Word count doesn't matter as much as user satisfaction to Google

[goodroi]

I just wrapped up another site review and the best performing pages are averaging about 300-400 words which is not alarming to me. This isn't the first time I have found this type of situation. Why do these light content pages perform so well? Because they satisfy users. For example imagine an online calculator page. Those pages don't usually have 800 or 1000 words per page. They tend to have a few hundred words and a functioning calculator and perform well.

I've actually come across sites that got penalized for boosting their word counts. In one situation they were blindly trying to make sure every page had at least 1000 words and that lead to ugly keyword stuffing, horrible usability and a total nightmare to their rankings. After we reworked the content to make it appropriate for human users, the pages rebounded.

I am not advocating people chase after short content but rather that we use common sense when trying to figure out how much content is the right amount. Using common sense, if you are writing a webpage on a complicated topic, it probably will need a big amount of content (you might even want to split it up between a few pages).

So what do you think? Why do you agree or disagree with me? What has been your experience?

[getcooking]

Thank you for this timely post. I literally just walked out of an SEO meeting with my company's top brass where I said pretty much exactly what you posted.

Don't write content for Google's sake (which is the mentality I'm trying to change at my company). Make the user happy. Answer their questions thoroughly (or satisfy whatever their need is) and do so in a user-friendly/easy to digest format. Don't skimp on content but don't draw it out either for the sake of keywords or word count.

The challenge I see most websites facing with content is that they don't take the time to understand what it is the user will want from their content (and as a result, they can't understand why their 100% unique and wonderfully written content doesn't rank in Google).

Just before I started with this company they started adding a lot of low quality, keyword stuffed, redundant content purely to try to rank. Today I showed them a revenue chart from organic-only traffic with an overlay of Google Panda/quality updates on it. It finally made more sense to them when they saw revenue drops after each Google update.

[tangor]

The content/topic will usually suggest "how many words", but when marketing, strategy, SEO and the like is given a higher priority over the content AND THE USER, mixed results will always happen.

Data point type content can be a few to perhaps a few hundred words. Social commentary might run a few thousand. Detailed manuals or ascertainment might hit 10,000. At that point the more logical question is "how many pages (if any)".

Yet, all the above means nothing if the USER walks away dissatisfied, that's the one metric which trumps everything.

[McMohan]

So to summarize what makes a page 'çool', even it has shallow "content in text" -

1. Above average time per pageview (Above average within the website compared to other pages and outside the website compared to peer websites)
2. High dwell time
3. Low incidence of pogo-sticking behavior when user clicks to the page from SERPs
4. Social media shares? I am not sure it is important and haven't seen anything definitive. perhaps varies from niche to niche.
5. Bounce rate? I don't think it is important if "Dwell Time" is in favor of the page.

So, the classical notion of Content meaning Text may be misplaced. If a slide of images or an embedded video can keep the interest of a user for say 3 minutes on the page, that slide or video may be equivalent to 800 words of content in text?

[edited by: McMohan at 7:13 am (utc) on Nov 8, 2017]

[browndog]

I've read some articles that just go on and on and on...it is the most frustrating experience. They just do it to pad out content.

I have some pretty long articles on my site and am starting to review them, do I really need to spend 1,000 words on the best type of metal for widgets when it's an article on how to repair X using a widget. I have always thought that more is better, but am starting to wonder.

I've also recently focused on providing both short and long answers on many pages. So if you really do want to know about the best type of metal for widgets, then you can, but for those skimmers, people in a rush they can skip to the info.

I don't think it's just about words, it's also readability. Really using h tags so that people can jump to the section they want. If their unicorn is up at 3 am crying, they may not want to know what widgetitis is, or how it was contracted, just what the symptoms are.

[keyplyr]

I've also recently focused on providing both short and long answers on many pages

An approach I've used from the start.

On info thick pages, I always use a very short summary in the first paragraph. Then in the subsequent content, go on to more length.

I've found this often becomes the snippet in SERP.

In rewriting many pages for voice search, those short summaries have now become short answers.

[EditorialGuy]

In other words, think like an editor, and publish for your audience. What a remarkable new concept. :-)

[seoskunk]

Perhaps at this stage goodroi would be useful to give your theories on how Google evaluates user satisfaction, then people can look for those signals before they prune good content. EG you hadn't upvoted yourself so I did it for you :)

[mack]

Google is smart enough to know when content is being created simply for Google. The calculator example was spot on, if user intent is met with user satisfaction then it's a win for the user and Google. Years ago there was always suggestions that Google was gauging user satisfaction (hitting the back button etc) but now with a large number of Google users being on Android I am pretty sure they have way better methods at their disposal. Google is getting smarter, and anyone who is still adding content that is not entirely for the end users benefit might want to reconsider their motives.

Mack.

[ByronM]

I think we're kidding ourselves if we believe content isn't created for Google. If it isn't created for Google then you get no traffic. Simple as that. You either pay to play with PPC marketing or you trade tailored content to play in SERPS... None of it is really natural anymore. We're all fighting for a piece of algorithmic pie and the worst part of it all is the pie doesn't even reflect natural human interactions anymore.

