Has the keyword planner search volume metric gone crazy?
-
I use the search volume found in keyword planner to score and weight my keywords in a similar way as Rand showed us in this WBF.
This week I've found that in many cases suddenly the singular and plural version of the keyword have the same search volume. This seems crazy to me as singular and plural is not the same, the intent is different but more importantly they behave very differently from each other when looking at their track record in Adwords (impressions, clicks, conversions, CTR, CVR etc...all different).
For example, here's a screenshot of 4 keywords (singular and plural versions of 2 phrases) with search volume captured a couple of months ago.
Now here's another screenshot of the same keywords taken from Keyword planner today.
Any ideas why this would be happening? Does it makes sense to you? It just seems buggy to me.
Thanks!
-
Well, it's annoying.
We actually got some feedback from our Adwords support saying "the Engineering team is looking to improve/change this behavior as feedback indicates it is confusing for customers. But for the moment, no details yet on what/how it will change."
So let's see, I suppose the more complaints they get from the PPC community the more likely they are to roll things back.
-
Another thing on the subject I noticed is:
For some phrases it will report the aggregate search volume, thus equal s.v. for the singular and plural. e.g. tel aviv hostels and tel aviv hostel show each a s.v. of 1300.
But "Jerusalem hostel" (s.v. reported - 880) and "Jerusalem hostels" (170) don't!!!
So W.T.F. Google trying to do here? Make us realize that they really really don't want us to use k.w.p. data? First take away the broad s.v. data leaving only the lesser valuable data of the exact, then having the bucket aggregation .... What's next? Don't they want us to show the true potential in Google searches to digital marketing clients??? Kind of hard when you have to either give false inflated numbers and say that there is no true accurate number these are all estimates that might even be very far from the real numbers.
-
Hi,
It is the same with Hebrew search phrases same s.v. for singular and plural.
The odd thing is that the search results for each are different. So I don't understand the logic here on Googles side, if you look at them as the same meaning, why are the search results different? they should aslo be exactly the same!
-
Couldn't agree more!
-
Thanks E_F. I've not heard anything back from Adwords support in Europe yet.
Wouldn't it be great if Google would also explain why they changed the methodology as the tool is no longer fit for the purpose it was originally designed for?:-)
-
Finally got a reply from the Adwords support team:
“there was a recent change in how average monthly searches are calculated in Keyword Planner and it is now expected that search terms that are close variants to each other will show the same aggregated search volumes” -
Yes I saw those articles and they're reporting similar findings as this post. Although nothing official yet from Google so I'm still keeping my fingers crossed for this to be a bug...Will be on the lookout for an official update note...
-
Breaking news on SEW is that Google made a change to the tool last week to combine search variants.Articles belowInstead of showing individual keyword estimates per KW or KW phrase, it now lumps in the data together which means that it will show identical estimates for both both. It's no longer possible to see individual estimates to check highest/lowest volumes anymore. All the data has been changed retrospectively alsohttp://www.thesempost.com/googles-keyword-planner-now-combines-keywords-for-search-volume/https://searchenginewatch.com/2016/06/29/googles-keyword-planner-tool-just-became-even-more-inaccurate/Adwords Keyword Planner now seems to combine many search variants, including:
- plurals with non-plurals for any word in the keyword phrase
- acronyms with longhand version e.g. SEO + search engine optimiZation + search engine optimiSation
- stemming variants: -er, -ing, -ized, -ed etc keywords (ie. designer, designing, designed)
- words that can be spelled with or without space (ie. car park and carpark)
- words with and without punctuation (ie. kid toys and kid’s toys)
-
If I had to guess, it seems to be an issue with "close keyword variations" which Google use for matching keywords to search terms in Adwords and the planner tool now somehow sums up the total search volume for all these variations. Which makes it looks like search volume has doubled or tripled or more depending on how many variations you're looking at.
Our PPC team is pushing for answers from our GG account manager but so far they say they are not aware of any changes to the tool and "believe the information to be accurate".
Will update here if we learn anything new.
-
Hi there
I'm seeing crazy inflated keyword volumes from Adwords KW Planner too. I keep a record of the extracts for these 300 KWs so have files going back to Sept-15
Not the same issue as with plural/singular highlighted by the other but...
-
an inflated "Avg. Monthly Searches (exact match only)" number in June versus all other previous months by 6x in some cases e.g. avg monthly searches for a particular KW was 2,400 up to April and now it's gone to 14,800.
