Banany są dobre!;D

nised24357 · Dołączył: 05 Gru 2023 Posty: 1

I’ll never forget the day I was sitting in one of our content labs and heard a guy say, “If I had seen this page 4 years ago, I wouldn’t have had to go to Home Depot for days.”

When he said in our audience interview - We knew we had something big for clients to invest in, not because it had the right word lengths and the right entity mappings, but because of any keyword research we heard from this guy Something you can’t even spot… Multiple trips to the hardware store are just disgusting… The better your content is, the less you’ll have to travel.

So when we decided to look at this update, I wanted Google to do a better job on this sentiment, algorithmically, and if anyone could, it was them. Boy, was I wrong Smile

Analytical method
Let's start our approach.

We selected 1,560,987 keywords and extracted the top 100 URLs for each keyword.

We looked at the following dates:

Pre-Update: August 12-14, 2022
Last updated: September 9-10, 2022
As a result, 31,550,933 URLs were analyzed among 2,703,210 domain names.

We run all of these keywords for our clients on a monthly basis, and we skew the data set based on the client's industry. We don't think this has a large enough impact on what we're seeing to warrant concern.

We rolled out useful content updates on Google in the weeks after we finished . Over time, I plan to get back to work on my hypothesis as more data continues to pour in after Google announced this change. But for now, this blog post is meant to show you some of the sights and sounds we see in our data so you can better explore your data.

Analyze survey results
Only 1 minute?

Big takeaways:

#1 - If you're not in the business of creating spammy websites with spammy links, then don't do anything.

#2 - From what we can see, out of the 2.7 million domains we pulled, very few high-quality domains were hit.

So my prediction was that something like this article in Forbes would probably hit, and boy, was I wrong.

Do you have 5 minutes? (or more)

If you want a more in-depth look at what we're seeing moving and not moving, then you can get our exported list of domains and URLs that have been hit or grown since the update. This might help you find some analysis of your own. We spent a lot of money to get all this data, might as well open it up for all of you to evaluate, and please share any findings you have with us on Twitter .

Note: I'm using the mammoth trust traffic metric as an assumption that Google isn't visiting any sites that appear to be genuine brands, which would tell me that big brands are safe for now, but should have a plan B in place in case they have some Low quality mass produced content.

Trust flow is a score between 1-100, where sites that are closely linked to trusted torrent sites will receive higher scores.

TLDR - for domains that rank in the top 100 for at least 50 keywords

34 = # of domains that have lost all top 10 rankings ( lists ) with an average trust traffic of 12.
1,526 = # of domains lost all rankings in the top 100 ( list ) with an average trust flow of 3.2
If you build a lot of content quickly and manage to get some of it into the top 10 and your domain has low trust and the content is “algorithmic”, be careful.

If you produce a lot of content quickly, mostly in the top 100 but not many in the top 10, and have extremely low domain trust, it might not even be worth doing because you'll be the most likely to be eliminated.

Note: There are many reasons why a domain may lose all traffic. If your data is dropping, diagnose the traffic loss before taking action.

For those domains with 1000-5000 keywords in the top 100 and bottom 100, they lost 50-90% of their rankings.

4 = # of domains who rank in their top 10 ( list )
80 = # domains lost 50-90% in the top 100 ( list )
For sites that lose most of their rankings,

For domain names that have 1000 - 5000 keywords in the top 100 and bottom 100 and get more than 50% in the rankings.

0 = # domains grew by 50% in the top 10 rankings
Among the top 100, 101 = # domains account for more than 50% of all rankings ( list ) with an average trusted traffic of 32.4
Domains that rose in ranking had significantly higher average trust traffic than domains that fell in ranking (32.4 vs. 3.2). Domains ranking at 50% or more do not (yet) receive new visibility on the updated Page 1.

