Author Topic: Keyword Cannibalization Hypothesis Making a Comeback  (Read 3311 times)

martinibuster

  • Inner Core
  • Full Member
  • *
  • Posts: 157
    • View Profile
    • Email
Keyword Cannibalization Hypothesis Making a Comeback
« on: January 30, 2018, 07:26:46 PM »
https://moz.com/blog/how-to-solve-keyword-cannibalization

In every case where someone claims keyword cannibalization, these people are pulling these ideas straight out of their behinds and calling it finger lickin' good.

No patents, no citations, no statements from google nothing to show that these ideas are plausible. Just more a## to mouth thinking.

Been seeing this in newb blackhat circles. What is it with blackhats nowadays, that they don't know sh## about SEO?

They're calling it Content Cannibalization.
Has anyone ever posted a decent debunking of this baseless hypothesis?

buckworks

  • Inner Core
  • Hero Member
  • *
  • Posts: 1191
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #1 on: January 30, 2018, 07:37:19 PM »
That article is from 2007.

Can you point us to something more recent?

ergophobe

  • Inner Core
  • Hero Member
  • *
  • Posts: 5597
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #2 on: January 30, 2018, 07:55:41 PM »
I don't think it's a majorly trending topic, but MB follows industry news and I don't. But seriously, who takes Warrior Forum seriously? Maybe I'm wrong. I don't know it, but I think of it as home to the "guru" internet marketers who launch info products, mostly about internet marketing. Perhaps I'm missing something though.

Keyword Cannibalization in 2018
https://www.warriorforum.com/main-internet-marketing-discussion-forum/1321578-keyword-cannibalization-2018-a.html

What Is Content Cannibalization and How to Avoid It
https://ninjaoutreach.com/content-cannibalization-avoid/

ergophobe

  • Inner Core
  • Hero Member
  • *
  • Posts: 5597
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #3 on: January 30, 2018, 07:59:37 PM »
It hardly seems like an epidemic and this guy's comments on content cannibalization sound very much in line with the song Roger has been singing the last few years.

Quote
Leo Weber, VP of Marketing at Austin, Texas based ProjectManager.com outlines the two “major phases” of a content modeling project:

If you start with your users' needs first, creating a content strategy becomes much easier. For example, if you own an email software company, which page would you want users to see when they search in Google for 'email software'? That would most likely be your home or product page. But perhaps you have 10 or 20 blog pages about email software that are showing in search instead of the homepage. "Your content model will need to identify these content cannibalization issues (where the wrong page is showing for a search) and ensure that your key pages are the ones getting in front of searchers,” says Weber. He went on to explain how this planning should come together into a basic architecture for your website, helping you to purge unnecessary content and focus on promoting pages that matter.

https://www.cmswire.com/content-strategy/content-modeling-what-it-is-and-how-to-get-started/


martinibuster

  • Inner Core
  • Full Member
  • *
  • Posts: 157
    • View Profile
    • Email
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #4 on: January 31, 2018, 01:35:56 AM »
>>>That article is from 2007.

Yes. I know that. I linked that to show who popularized that hypothesis and how long ago.

The point is that it's trending NOW and it is a very hot topic NOW (all over again). 

Here's an example of a recent site. There are a LOT of these newb blackhat sites promoting this old concept as if it's something new. Finger way up their behinds. These are the newbs in the Matt Diggity/Brian Dean orbit.

https://www.fatrank.com/content-cannibalization-google-penalty/

So I'll ask again:

Has anyone posted a definitive takedown of this half baked hypothesis?

All SEO Ideas Must be Challenged
It's really annoying that people are allowed to publish garbage ideas and not be challenged, like the Phantom Quality Update hypothesis (every couple months they pop up saying the data shows that Google is targeting over optimized bad user experience sites again, again again again and again...)

It was a real sin that Rand Fishkin went unchallenged during that time he was promoting the idea that Facebook likes are a ranking factor. The community did not stand up to his ridiculous ideas and a great deal of harm resulted. One guy was on SERoundtable saying he worked for a company that had him working every day on the company Facebook page because they were fully invested in the idea that it influenced their Google rankings.

Nobody challenged the phony Phantom ideas until I called bullshit on Phantom last year on WebmasterWorld. I saw a WMW Google Mod parroting that stuff on a thread about Phantom, that's how deep those crackpot ideas had penetrated this industry. Now we are hearing less and less about Phantom and it's not even being called Phantom that anymore, thank goodness. I really think it's important to the community at large to stand up to the half baked ideas this industry has been churning out.
« Last Edit: January 31, 2018, 01:57:01 AM by martinibuster »

ergophobe

  • Inner Core
  • Hero Member
  • *
  • Posts: 5597
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #5 on: January 31, 2018, 07:15:33 AM »
From the article you link to:

Quote
But only the advanced SEO companies know about the hidden secret of the content cannibalization Google penalty.

Okay, so who can take an article like that seriously? That's like talking about a duplicate content Google penalty. People who don't know the difference between failing to rank and being penalized should not be allowed to call themselves SEO (let alone "the advanced SEO companies"). But there have always been people selling SEO who have no clue about basics (correlation vs causation above all, but simple logic in general).

But the original article you link to is common sense. If you have pages with similar content and similar anchor text and similar titles, you effectively have a duplicate content problem, which is to say a basic information architecture problem.

You don't need patents, citations or statements from Google to figure that out. All you need is server logs, a wall and a bunch of post-it notes (assuming a small site of course). For bigger sites with a million pages, then you have a serious analysis challenge, but the fundamental problem is the same.

