Thread by @darth_na

@si1very Yes/No/Maybe?

The distinction is important, because a penalty is an applied negative.
Getting hit with a penalty for things like mass-cookie-cutter or stolen content is vastly different to G filtering out X of Y pages.

>>>

@si1very >>>

Then there's the issues of vocabulary/ & specificity.
* Internal duplication
* External duplication
* Copied/Stolen content
* Highly similar content
* Cookie-cutter-content
* Boilerplate content
* Low "original" content
* Mashups
* Enhanced mashups

Each is different.

>>>

@si1very >>>

If you've got 50+ pages, and have some canonical issues (such as protocol, domain/TLD, paths/queries etc.) - it's problematic.

1) Diluted and/or Dispersed values typically lead to lower rankings.
2) Auto-canonical/Filter may lead to fluctuating rankings

>>>

@si1very >>>

But typically G will resolve to one version and push that (in most cases).
These can be fixed with Hard/Soft canonical methods (301s, CLE/CLR etc.)

But that's not quite the same as external dupes - esp. with things like syndicated content,
which you seldom control!

>>>

@si1very >>>

For external dupes - you may well be stuffed (better off to alter your version (and learn Not to syndicate full/exact versions! (Not even if they provide a link and use a CLE/CLR!))

Then there's nigh-identical and cookie-cutter,
again ... different!

>>>

@si1very >>>

Nigh identical and cookie-cutter are often both intentional - and serve similar aims...
... to get multiple "versions" ranking and earning traffic.

The main differences are the quality, the quantity, and the type.

>>>

@si1very >>>

There's seldom a good reason to have multiple versions of an informational piece.
Typically highly-similar/nigh-identical applies to "products" and "services".

Legitimately - you may have numerous versions
(such as different colour iPads, with different RAM etc.)

>>>

@si1very >>>

If done properly - these are canonicalised.
G will then attempt to show the most relevant out of the set to match a query.
(if someone searches for "car sales in X", G may show your X page, even if the designated canonical is the Y page!).

>>>

@si1very >>>

Cookie-cutter pages though are often in greater quantity, with only the "keyword" different.
Worse, the quality of the content is often lower ... G doesn't want it, and may penalise it, or the "quality" is marked down (lower crawl/index rates).

>>>

@si1very >>>

Then there's the confusion over things like "boilerplate".
Having a % of content that is identical across pages is seldom a cause for concern.

Really, the only time it's an issue is if you have little other content (and/or it's not original/unique).

>>>

@si1very >>>

Then we have "mashups".
Be it your own content, or from external sites ... multiple bits of content that can be found elsewhere.

Sometimes this can be good for users (curated, relevant, useful) ... but often it's subpar, and if G see's poor reactions, it tends to die

>>>

@si1very >>>

Google's reaction to "duplicates",
and the impact it can/does have on:
* indexing
* crawling
* ranking
can and does vary - depending on the nature, quantity and quality.

>>>

@si1very >>>

Having less than optimal value dispersal
is not the same as
G refusing to show a % of your site,
or
G removing a % of your site from it's index.

Different results, caused by (slightly) different problems, with different solutions.

>>>

@si1very >>>

So … yes - in some cases, it's moot.
In others, it's an important distinction.

What matters is understanding the problem,
knowing the cause
and applying the correct solution.

It may require canonicalization, or altering content,
or simply deleting spammy garbage.

>>>

@si1very >>>

And the only way that's going to happen is if people use correct descriptions.
Things like "boilerplate" aren't duplication.
Highly similar/expanded mashups aren't duplication.
Even cookie-cutter-content isn't technically "duplication" (as they aren't identical).