Analysis walkthrough

winner
This deck .com page increased its first page ranking by 51%. https://www.decks.com/how-to/articles/best-composite-decking-materials-options

observe:

custom picture
area of authority
There is no discernible pattern in the content
Deckbros’ Trex Deck Problems page’s ranking on page one increased from 56 to 563 keywords, a 905% increase. https://deckbros.com/trex-deck-problems/

observe:

Content structure starts with a formulaic pattern and then expands into longer paragraphs
The first page ranking of the Deck Building Cost Guide from the Homeserve page increased by 367% (from 823 keywords before the update to 3848 keywords after the launch).

https://www.homeserve.com/en-us/blog/cost-guide/build-deck/

observe:

The content structure feels like it was written by a real person and is less formulaic
loser
Decoalert’s 10x10 bridge cost page’s first page ranking dropped from 111 keywords to 32 (-71%). ( ahrefs ) Their 20X20 posts also saw a 62% drop in ranking on the first page.

https://decoalert.com/how-much-does-it-cost-to-build-a-10x10-deck/

https://decoalert.com/how-much-should-a-20x20-deck-cost-20/

Warning: Beware of URLs that have lost all rankings, we have seen moves from http to https, or new URLs with the same content due to updates or refreshes, so before sounding the alarm, look for URL boosters for keywords to come Offset URL loss.

observe:

Large directories with size ranges (may not work for a single query)
Domains with lower authority/trust may not be successful using this method
A formulaic approach to content structure (a question title followed by a certain number of words, a question title followed by a certain number of words, rinse and repeat)
May be easy to spot, may be of no value to readers
Organic page count increased from ~2k in January 2022 to ~8k in August 2022
Estimated organic traffic drop after helpful content update
Organic keyword drops on page 1 after helpful content update
The San Diego Contractor News page lost 96% of its first page ranking (in our data set), from 65 keywords before the update to 2 keywords after the update. https://sandiegocontractor.news/how-much-does-a-20x20-composite-deck-cost/ ( ahrefs ) Observation:

Huge catalog covering a range of sizes
Patterns in the structure of content that may be identified as automation content
Estimated organic traffic and page 1 keyword drops after helpful content updates
After the update, Household Advice’s comparison page ranking dropped by 95% (from 821 keywords to only 42).

https://householdadvice.net/timbertech-vs-trex/ ( ahrefs )

observe:

Big catalog
Content structure follows a formulaic pattern
Estimated organic traffic and first page visibility (number of keywords) fluctuate around various Google updates
Forbes Prediction
wrong.

See “#1 Stay in Your Lane” in my Helpful Content Update Forecast post.

To me, I think Forbes, a financial website, writing about picking the best tires can be an example of people getting out of their own lane. I was wrong. This page is still in the top 3 for best truck tire queries.

Forbes' Page 1 performance jumps 8% after update

many winners

There are few losers

What we know about useful content updates
Try tagging content
Updates will be contributed to ensuring that search results contain original content written by people, for people .

In the same way I showed some directory examples, you'll want to be keenly aware of what could be considered automation content. I think Kevin Diggle's case study is one of the best ways to test it on your own data to see how "automated" it sounds.

One of many ranking factors
Useful content classifiers are one of many signals that lead to search rankings.

Google says this update will have no impact even if it is deemed unhelpful.

“Sites classified as unhelpful content can still rank highly if there are other signals that human-centered content is helpful and relevant to the query.”

Why would this be? There are a few reasons, but if most of your content is considered "helpful," a page that might trigger the algorithm may be spared since you'll generally be writing helpful for people's questions Answer. Personally, I can see brand coming into play here as well, bigger brands send a different signal and Google can handle it well vs. launching a brand new content site with thousands of pages overnight.

Continuous learning
Classification is automated and uses milliliters

This model is still in its infancy in terms of the ultimate impact it may have on content classification and search results. We know Google has some of the smartest AI and machine learning talent on the planet, and we also know Gmail has been in beta for 5 years, which means Google can easily ship something and improve it over time (whereas Apple is obsessed with Just do it right when you release it).

Site level instead of url level
Potential site-wide impact

The ratio of content a website is classified as unhelpful to content that is helpful may not have been fully assessed. It may happen that a website has a large amount of content that is classified as useful, but over time, a large amount of "useless" content may affect the content that is currently performing.

3 things you can do today
Develop a hypothesis about “useless content” – what traces of useless content will it leave behind, high bounce rates, 1 page exits, etc., whatever it is, develop one and monitor your content to see if you feel like you’re off balance.

I said it above and I'll say it again, another way to talk to your customers is to use a tool like HotJar to put a question at the end of your content asking customers if they can rate how helpful it was. (Warning, when we do this on our site we get very few responses, but some insights are better than 0).

[Tip]
Learn how to optimize for people: Why user research is important for marketers
Be really good at understanding how your customers’ language has changed. This article was written 6 years ago but the ideas still apply. Ngram Reviews. Understand the Pain Points and Write to Them: How to Write Better Ad Copy Using Review Ngrams.

Article Source By : Database Sale