aakk9999 - 11:13 pm on Sep 20, 2013 (gmt 0)
Adwords Keyword Planner Tool (KTP) shows the average exact monthly searches for a keyword phrase entered in "Enter keywords" box.
So if you are looking for the volume only, it does not matter whether you enter keyword phrase as [kw1 kw2], "kw1 kw2" or just kw1 kw2 since the match type (broad, exact, phrase) only affects figures on CPC and competition.
Note my previous emphasis on "average". If you are comparing this data with your visits for a month, you need to take the visits for the whole year and average it per month. Otherwise you may be taking the clicks from a month that in fact has low search volume versus the average monthly searches, that are influenced by higher search volume and the CTR may look lower than it actually is (or vice versa).
I have done some study comparing Adword Keyword Planner, Google Analytics and WMT Query data for a branded name search that has no competition (nobody bids/tries to rank for this brand name) and where I am very certain that the site always ranks #1. Below are some figures to compare.
Keyword Planner Tool avg. monthly searches: 390
Because of the seasonal trend and the problem in calculation explained above, I compared the figures for July 2013:
Keyword Planner Tool July searches: 1000
Google Analytics visits July, for this keyword, Google only: 870
Google Webmaster Tools Web Impressions: 2000
Google Webmaster Tools Web Clicks: 900
Google Webmaster Tools Web Avg. Position: 1
Google Webmaster tools Web CTR: 45%
Google Webmaster Tools Mobile Impressions: 700
Google Webmaster Tools Mobile Clicks: 90
Google Webmaster Tools Mobile Avg. Pos: 1
Google Webmaster Tools Mobile CTR 13%
I have not included Image search figures as I have only selected "Google" in KTP which does not include image/map etc search volumes.
A bit of analysis:
Firstly, the number of impressions is almost 3 times the number of searches. My speculation is that this is because if a site has more than 1 result in SERPs on the 1st page, each result count as an impression. Similarly, if the visitor clicks through to 2nd and subsequent page and there is SERP entry there, it also counts as an impression of that same search query. Hence 1 search (from search volume) can result in more than 1 impressions reported.
Second interesting thing is to compare the number of GA visits with the number of KTP searches and number of WMT reported clicks (web + mobile).
GA visits: 870
KTP searches: 1000
WMT clicks (900 + 90) = 990
I am reasonably happy that the number of searches as reported in KTP versus number of clicks as reported in WMT pretty much ties up. As this is a branded search, it is extremely likely that whoever searched for it, clicked through to the site, and CTR of 99% (based on the number of searches) ties up (although the figure is also a bit skewed since one search may have more than one WMT click, however from the experience of following the visitor navigation pattern through Statcounter, this is a rare occasion.)
WMT reports more clicks than GA reports visits for this keyword. The reason for this is *probably* because of "not provided" searches (although I am not entirely sure whether GA definition of visit of "interaction with the site within certain timeframe" would also impact this figure).
From the above example it is also obvious how WMT reported CTR is completely misleading. WMT calculates CTR based on impressions. This makes sense for Adwords where Google has a policy of not allowing multiple ads from the same site, but does not make sense in Organic where the same site may have more than one entry in SERPs.
For example, having 3 impressions (because of 3 entries appearing on page 1) and only one click through to the site would result in CTR being reported as 1/3 = 33% whereas the more useful number would be to calculate CTR based on number of searches in which one or more results are shown (e.g. 1 search, 1 click = 100% CTR).
Drilling down in WMT by clicking on keyword will give impressions, clicks and CTR for every position in SERPs and these numbers are a bit more realistic.
As it can be seen, there is still lots to guess despite figures given. But what I believe, based on the information I have studied, is that:
- the exact searches per month figure reported in KTP is reasonably accurate
- impressions are skewed (overstated) because of multiple site listings seen within one search
- hence the CTR is skewed (understated) as it is based on impressions
- searches versus visits figure from "google organic" does not tie up because of "not provided" (and maybe also because of GA definition of a "visit")
- comparing number of monthly visits with avg. monthly searches does not work unless you average your monthly visits for the last 12 months