Is Google AdWords Keyword Tool Now More Accurate?

good spots… going to play with some of my ow notes ad see if my KW sets have changed…

Damm it , hope they give me back all the money i spent on keyword rich domains recently

I was checking out Gary Taylor’s website (www.3ac.co.uk) after Think Vis last night and he claimed that “‘Fly fishing books’ gets about 3,600 exact [local] searches per month in the UK.” . that was on May 31st. It is now 210 exact local searches. Bit of a drop. funny he didn’t mention that and nobody picked him up on it at Think Vis.

let me check my kw, any info about next pr updates?

That’s interesting! Was having an issue this morning where traffic volume displayed in old keyword tool was grossly inflated. Will be using new keyword tool from now on.

Still a lot lower than volumes being shown in Traffic Estimator from an Adwords account though… the Traffic Estimator figures seem to look much more accurate based on a very small/quick look at CTR (not scientific at all)

Great post guys. We’ve checked some keywords and there seems to be a huge difference in Local Search Volume and the statistics from our Adwords campaign. We had way more traffic from our Exact Adwords campaign then the Local Search Volume suggested.

Thanks for the updated post. I’ll go check over a few of my keywords now and see if the numbers tally more accurately with the analytics data.

Also a nice example why the Keyword Tool isn’t more accurate:
Keyword: macbook
Country: Netherlands
Language: Dutch

Results: 0

@Mark – sorry to hear that dude

Goood spot – for some of the terms I check, the numbers reported they’re reporting now appear to be more accurate (and much lower). Be interesting to hear if other people are finding the same thing.

Just checked a term for a client, funny how it’s dropped from 246,000 to 1,300. Thankfully I didn’t trust this figure in the first place as it didn’t correlate with other data!

If nothing else I’ll have to start using the new tool now as I hadn’t actually noticed the difference between the beta tool and the legacy tool and had continued using that recently.

The logged in and logged out values now seem to be the same as well (unlike when I wrote http://www.malcolmcoles.co.uk/blog/why-you-shouldnt-use-googles-keyword-tool-for-seo/)

Can anyone say class action lawsuit?

Lots of data crunching to do to compare… good spot

What I really want to know though is why your topsy button seems to be working today… at least a little, and mine isn’t

Excellent spot, and after a few checks, I think YES it is on some terms….

I clicked Submit too fast. Anyways, I’m not sure what your point is here in this post. I can say, nothing in my recent testing has changed. Perhaps everyone is talking about something different. This is all I see it?

The beta keyword tool and previous interface keyword tool. I still see what were 500,000 volumes (exact) showing 1000 or less with the beta tool. That is what I’m talking about. I don’t get it how this can be considered more accurate or how people are agreeing that something has changed since your last post about this.

I think you need to clarify what you’re talking about. The 2 versions of the keyword tool are so different in terms of returned data, that it is simply inexcusable on Google’s behalf to not address the issue.

So you have a domain you are selling and you think it’s worth $5000. Yeah? Check now. Maybe that domain shows a volume of 250 people a month. Now what? What about the domains you may have bought based on returned data that is well, simply a completely and utter joke. Is that ethical? Is that proper? Is that not worth an official comment from Google?

@Andy it wasn’t working earlier m8 – it’s probably still out, 5 seems low.

Sorry I’m not trying to spam! I just reread your post. I’m curious though, why do you think people will use the tool for their online business? I’m asking because how can you say whether the “new” data is accurate and the “old” data was wrong? How can you say that the data is accurate at all? I’ve seen enough in my own testing to say that I cannot, and nobody really should be basing anything on what data they are seeing. People can do as they please, but I would put an abrupt stop to using the keyword tool. At least until something is said officially or until the variances are explained or corrected.

It’s fine to have inflated numbers or deflated numbers. The issue is, keywords are being flip flopped in terms of where they rank. It can show a previous #4 searched term in the #1 spot now. And you think people should be confident in the data?

Data that you cannot trust or that has show flaws is as useful as no data at all. It’s as simple as that. If you can’t trust it, or it isn’t reliable, then it’s a mistake to even contemplate using it.

Thanks for bringing to light your findings and helping the discussion come to the forefront. I see a lot more rumblings now about this. Cheers!

Yup. Keyword research is crucial. I have in the past found the Google sktool to be pretty accurate as it matches up with our traffic and I’ll be comparing the sktool numbers to the new tool’s numbers when it comes out. As for Google explaining more thoroughly where these numbers come from, I second your interest in knowing more! It’s kind of weird that they aren’t clearer about that since yeah, inaccurate numbers can really screw you over.
First comment from me, but I’ve been reading the blog a lot lately and really enjoying it, so thanks 🙂

gman, have you been buying lots of domains based on the old tool’s figures or something?

our experience has shown for years that the numbers on the old tool were 20-30x optimistic, when compared with what #1’s for “significant” kws actually deliver.

what it did do ok however was to give you comparative traffic levels, ie if it said kw 1 was 20k and kw 2 was 15k, then a position one for both of those would turn out with kw2 delivering 75% of the kw1 hits, albeit at around about 1k or less.

the new tool, multiplied by Brent (Market Samurai)’s 0.54 factor, multiplied again by your position factor (ie 40% for pos 1) is now estimating 2-4 times high from our stats.

as this is actually the closest we’ve ever been to being able to accurately assess likely numbers, I’m actually fairly happy with that.

but then I haven’t been buying domains based on the old figures.. 🙂

I’m not speaking so much for myself. I depends on your understanding of statistics. I could care less about the actual number, so long as that number is correct relatively speaking. Is a result of 500,000 vs. 1000 something to quibble about? In my world, yes it is.

