Wordtracker Keyword SEO Tool Spammed By Brute Force Hacker?

While researching SEO keywords in Wordtracker, I noticed obviously spammy results compared to Keyword Discovery or Google Keyword Tool.  This is not the type of website keyword spam that gets sniffed out by Google Panda filter, but rather keywords that must have entered the Wordtracker database through some kind of bug or hacker virus. It makes you wonder, if the data is compromised how reliable are the numbers for normal looking keywords?

Wordtracker SEO keyword tool spam resultsHow Good is the Data From Your Keyword Research Tool?

The most important thing for any keyword research tool is its database. No tool can possibly capture 100% of searches performed.  Any keyword tool that provides search volume must capture searches from users that in some way allow their searches to be tracked.  This may be a known panel of users, or users that install a browser toolbar or use a service that in the terms of use agreement (which no one reads) requires users to allow their searches to be tracked.  The fact is, since searches are captured from a limited number of users there is likely to be some skew or level of inaccuracy in the data collected compared to the much greater number of actual searches performed worldwide.  Bells and whistles aside, the quality of the keyword database defines the quality of the keyword research tool.

Example of Keyword Spam in Wordtracker

It appears to me that the Wordtracker database is capturing searches that include a substantial number of systematic searches, probably performed by a program on someone’s computer that is supplying data for Wordtracker’s keyword search volume database.  When researching keyword search volume on “IT security”, a terms I would expect would have a pretty decent volume, slight variations of the same 18 word phrase ranked above and below IT Security.  The variation in the 18 word phrase was only 2 letters.  If this were a common typo for what you would think would be a popular phrase, I could dismiss it.  But the keyword phrase is 18 words long!  And each variation was searched about 100 times!  The 2 letter variation makes it look like some kind of brute force hacking attempt.  Brute force hacking is when a program is used to guess a password by trying every possible character combination.  In this case, it appears like a program is attempting to fill a search form using character variations of a phrase copied from a login page.  Possibly a virus on one or more computers included in Wordtracker’s user database was attempting to hack into a login form on a page that also had an embedded search engine form.

Keyword Discovery results verses WordtrackerWordtracker Verses KeywordDiscovery

A glance at KeywordDiscovery results for IT security show absolutely no phrases similar to that spammy phrase found in Wordtracker.

The Right Mix of Keyword SEO Tools

I’m a regular user of Wordtracker and KeywordDiscovery SEO keyword research tools.  I use the paid version of Trellian Keyword Discovery and the free plugin for Wordtracker SEO Blogger.  I used to use the paid version of Wordtracker from 2004-2010.  I had access to Keyword Discovery, but never really got into it because the interface was not intuitive and hard to find your way around without clicking and trying everything. I switched to Keyword Discovery in 2010 when I worked as the search engine optimization research & development technician at Medium Blue, an SEO agency in Atlanta.  I basically switched at that point, because that was the keyword research tool they were using.  They had the enterprise package, allowing unlimited keyword lists. Even now as SEO manager at BKV (full service advertising agency in Atlanta), I am sticking with Keyword Discovery as my first choice.  However, I am training my team to continue using Wordtracker to inform their keyword research.  Wordtracker seems to pick up keywords that KeywordDiscovery does not.  The search volumes don’t compare apples to apples across Wordtracker and KeywordDiscovery, but you can still get an idea of relative search volume compared to other nearby terms.  Google’s Keyword tool can give a third opinion.  But here again, trust your gut when you see outrageous estimated search volume that looks too good to be true.  Google holds its data close and can provide anything it wants as an “estimate”.  Especially if it entices PPC customers to spend more money on keywords that Google says are going to be popular.  Therefore, I go to Keyword Discovery first, and look to Wordtracker and Google Keyword Tool to get ideas for expanding my list.  I then dump those keywords into Keyword Discovery, export search volumes as Excel and include a column for Google Keyword Tool monthly search volume.  This way I can look at both counts side by side and prioritize which keywords to target.

There are 3 comments for this article
  1. Mal at 2:53 am

    Hi Greg,

    I’m Mal and I work with Wordtracker.

    It’s certainly an interesting set of keywords you’ve found here – and you’re right, it’s clearly spam. Deeper analysis shows that it’s largely the same string of text that’s being used, prefixed by a different word each time.

    We do have a ten-stage filtering process which is remarkably effective at removing most spam, but it is true that sometimes things do slip through. We’re not able to manually test every keyword that comes through our system, so it’s a good thing that we have diligent users who let us know when situations like this do arise.

    It’s worth mentioning, however, that while SEO Blogger (the free tool you’re working with) is affected by this, our subscribers are able to easily filter out these offending keywords and still have just under 1,600 search terms available to assess.

    The phrases you’re seeing in the database are there as a result of someone external to Wordtracker setting up a script and pointing it at the keyword entry field of one of our partner search engines and the keywords the script is using being passed on to us. The likelihood of it being the result of a virus in our own data chain can be ruled out through our own monitoring and analysis services.

    I’ll be looking at cleaning these phrases from the database – it’s an unusual set of unwanted keywords, and it’s rare to come across a niche that’s so heavily spammed – this could easily have happened over the course of a couple of hours.

    In terms of your comment about search volumes not comparing across different tools, this is normal and to be expected – every keyword research tool (including Google’s) uses a sample of data, and so those samples will vary from tool to tool. As you’re undoubtedly aware, they’re there to provide insight and guidance rather than gospel truths about search behaviour – so you’re exactly right to use multiple data sources for comparison.

    Do feel free to contact us directly if you have any other questions about this – you can reach us at support@wordtracker.com

    All the best,

    Mal

    • Gregory Lee Author at 12:28 pm

      Hi Mal. Thanks for letting us know you are cleaning up this spam from WordTracker results. How would you say your accuracy compares to other platforms out there?

  2. Mal at 6:15 am

    Hi Greg,

    yes, we’re on the case with this and the work should be complete fairly soon.

    In terms of your accuracy question – it’s a slightly difficult one to answer directly. No keyword research tool (even the ones delivered from major search engines for PPC research) can deliver results that can be described as pin-point accurate. There’s no ‘gospel truth’ available to SEMs

    Any tool used for keyword research is intended as a guide. We do recommend to all of our users (whether for the free or subscription tools) that it’s better to look at the relationships between the keywords rather than the raw numbers themselves (although the numbers, particularly the competition figures, do give a deeper insight) – it gives a much clearer overview – and that applies to using any KW research tools, not just ours.

    In terms of Wordtracker’s accuracy, though, we have a representative sample of US searches that thousands of customers and free tool users have found helpful over the years.

    I haven’t spent a huge amount of time recently researching other platforms, but I do know that many of them rely on data from Google AdWords, which, while it’s useful in many respects, has a heavier emphasis on PPC research than organic search, which is where Wordtracker’s strength emerges, particularly in terms of providing access to the long tail of organic search – the very language that real people use to find solutions to their web questions, be that as part of a buying cycle, or just a search for information.

    Do let us know if you have any other questions.

    All the best,

    Mal

Leave a Reply