[EditorialGuy]

I think we're kidding ourselves if we believe content isn't created for Google. If it isn't created for Google then you get no traffic.

We're all entitled to our opinions. Mine is that much of the teeth-gnashing that we see here is from SEO-driven site owners who have spent the last 10 or 15 years creating content for Google and are finding that their efforts don't work as well as they did back in the day.

[sdksjdksjd]

So what do you think?

For web professionals this was obvious since 2011 - Panda Year.

[ByronM]

We're all entitled to our opinions. Mine is that much of the teeth-gnashing that we see here is from SEO-driven site owners who have spent the last 10 or 15 years creating content for Google and are finding that their efforts don't work as well as they did back in the day.

If they didn't optimize for google at all these sites wouldn't have lasted 6 months yet alone 10 to 15 years. I could do some A/B testing and updates to these old sites to recover most of them... but writing new content on new domains.. its a miserable experience these days.

[tangor]

The playing field keeps changing, as well as the so-called SEO. Is there any question why this topic keeps coming up?

[riccarbi]

In my experience, at least in Google's inscrutable mind, a 300-word nicely written article is better than a 1,000-word badly written one. A 1,000-word nicely wtitten article is better than a 300-word nicely written one.

[satchmo]

I have read recently that people are preferring long pages, rather than paginated articles. They don't want to wait for successive pages to load. My site has some fairly technical articles, and given their nature, they cannot be written as short articles. Any thoughts?

[lucy24]

My site has some fairly technical articles

How heavily illustrated are they? Can you delay loading some of the supporting files? When I divide something into multiple files it's not so much because I think a browser can't handle 2MB of HTML, as because each of those megabytes comes with another 10MB of illustrations--and why dump that on either the user's connection or my server's bandwidth?

Is there lots of good in-page navigation? At the minimum, make sure there's a Back To Top available at all times. I'm still not sure whether in-page breadcrumbs are a good idea (each h5 leads back to the previous h4, which in turn leads back to the previous h3 and so on).

[EditorialGuy]

I have read recently that people are preferring long pages, rather than paginated articles. They don't want to wait for successive pages to load. My site has some fairly technical articles, and given their nature, they cannot be written as short articles. Any thoughts?

There was an academic study done years ago (which WIRED reported on) that showed the opposite: People preferred to have articles broken up into shorter chunks (a.k.a. pages), and in user testing, they perceived a long article with multiple pages as being shorter than a somewhat shorter article that required scrolling.

High-resolution monitors and mobile (with easy swiping instead of scrolling) have obviously changed things somewhat, but there are still advantages to using multiple pages when there's a good reason for doing so. Let's say that you've written a 5,000-word article on the various types of widget-propulsion systems. Wouldn't it be reasonable to have an introduction plus a page for each of the major subtopics (gasoline propulsion systems, diesel propulsion systems, electric motors, hydrid systems, etc.) instead of making the reader scroll or swipe through 5,000 words if he isn't interested in all of the subtopics? As a bonus, the subtopic pages might rank better in search results for specific searches ("gasoline propulsion systems for widgets").

IMO, it all comes down to editorial judgment. What's likely to work best for the reader in the situation at hand?

[browndog]

I absolutely loathe content posted over multiple pages. I'd rather scroll than click, but that's just me.

If I'm writing about breeds of unicorn, I will create one page per breed.

If I am writing about unicornitis, I will cover it on one page (even if it is detailed and long). But I treak the article up into shorter sections, with links at the top for people to scroll to the relevant section, and as I have mentioned before, I ALWAYS do a 5 line summary at the start of the article for the TL/DR people.

I have recently installed a rating widget (actually, somebody on here mentioned it). It won't stay up forever, but what it is doing is giving me feedback. I felt really stung yesterday when one article was given 1 star=awful. But had to put my bruised ego aside and ask myself why that was so (and sure, it's only one person, maybe they were having a bad day or maybe it is a bad article)...and look at how to improve the article. It is actually good to give my readers a chance to rate my content, which provides me with the opportunity to work on articles that may not get good feedback.

[ByronM]

I do long (1000+) posts with table of contents that has links to the major sections. I can't stand paginated articles - the only purpose is to serve more ads.

If everything was so easy to convey in 300 words, we wouldn't have forums like this where even a single topic has two pages of comments.

Also, 300 words after going through tokenization is barely enough to understand what the on page content is about especially if other pages have more depth.

[EditorialGuy]

I I can't stand paginated articles - the only purpose is to serve more ads.

That's simply untrue. As Browndog said:

If I'm writing about breeds of unicorn, I will create one page per breed.

Assuming that there's a reasonable amount of information on each breed of uniform (i.e., enough to justify a dedicated page), this is far more reader-friendly than having a single page with 10,000 words about every breed of unicorn. It's also more searcher-friendly, since the person who's searching Google, Bing, Yandex, etc. on "wire-haired Alsatian unicorns" will be able to click directly to a page about that breed.