High or low doesn’t matter. Ranking does. If you say red shoes is #1 searched term and blue shoes is #6, then fine. But what is happening is now, that Google is saying “well actually blue shoes is #1 and red shoes is #7”. The data is about that reliable. The actual volume of searches is statistically insignificant so long as it’s properly represented.

To each his own. If you are okay with completely different results depending on which tool your using (new vs. previous) then good luck. How do you know which one is correct? Google said nothing about is so far so we are all guessing. I’ve done enough testing and sampling over the years to know the data simply CANNOT be trusted. My opinion based on my experience. Everyone can do what they want. These examples here show one or two tests. It’s much more complicated and corrupt than that imo.

@ Sean

Hi Sean

I did mention the huge difference in the presentation and warned people to be careful. My example however was with “paper party bags” dropping from 5,400 on the one tool to 440 on the other

See the post below, its not an exact science:

http://www.3ac.co.uk/exact-searches-not-an-exact-science

I would love to hear Google give a clear cut explanation too

@kev grant. Not sure why you would put a smiley face about people buying domains based on inaccurate data that Google was publishing. That’s funny to you? I don’t get the same “joy” or “chuckle” out of a lot of wasted money and misinformed business decisions based on again, published data from Google. The name you can trust right?

Again, you find humour in it. I certainly don’t see humour it. The smiley is slightly more than what I would consider to be arrogant and pompous. Let’s leave it at that. My advice? I say wipe the smile off your face. Karma is one of those things in life that if you believe in it, then you will never be offended by jerks. You know their day will come.

gman, you misunderstand.

it’s not about enjoying others suffering or loss, it’s about being happy that Ive been in the game long enough to understand that those numbers were always massively optimistic and not base any direct transactions on them.

sorry if that’s not the case for you and others.

Fair spot and @KevGrant, your wisdom is apparant.

I didn’t think the keyword external tool was ever designed for an SEO-crew audience. The only thing it’s ever been reliable for is getting an idea about how much budget 100% I.S. will need for any given keyword (i.e. show all columns, find CPC, times that by the search volume, done).

If a specific keyword is literally the lifeblood of your business, or your planning to pay off hooker/coke debts, run an impressions test for a week or two. Get your own data, it’s not that difficult, and it might save your kneecaps.

PEACE

Excellent discussion on Adwords stats,being new to domains and SEO i have been wondering why one keyword may show 200,000 then 900 next month.
Search volume is important for an investor looking for ROI with a domain purchase but a good keyword domain is more than numbers it is an important description of the company,a good keyword domain is still the best way to drive business,the web is still young and being that keyword on the web will only increase traffic but also relevancy to customers as long as you have the relevant domain, high search or not.
It would be nice to know true search volume to gauge prices for domain sales, but also what you pay for Adword campaigns,so as long as you have a descriptive simple keyword that is memorable then you are ahead of the game.
PS.Google should explain why the stats have so much discrepancy we deserve that for so blindly following them all these years.

So, why was it so innaccurate in the first place? How is that 19,000 please less search fish pond supplies, lol? I

Timely topic, David.

Just last month, the Adwords Keyword Tool was showing 2.24 million searches for a particular phrase locally AND globally. I thought then “this can’t be right?” Turns out it wasn’t. Now it’s showing 12,100 locally and 90,500 globally. But after reading yours and every one else’s post I doubt even these figures are anywhere near accurate. Who knows?

I call this issue ‘the lottery effect’. No one will ever give you information that will make you money, for free or so easily; they are more likely to cease the opportunity themselves. Like the saying goes, if it seems too good to be true, it usually is!

I think the only SEO tool worthy of a look is Google Insights as it shows you a trend for the search terms and tells you flat out if it has enough data to support its results. Everything else is completely unreliable and a waste of time.

Do you think this increase in accuracy is down to Google instant being rolled out? I would also love to get your take on the ramifications for Google Instant, with respect to seo.

Regards

Mansoor

@Kay – good point, if it doesn’t show anything in Google Insights, it probably isn’t worth going for – plus you can benchmark it against another keyword you rank for.

Great call-outs. Analytics was already showing attribution higher than the keyword tool’s reported search volume for some terms, now that difference it going to become even worse. Ridiculous when my functions are returning a 1000% CTR.