-
And also, Google have also retrospectively updated each of the 12 previous months average search query volumes by keyword.
Here's an example attached. Hope you can read it
Out of the ~300 KWs I track, 40 keywords have 2x'd and higher their average monthly keyword volume. 81 KWs have increased average monthly search volumes by 50%
To me it looks like either it's either a bug (or new method) in how Google count average monthly searches or they haven't updated their KW volumes in this tool in the last 12 months.
-
-
Definitely some odd ones - I think Google may be conflating certain keywords, and removal of the ability to see exact match vs. phrase/broad match is definitley an issue, too. In any case, we're sorta stuck with their data. Moz is collecting some additional search volume information via clickstream sources and including that in our buckets for KW Explorer, but that only applies to the US (and won't give precise numbers since we can only get sampled data).
-
correction..."masters in accounting" vs "masters of accounting". It's Monday and the Warriors lost.
-
We're seeing this as well. It doesn't seem as simple as plural vs singular.
"MBA" & "Master of Business Administration" now have the same search volume. "MBA" had 110,000 before and "Master of Business Administration" had 2,400. These variations also have 110K searches/mo now: "Masters Business Administration", "Masters of Business Administration", "Master in Business Administration". Seems like they are bucketing those as the same keyword but then "masters in accounting" is different than "masters in accounting".
-
Yes you're right, the example I provided is not as crazy as what I've seen in our other markets (countries) and I'm fine with Google using buckets as long as I can understand the relative search demand of a keyword compared to another. I don't think we should take the search volumes given by Google as gospel but it's been helpful in the past as a comparison tool.
So what about this - here is a list of plural and singular versions of keywords in Sweden, Denmark, Germany and Switzerland. Something has definitely changed and in many cases relatively small keywords suddenly look huge.
If you run the keyword variations in Google Trends you'll see that they don't have the same search demand. I also had a look in Adwords for the plural and singular version of one of our biggest keywords, this is what their impressions look like over time and now they have the same search volume in keyword planner.
I haven't seen this in all countries but it seems to be happening in the US as well (if you have old search volume data lying around, have a look and see if you get the same increase for the less popular version).
If Google has decided that plurals and singulars are the same and that search marketers should treat them as such, I can learn to live with that as long as I see consistency in this approach (which I don't, France for example still shows different search demand for singulars and plurals). Google should also show the same SERP results for these keywords, which they don't and I believe this is because the intent is (slightly) different between plural and singular search.
I think this is a bug, perhaps I have overlooked this but I can't recall seeing the option "Only show ideas closely related to my search terms" before in keyword planner and if this is a new feature it might be what's breaking the tool. Just guessing here of course but the reality is that turning this setting on/off changes absolutely nothing for me.
So what do you think? Is this an issue for anyone else?
Thanks for your insights and suggestions.
-
The numbers you received in your second screenshot are the same ones I'm getting, and they're pretty similar to what I see in Moz's Keyword Explorer (which bolsters AdWords data with clickstream serach data). I don't know that AdWords is going crazy though - the first screenshot you showed had keywords in the 4-600 range that now show ~1,000 searches? That's not a massive swing, and we know Google uses buckets, even though they show numbers (as Russ Jones pointed out here: https://moz.com/blog/google-keyword-planner-dirty-secrets).
It wouldn't surprise me if these keywords are just on the edge of the buckets Google's defined, and thus swing between one volume number and another.
-
Are you referring to the setting in the keyword planner called "keyword option" and the choice "Only show ideas closely related to my search terms"? Is this a new option in the tool?
Anyway, I get the same result regardless if this is on or off.
It definitely would make sense if there was a setting where I could choose to look at exact match only but I can't seem to locate it. (EDIT: and Google has not changed the definition of search volume in the tool: "The average number of times people have searched for this exact keyword based on the date range and targeting settings that you've selected.")
-
I think the biggest difference is in the match type of the keywords. In the end they're used broad in this case which would make sense that the singular and plural could be the same. Usually when you would have an exact match you're going to see a difference in volume.
Got a burning SEO question?
Subscribe to Moz Pro to gain full access to Q&A, answer questions, and ask your own.
Browse Questions
Explore more categories
-
Moz Tools
Chat with the community about the Moz tools.
-
SEO Tactics
Discuss the SEO process with fellow marketers
-
Community
Discuss industry events, jobs, and news!
