that one and have a relative-to-absolute correlation point.
Here is such a comparison.
Where Google Insights can compare up to 5 terms, Yahoo Clues does a maximum of 2. Given the amount of data they need to put on the page that makes sense.
A comparison like the one above helps you get a feel for your direction. “Christmas songs” does better than "Christmas music”.
2. DEMOGRAPHICS
Includes everything and the kitchen sink 
Click on one of the segments to get extra query details presented in a natural language format.
Demographics displays both gender & age by default but you can zoom in on either.
The value of the market research is again in comparing.
Who are you targeting with this keyword? Who is your web site or this specific landing page for? Who is the product aimed at? The demographics age/gender view allows you to zoom into your market segment very easily.
3. INCOME
Search audience divided by income segments.
Yahoo! Clues calculates this category using anonymous aggregated zip code information from Yahoo! Search matched against per capita income data from the US Census Bureau.
-- Yahoo Clues Help
Again, click on each segment to get some extra details delivered in a natural language format.
If you’ve followed along we’ve already learned that our main audience is between 25-54, female, and with a yearly income of up to $25K.
Comparing by income we don’t see a big difference between christmas songs and christmas music: the few % difference between 25k and 25-70k is not statistically significant enough.
On the other hand, let’s say we would compare christmas songs with christmas concert:
Offsetting search volume against income bracket can help improve your profit margins.
4. LOCATION
US only.
To avoid the most populated states always claiming top spots, Yahoo! Clues factors out relative population differences between states and shows a ranked list of the top 10
-- Yahoo Clues Help
Note that all data is in relation to the search: not in relation to searches to each other.
Compared to christmas music we see very little difference but comparing christmas songs with christmas concert brings up huge differences.
5. SEARCH FLOW
This section provides a list of the most probable previous and next user searches from an analysis of aggregated search patterns across Yahoo! Search. The top item under "Previous Queries" represents the most probable search term people tried before your search term. Similarly, the top item under "Next Queries" represents the most probable search term people tried after your search term.
-- Yahoo Clues Help
One of my favorite views. We’ve seen this data before in Yahoo’s search interface.
Having it laid out in previous queries – current query – next queries is a unique view however and gives great insight in the search funnel. Dpending on how specific a keyword you’re researching you can get a good insight in query repair here.
6. RELATED SEARCHES
Seen that too in Yahoo (and elsewhere). The kitchen sink aspect of the page, I guess. Nice anyway
This section provides a list of the most common related search terms across Yahoo! Search and is not limited to the user search patterns displayed in "Search Flow"
-- Yahoo Clues Help
7. BONUS: POPULAR SEARCHES
By today, past 7 days or past 30 days. (Milli Vanilli? What is this; the 1980’s?!)
YAHOO CLUES & GOOGLE INSIGHTS COMPARED
| Yahoo Clues | Google Insights | |
| Search Volume |  |  |
| Compare Searches | ~ 2 | ~ 5 |
| Time Range | past 30 days max. | 2004 ~ present |
| Demographics | Age, Gender, Income | - |
| Location | US only | world wide |
| Search Flow |  | - |
| Related Searches |  |  |
| Filter | - | web, image, news, product, country, time period, category |
Obviously there is no need to choose and the two tools can live together to complement each other. Together they create a wonderful keyword and marketing research feature set.
Going feature by feature, Yahoo Clues vs. Google Insights, I must say that Yahoo Clues provides more valuable market insight. The demographics and income data is priceless; it can really help to home in on a market. For your own segments you should definitely go back every 30 days to repeat your search(es) and scrape that data into your marketing files.
Google Insights on the other hand is nicer to get a long term view on search trends: whether a search is going up or down over the years vs. just the past 30 days.
I have to pick a winner, you say? OK, if I can only use 1 then I’m going with Yahoo Clues: in the end it has a more inclusive, allround marketing & keyword research feature set than Google Insights would … as the only tool one would be allowed to use.
Which do you prefer: Google Insights or Yahoo Clues?
What a great tool from Yahoo very insightful especially the ability to look more into the demographics. I can only hope that Yahoo! Clues will eventually provide results for the UK
Very insightful tool! Considering that Yahoo! and Bing together only make up about 10% of the search market (compared to Googles 85%+) – I wonder how accurately the Yahoo! sample represents the general populations?
Cheers,
Kris
Good point, Kris. I think it’s a case of the law of big numbers. If they’d only have 10% of North American traffic we’re still talking about 26 million people. That’s a pretty decent sample size.