The search figures problem has been flagged up before – this thing though it seems perhaps G has realised something is up. I would love to know how many people use dthe Adwords search figures to base their keywords on.
My other point is that there are now quite a few SEO apps out there powersuite, webseo to name but 2 – Im pretty sure their keyword research tools use the same figures that have been flagged to be dodgy .

There may well be hundreds of thousands of sites out there waiting for non existent traffic to arrive

There are rumors that Google is going to be forced to compensate all advertisers for opportunity cost – where they wasted advertisers’ time by intentionally misleading them with falsely inflated viewership data.

An example of this would be where you believed that 100,000 people were seeing an ad for your website when, in reality, only 1,000 people were seeing it. Meanwhile, in the time you wasted advertising your site to almost no-one, a competing site became wildly successful because they weren’t tricked by Google.

A few insiders in the legal department at Google are predicting that compensation claims could be in the billions of dollars.

Guess I should add my 2 cents in Personally I fell in the boat that took exact results and tenthed them (ie divided by ten) to get an idea of search volumes.

Another point is that this is an AdWords tool – that means it covers and calculates data from all of Google’s networks as well as all AdSense publishers.
Ex: if someone searches for red shoes as gman said and finds a site that also has AdWords displayed for red shoes, that would skew the data.

I would still say half the current data….

Plumber Sydney – my clients site ranks #1 in Australia and only sees around 40-50 visits monthly while the Exact Search on the tool suggests 1000 searches

Even if #1 should give you 10% – that’s double what they are actually getting.

Just my 2 cents 🙂

The search volumes given by Google shouldn’t be considered 100%, neither before nor now

Lawsuit by advertisers. Google can not said is approx. data because is a tool provided from their business and make a huge profit from small business to big enterprises.
I can’t undestand how many people support this kind of business closing their ayes to other alternatives.

A google employee has responded to this issue at

http://www.google.com/support/forum/p/AdWords/thread?tid=5241a33035617b73&hl=en

Apparentely as mentioned previously in this thread, he says the old tool takes it’s data from google and it’s search partners where the new tool is just google. Very misleading

I’ve been using Wordtracker recently to compare search volumes, this has enabled me to gauge potential search volumes more accurately – for example take the keyword ‘Used Mercedes’

Old Google keyword tool, the local UK exact searches are 301,000 per month

New Google keyword tool, the local UK exact searches are 3600 per month

Wordtracker exact average (say 80% is Google traffic) – 5500 per month

The above example shows that the new Google tool is more accurate than the old one.

Google has been deceiving its advertisers since it started providing data.

Now they are trying to minimize liability and coming good (we hope) with search volume data is probably part of their cover-up plan.

In hindsite, wordtracker data made more sense from day 1…a shame it is so expensive.

Looking at this and Keyword volumes have dropped considerably! from 100’s of 1000’s to less then A thousand in some cases!

However from looking at analytics data and using conversion rate click through stats I would say the new fiigures are perhaps just around 60% or 70% of the true figures, It seams google keyword tool has gone from really high unbelievable volume data, to extremely low volume data. But it does seam more accurate.

This has a large inpack on what keywords should be considered.

I’ve always used exact match type to select keywords but now with these low volumes I am thinking it may be best to use broad match type ? With the hope in possibly creating more scope for keywords choosen.

I have tried the google keyword tool and it always seems to give inflated results, I will try wordtracker as fergus clawson suggests, are there any other alternatives to the google tool?

Just wanted to say thankyou to David and his informative blog posts – and everyone who contributed. I too am based in the UK where Google is the dominant player so have really appreciated this read 🙂

It was good but now it`s best and too much accurate now thanks 🙂

I know many many guys bought a long tail domain bases on the Keyword Tool.

I dont believe the numbers googles adword keyword tool produces, I belive the real number of searches are lower. I also find it produces americanised search results for searches that are supposed to be local to the uk.

I have seen some inflation but I also try to be realistic with keyword research. I think a good head on your shoulders can go a long way in figure out those keywords that are likely inflated. Other times, look at other similar keywords in the niche to see what the volume looks like on those. A niche that has a good demand will likely have several keywords with a high volume of monthly searches.

Aww Snap! Looks like marketing consultant is back up to 60,500/27,100. I don’t trust it. I use it to compare relative values, but I don’t trust it.

How does localisation work in all of this?

Presumably you still have to be careful with the results that you get from the tool because the search results will be skewed by Places entries or the natural inclination to show stuff from nearby.

What do the experts think?

Gareth

I really don’t agree with you. I did compare the visits on my site and AdWords Keywords Tool and it match. Plus, your data, it is clearly not enough.

I wish someone would invent a really reliable keyword tool, they all seem to have weakness in their stats which is of no use to the webmasters !

I have used the adwords tool but rarely did I come across such a huge variance (“marketing consultant” 110,000 to 1,000). This is even more trickier because we know such a keyword has the potential to be searched many times. So there is no way of doing a “guestimate”. The current versions of many of the keyword tools are more accurate due to the fact that there are many companies in the business of doing keyword tools!

Is Google AdWords Keyword Tool Now More Accurate?

60 Comments

Get in Touch