-
Digital Marketing
Chat about tactics outside of SEO
-
Research & Trends
Dive into research and trends in the search industry.
-
Support
Connect on product support and feature requests.
Related Questions
-
Keyword Research - Min "Volume" recommendations
Hi, I've began doing some keyword research, the terms I am after are yielding "11-50" Volume in Keyword Explorer. Should I bother optimizing for such a low volume? Looking for existing member / community feedback. Thank you.
Keyword Research | | idlwebinc1 -
How to estimate search volume for multiple long tail queries?
I am comparing Moz, WordTracker and Google keyword research tools and trying to figure out how I can compare the aggregated search volume for a page which targets several similar long-tail queries. For instance, according to one tool, the query 'printable guitar sheet music' has 125 monthly searches, but many small variations of this also have a similar amount. Do I just make a reasonable list and calculate them? Or do I just use "125" as an abstract number of measurement with no concrete reference to a **real **number of searchers, but useful in comparing one keyword to the next, and assume that all key words have a similar degree of consorting variations? I seem to remember Google used to give away some pretty damned useful information in the keyword search tool, including monthly search volume. But now it looks like the metric is not actual search volume, but a 1-100 score based on the search volume. So my questions are: Should I try to convert these numbers into real search figure estimates or just use them for comparing one to the next? Should I worry about trying to combine the aggregated variations of long tail keyword searches, or should I just choose one good keyword and let that represent all possible variations?
Keyword Research | | guitarsites0 -
Broad Match Search Volume
Hi All, In Google keyword planner, "broad match" option has been removed. I want to know that is there any other method in Google keyword planner or any other tool or any technique through which we can find out the "Broad Match" search volume??
Keyword Research | | RuchiPardal1 -
How does SEOmoz account for articles (a, the) and transitions (in, on) in keywords?
For instance, if I want to track traffic for a sports memorabilia store in Miami, and I only have "sports memorabilia shop Miami" as a keyword to track, will it track traffic of people who visited the shop's site after searching for "sports memorabilia shop IN miami"? Or would I have to use both "sports memorabilia shop Miami" AND "sports memorabilia in Miami"? I know Google doesn't show results with "in", "and", "the", etc., but Google Analytics tracks them, so I'm curious about how SEOmoz does it. Thanks in advance.
Keyword Research | | simply1800 -
Using a country in your keywords
My company has recently started offering services in South Africa - which is great - same language - lot of potential there. But I suspect our keyword research is suffering from "thinking like a foreigner". i.e. "buy cars South Africa" rather than "buy cars". Wouldn't google.co.za naturally assume: 1) Anyone using their service for "buy cars" is looking for "buy cars South Africa" (I know it's more complicated - Google will give different results the more it can figure out about your location) 2) That any co.za site optimised for "buy cars" is automatically 90% optimised for "buy cars South Africa" too - so isn't a productive long tail query (against something like “buy sports cars”) I appreciate that some companies will be TARGETING foreign browsers (i.e. holiday accommodation South Africa) so will need to explicitly optimise their on-site for the country for “foreign search engines”. But I'm asking about companies in South Africa targeting people from South Africa.
Keyword Research | | Ali310 -
Keyword search tool or API
Anyone have a suggestion for a simplistic (scaled down) version of a keyword research tool that I can put on our website? We need our clients to be able to perform some basic keyword research on their own without getting lost in the details of something too robust. We would like to control aspects of the tool, capture the data from the site visitor and move those results (from their searches) into a database or other web application.
Keyword Research | | webindustry0 -
Is there an ideal ratio of keyword difficulty to search volume?
I used the keyword research tool to pull data on potential keywords, the report returns a percentage to represent the competition and difficulty of a keyword. Is there an ideal ratio of percent difficulty to search volume
Keyword Research | | AdviceElle0 -
Two for the price of one: Can I rank for multiple keywords when only targeting one keyword?
If I'm optimizing for a specific keyword, is it accurate to assume that by ranking for that specific keyword that I will also be able to rank for similar or root keywords merely by ranking for the original keyword? For example, if I'm targeting 'free online bucket list' is it safe to assume that I will also be able to inadvertently rank for 'online bucket list' or 'free bucket list'? Can I assure clients of this? Or if I'm targeting 'Colorado grocery store' should I also naturally rank for 'grocery store Colorado' and not need to make both of these my targeted keywords?
Keyword Research | | derrickkuhn0