[satchmo]

In the case of my technical articles, they do not easily break down into separate stand-alone sections. So it's either pagination, or one long page.

[lucy24]

the person who's searching Google, Bing, Yandex, etc. on "wire-haired Alsatian unicorns" will be able to click directly to a page about that breed

A few years back, it was noted that Google, at least--I don't know about other search engines--will sometimes link to fragments. (And then, of course, they went https and stopped disclosing any information, so now there's no knowing. If I can figure out how to extract the information from piwik, I will do so.) So if the material you're looking for begins 3/4 of the way down a long page, the searcher may still be sent there directly.

[EditorialGuy]

So if the material you're looking for begins 3/4 of the way down a long page, the searcher may still be sent there directly.

I'm not seeing that in my searches, but it would be a sensible thing for a search engine to do.

In the case of my technical articles, they do not easily break down into separate stand-alone sections. So it's either pagination, or one long page.

In that case, I'd go with the long page, or at least offer a "View All" option.

[browndog]

Also, no two users are the same.

If my boss tells me his cousin has widgetitis, I might do a search, but I want one line on the topic (which is covered in my brief summary at the top), if I am diagnosed with widgetitis I want an in-depth article on what it is, how it happened, what the diagnosis will be, the treatment and the outcome.

If my friend tells me he's bought a Siberian unicorn, I'd want a quick photo and outline of the breed, if I am considering a Siberian unicorn for my family, I want to know the cost, if they're good with children, how large they get...(which I outline in single sentences at the top), and I'd also want more in-depth information, such as their personality, any health issues. Again, same topic for both searches, but completely different outcomes.

I like to think I don't write for Google, but I can't say that with 100% honesty. What my goal is right now is user satisfaction, so that they will share my content on FB, Twitter etc., because I am realising Google is not everything, and I can get lots of traffic by social shares. But that will only happen if the users are happy with the information I've provided.

[lucy24]

it would be a sensible thing for a search engine to do

After further experimentation: If you use piwik, apply this pattern to raw logs (%23 means #) for the site where your piwik files live:
%23\w+&urlref=https?%3A%2F%2Fwww.google
"urlref" is the direct referer for the current request--happily, in piwik it comes immediately after the URL in the query string--not to be confused with "_ref" which is the link that brought the visitor to your site in the first place. About 2% (1/50) of google referers point to a fragment. I didn't check for other search engines.

Dang. Now I need to figure out how to incorporate this information into my ordinary log-wrangling.

[ByronM]

That's simply untrue. As Browndog said:

From my experience and testing and *my* perception it is true. Pagination is there to sell more ads. For most Journalism, there isn't a high ranking site that paginates their page. Tech pages seem to LOVE this for hardware reviews but honestly i'd prefer longer form single page reviews with lazy loading elements.

Where pagination may "Work" is joke/image/meme sites... but then again, it only works because they absolutely need 25 ads per page to make 1 cent.

Pagination works for search, comments, blogs, forums & community pages to a degree but oddly enough some of the top plugins are infinite scroll browser plugins or scripts to fix it.

and now that i think about it.. for most people if its not page 1 of search they don't go next either.. i can't see any value of pagination.

But that's me and my experience and my readers experience and everyone i know and everyone i hear about :)

[EditorialGuy]

But that's me and my experience and my readers experience and everyone i know and everyone i hear about :)

Well, it hasn't been my experience, and our site's traffic statistics suggest that it isn't true in our case. Then again, we don't paginate to fit an ad template as so many media sites do. We paginate for editorial (a.k.a. "content structure") reasons, in much the same way that Wikipedia does. (If you look up the dog breed "widget terrier" in Wikipedia, you aren't taken to a single page with 500 or 1,000 words on each of 100+ dog breeds, you're taken to a page about widget terriers.)

[browndog]

I don't create separate articles to display more ads. It makes no sense to create one enormous page covering all the unicorn breeds. Firstly, each article is approximately 1,000 words long, secondly it wouldn't work with my page structure.

At the top of each unicorn breed I cover the basics.

Origins
Eye colour
Coat colour
Horn length
Cost

So you can get all the basic information you want from those long lines, but they're not detailed

I then provide an image slider with several images of Siberian unicorns.

I then break up each topic into paragraphs. History, appearance, temperament etc.

In my niche, there are approximately 50 breeds of 'unicorn'. Who in their right mind would read a 50,000 word article?

One non-profit medical group in the US splits articles over multiple pages with one or two lines per page. They don't display ads, and I really see no reason for doing this. I get to about the second page in and give up. It is so frustrating.

My competitor also does it, not quite to that extreme, but most of their content is spread over 2 pages and it's not that long. I don't understand why they do it.

[EditorialGuy]

Sometimes it's tricky. I've consolidated some of our older and longer multi-page articles into fewer pages, and even with internal links via bookmarks, some of the pages just feel too long to me. There used to be a rule of thumb that said an ideal page length was 2-1/2 screens long. The current ideal might be longer than that, but mightn't 10 or 20 screens be too long for easy digestibility?

It's worth remembering that scrolls gave way to books.