Responding to drastic change of intent for a keyword

[bernard]

How do you respond when a keyword changes drastically in intent and you lose rankings?

Example:

A page that has an aggregation of products and a filter/search function, that is aimed at both "Cheap" and "Sale" keywords. You have both in the Title tag such as: "Cheap widgets - find widgets on sale here"

Google changes intent, so that now the "Sale" keyword is now something different. In this case, cheap is still doing ok with the collection of widgets, but "sale" has now turned into mostly ecommerce intent with a few info/best intent.

Then what?

A few options:

1. Keep the page as is and work on linkbuilding, while waiting for intent to hopefully change back

2. Keep the page and Title as is, but change content, try to more match the new intent, while not deleting the existant content

3. Split the page into "cheap" and "sale" pages and have different intents

What do you choose?

I am leaning towards 3.

For some time, I've seen this play out, in that the pages optimized for "best widget" increasingly rarer rank for "widget" and likewise "widget" and "best widget" rarely rank for "sales", "discount" and "cheap".

So actually, you would do best to split your product keywords into:

1. Widget main - a wikipedia like site with links to subpages and maybe one "best product" for each category (not pillar)

2. Best widget - product reviews and pretty much only that

3. Cheap widget - a list of actual cheap widget, presented as list or search/filter

4. Widget sale/discount - a page with current, updated sales, and perhaps, links to actual sales pages with the advertiser

The thing is, we don't know if these current intent categories are subject to change. Maybe Google decides that for a particular keyword the "widget main" keyword intent should be "best widget", because the widget is not served by wiki like content. This could be some very commercial keyword as an example.

What do you think about the above?

 

[secretagentdad]

Strong affirmative vote for

3. Split the page into "cheap" and "sale" pages and have different intents.

I always error on the side of be narrower.
It allows for more focused tweaking of the messaging to improve experience or conversions.

 

[bernard]

Strong affirmative vote for

3. Split the page into "cheap" and "sale" pages and have different intents.

I always error on the side of be narrower.
It allows for more focused tweaking of the messaging to improve experience or conversions.

I agree.

I guess, I am cautious because of Panda in the past.

Wasn't Panda about punishing that kind of micro targeted content?

You can of course, and I think we should, argue, that the keyword types mentioned above are pretty major types.

What then about smaller keyword variations such as widget size or widget color or widget material, because right now, I'm seeing Google really like those variation pages like "Coffee table 90cm" and the like. Now that, I would be sure would be Panda, but of course, if you actually deliver 90cm coffee tables, then you helped fill that intent, you didn't fluff up your page with senseless content.

 

[Ryuzaki]

I agree. #3 is your safe bet that defends you against future changes in intent. If you laser focus on the intent, then you'll be fine:

• Best = the top choices regardless of price
• Cheap = the least expensive choices
• Sale = the better choices that are temporarily a lower price (eComm results)

These are definitely of different user intent and I'd expect Google to break up the SERPs like this more and more in the future in order to better serve the user's needs.

Wasn't Panda about punishing that kind of micro targeted content?

Panda wasn't about micro-targeted content. You're probably thinking of the Farmer update that went after sites like eHow and Ezine Articles, etc. The whole Demand Media fiasco. They were pumping out very similar articles with slight variations for every slightly different keyword. It was like manual spinning / ad-libbing. They made gobs of money in the meantime, but it was definitely basically "doorway pages" to some degree.

Your last paragraph seems 100% fine and legit if they're eCommerce results. Not if they're content-based results like "the fastest way to blank" and "the quickest way to blank" and "the speediest way to blank".

 

[bernard]

Would the latter type (size, color etc), also apply for affiliate sites?

I have a new site basically doing this, but by listicles "23 coffee tables made of wood" etc.

 

[Ryuzaki]

I think you'd be fine doing variations like "23 coffee tables made of wood", "13 DIY coffee table ideas", "17 IKEA coffee tables", etc.

You're not going to have 1,000,000 posts that are all variations of slightly different keywords. My example above in this post are all different user intent that will demand different content. The Farmer stuff was literally trash with very little differences, made to be keyword optimized and nothing else and rake in 5 visits a month, times 1,000,000, with Adsense on the page.

You have to be doing some really ridiculous crowd-sourced or computer-automated crap to get hit by any of those older filters. It's nothing you're going to run into if you're building high quality sites.

 

[bernard]

Another thing that I've struggled with.

Why does Google show a subpage that is the opposite of the search?

Example, if you search for "soft pillow" Google might show pages for "firm pillows".

Is it some weird AI minority report thing going on, where people who eventually search for soft pillow, might actually have a neck problem and discover that they need a firmer pillow?

 

[Ryuzaki]

@bernard, I've seen several cases with different causes.

Sometimes you have pages for "soft pillow", "firm pillow", and "medium pillow" but one of them has way more links, thus Google thinks it's a more important page to share. This usually gets fixed by getting some targeted anchor texts to the other pages for their phrases.

I've seen where people will write about firm pillows and compare them to soft pillows, thereby confusing Google to a degree. Combine this with the scenario above and you can have this issue for sure.

It could also be a case where Google is struggling with the "adjective + noun" understanding in their natural language processing, especially where there's no entities involved (like proper nouns). If for some reason they can't get a good fix on the adjective part, they over focus on the noun part.

It's typically a case by case basis you have to diagnose though. It's hard to throw out all of the possible scenarios off the top of my head.

 

[bernard]

It's definitely rare, I think you're correct about language processing here, it's not in English either.

The much more common, and infuriating case, is when Google prefers the "soft pillow" page for "pillows" keyword instead of the pillows pillar content.

 

[Ryuzaki]

The much more common, and infuriating case, is when Google prefers the "soft pillow" page for "pillows" keyword instead of the pillows pillar content.

In my experience, this tends to be more of an issue with matching the intent of the SERP. One might assume a keyword like "pillows" needs a Wikipedia style article explaining the history and development of pillows, etc. But really, intuitively without even looking I can tell you in the USA anyways that since that's a short-tail product, it's going to be broad eCommerce results like Amazon, Bed Bath & Beyond, Target, Wayfair, etc.

So if you write an encyclopedia style post, but you have an ecommerce style result for "soft pillows" or a comparison or review style post for "best soft pillows" then I can see why Google chooses those over an informational piece of content. The intent is closer.

When you've done everything right, however, and they just flat out choose the wrong page, I'd go back to the "which one has more external and internal links" and "is the soft pillows better optimized for 'pillows' than the pillar page by accident" scenarios.