Market Samurai inaccuracy?

So when I'm doing KW research with Market Samurai, I've noticed recently that results are WAY off on monthly searches (with exact selected). I have a KW that according to MS gets 100 searches per day EXACT. When I plug the same KW into G keyword tool, it says it only gets 390 exact searches per month. What gives? I thought MS got their info from the big G anyway?

Market Samurai is far more accurate than Google!

I want MS to be right, but in reality I've been using it since it came out, and haven't had the best results. It's accuracy to me is questionable at best for KW research. But since MNF is on the DNS list and I don't wanna cough up $97 bucks for it, I'll power through.

What you may want to consider is Market Samurai's data is based around the exact Keyword - rather than including the keywords in phrases and in different ways (Phrase to Broad) ...

For example, type this in Google:
"Cheap Laptops"

You'll get a completely different statistic result than just typing in:
Cheap Laptops

Hope that helps.
Lynn

Did you check with the new or the old interface of Google Keyword Tool? Market Samurai picks data from old interface. I don't know the reason but the new interface just gives a different results. Be sure to check with the old interface too.

Hope that helps!

This is a known issue from inside Google. They have been working on resolving it since June.

It centers around a large discrepancy between the API data and the logged in KW tool. It doesn't *just* effect MS. It impacts all software that uses the API.

I cannot post a link yet but it is easily found on Google if you search for "Difference in search volume between API and Keyword Tool?" and look for the Google Groups discussion for the Adwords API forum. Engineers have commented on that thread and they want everyone to know they are on the case.

There is no time line for when it will be finished.

This has large implications for every one using these tools to find new niches. You have got to be extra careful now that we have polluted data sets.