Adam C

  • Inner Core
  • Hero Member
  • *
  • Posts: 578
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #6 on: January 31, 2018, 12:19:46 PM »
This is a long read to throw in at a tangent, but I think it speaks to the same concept

https://moz.com/blog/tag-sprawl

This concept is something I have some experience of and have seen positive results from when executed across large scale sites.

I can totally believe the keyword canibalization concept is considerably less of an issue for smaller sites, but when you're in the 100s of thousands, or millions of pages, there can be wins from addressing it.

(Sorry to fight a Moz article with another Moz article - but this is pretty decent write up of the concept IMV)

ergophobe

  • Inner Core
  • Hero Member
  • *
  • Posts: 5597
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #7 on: January 31, 2018, 09:17:33 PM »
Roger, you've been singing about focusing on user experience for the last few years. To me, this is a user experience and information architecture problem that manifests itself in poor results in organic search. Again, the talk of a "penalty" is just ignorance.

So what exact hypothesis do you want to see debunked?
1. That poor IA hurts your ranking?
2. That overlapping content results in lower rankings (link dilution, etc)?
3. That overlapping content results in lower overall traffic (as a result of effect #3 being less than compensated for by additional pages)?
4. That Google has an actual penalty for sites with poor IA?

It sounds like you're mostly trying to debunk #3.

I have heard of many cases where people did content audits (specifically to solve this problem) and saw greatly improved results and cases where people did content audits and streamlined their site and had no effect. So I think a serious takedown of hypothesis #3 is going to have a lot of conflicting factors.

ergophobe

  • Inner Core
  • Hero Member
  • *
  • Posts: 5597
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #8 on: January 31, 2018, 09:26:39 PM »
This is a long read to throw in at a tangent, but I think it speaks to the same concept

https://moz.com/blog/tag-sprawl

Good read though. I've certainly never worked on anything that required calculating a Jaccard Index (or automated methods in any form), but certainly we've all seen sites with 400 tags that can manually be brought down to 40 and the resultant bad user experience and often bad search experience that results from tag sprawl.

Even on small sites, it frequently results in a crappy thin page outranking your main page on the topic, which is a bad user experience.

Adam C

  • Inner Core
  • Hero Member
  • *
  • Posts: 578
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #9 on: February 01, 2018, 02:21:17 PM »
>>never worked on anything that required calculating a Jaccard Index

We didn't go down exactly the same path, and chose a method not listed in that article that is probably easier to digest if you don't have a phd in linguistics to tackle the same type of problem.

A post for the IC when I get some time.

Rumbas

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 1951
  • Viking Wrath
    • MSN Messenger - rasmussoerensen@hotmail.com
    • AOL Instant Messenger - seorasmus
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #10 on: February 01, 2018, 02:33:03 PM »
I hate the word cannibalization, but I DO agree that the concept that deluting a topic with a ton of pages covering the same topic. We've seen some good results pruning content and merging pages for clients.

ergophobe

  • Inner Core
  • Hero Member
  • *
  • Posts: 5597
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #11 on: February 01, 2018, 06:06:57 PM »
I hate the word cannibalization

I think that gets to the heart of Roger's pique. What people are talking about is bad canonicalization and bad information architecture, but if you need a new topic for your SEO blog, you can't be using terminology that goes back to the 20th century.

martinibuster

  • Inner Core
  • Full Member
  • *
  • Posts: 157
    • View Profile
    • Email
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #12 on: February 01, 2018, 06:35:48 PM »
Not a Google Penalty
My first concern is the propagation of the idea that this is an actual Google Penalty. 

Not an Actual Diagnosis
My second concern is that this hypothesis is deceptively a catch-all that encompasses a wide range of actual problems. The solution (paring amount of pages) is generally the same. But the root cause is not.

Ignorance is Not Blissful
My third concern is that some SEOs don't care whether this is a catch-all that masks the real reason why a site has lost ranking.
Some may say it's being pedantic to want to identify the real cause but that's just rhetoric in defense of a bad hypothesis.  Why cling to an overly general explanation that does not at all identify the cause of a problem when you can know what the problem is and accurately diagnose and treat it? It's like saying it's enough to treat a flu by resting and staying hydrated when the alternative is to understand that there is a thing called a flu season and  you can take a flu shot and avoid the flu altogether. There is value in knowing why things break.



ergophobe

  • Inner Core
  • Hero Member
  • *
  • Posts: 5597
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #13 on: February 01, 2018, 06:59:15 PM »
Some may say it's being pedantic to want to identify the real cause

Not if you want to fix it. You might equally take issue with what I said above, namely that the "causes" of "cannibalization" are usually
 - poor canonicalization
 - poor information architecture

Those buckets are still much too large to describe a problem in a way that lets you fix it. Is the "tag sprawl" that Adam posted about a canonicalization problem or an IA problem? Could be both or either. Maybe it's neither. At the end of the day, the root problem could be a UI problem - autocomplete/autosuggest would get rid of a lot of tag sprawl. So the sprawl is an IA problem, but the *cause* of it all is a UX problem.

So yes, it may be a bit pedantic, but it could be food for a good rant.

BoL

  • Inner Core
  • Hero Member
  • *
  • Posts: 896
    • View Profile
Re: Keyword Cannibalization Hypothesis Making a Comeback
« Reply #14 on: February 03, 2018, 12:14:38 AM »
I saw a bit of the FB thread on this.

It seems a few people confused domain clustering with cannabilzation.

Seems like a lost cause arguing in some circles as the general populous will prefer the personalities they know and their position. In the end, is that a bad thing?

Noticed some talk in the same circles about LSI and TF/IDF. the obvious question would be... where are they getting their global values for that calculation?

In any event, there are folks in those circles that make decent money so it's hard to argue the technical merits.