Drop in ranking after editing article

[Saffron]

The past two months I've been reviewing analytics and articles which have dropped more than 20% (from 2012, my best year) have been getting a review. I do think I have greatly improved the quality of the articles but most of them have dropped in ranking. Has anybody else edited older content and noticed this happening?

I have no plans to change them back, the quality of the articles has significantly improved (my writing is much better than it was when they were originally written). Just curious I guess, feels like I'm actually causing more harm than good, despite the fact the articles are significantly better after the updates.

[Globetrotter]

I update articles all the time. At first you will see a little drop, but after a while you will get your old traffic back. If the content is really good you will get more traffic.

If you edit a lot of articles in the same section it takes longer to recover.

[Saffron]

Sorry, that was really badly worded before. It was only 6.30am when I typed it.

Thanks for your response, it's good to know it's not just me.

I've actually switched to editing the ones with very few visitors so that if they take a hit downwards, it won't hurt me so much.

[EditorialGuy]

Like Globetrotter, I update articles regularly.

Some of our highest-traffic articles (which rank at or near the top of Google's SERPs) have been edited several times a year for a decade or longer.

[aakk9999]

If we (simplistically) split the evaluation of the article between:

- on-page signals
- backlinks signals
- and user engagement/usage signals

Then if the article sufficiently changes, Google will re-evaluate on-page signals, but at the same time it may decide to discard or devalue historical engagement/usage signals and start to collect these again.

If usage signals were positively affecting the article, then by changing the article sufficiently, these usage signals are reset/devalued, which could influence the initial drop. As the time passes the usage signals fold back in, which (in the case of improved article) can explain not just the recovery but the improvement.

I think that different pages are influenced in different ratio with these three groups of signals. I suspect that pages with a strong backlink profile are less suspectable to drop based on resetting user signals.

This can also explain the move in the oposite direction - we have heard cases where the page has been changed and improved in ranking, just to drop back or even drop back further.

[Saffron]

I am hoping this is a temporary drop. Out of the 9 I'm monitoring, 6 have dropped and 3 have risen. I've edited a lot more than that, but not watching the rest.

Hopefully it will correct itself in time and I'm not boring my visitors by increasing the amount of information on the pages.

[Planet13]

Hopefully it will correct itself in time and I'm not boring my visitors by increasing the amount of information on the pages.

Don't hope... test and measure!

You can use events in analytics to track visitor reactions to your page.

If you think your pages are too long, then maybe set up events so that analytics fires an event when a user scrolls past a certain point in the page, or spends a certain amount of time on a page.

You can go ahead and call my thinking one-dimensional, but I believe that the better your page answers a query, the better it will rank (over time).

Keep your users in mind and the rest will follow.

[Planet13]

Oh, one more thought:

How many total articles are there, and did you edit all 9 of them at pretty much the same time? or did you stagger them out over a period of weeks?

If 9 articles makes up a significant percentage of the total articles on the site, and if you did them all at once, that might have SOMETHING to do with it, too.

[JD_Toims]

Out of the 9 I'm monitoring, 6 have dropped and 3 have risen.

I'd revert one of the six that dropped to the old version to see what effect it has on the page's rankings.

Added: Based on the testing I've done and what I've seen, user metrics [time on page, page views per visit, bounce rate, etc] aren't at all what they're touted by many to be as a ranking factor, so I wouldn't concentrate on those numbers as much as figuring out what content ranks today.

[Globetrotter]

@JD_toims if it are not user metrics what else could it be?

There is some sort of relation between the user metrics and the amount of traffic you get, but there are exceptions.

What I can say is that if you have terrible user metrics (from search engine traffic) you don't get much visitors. The other way around is not always the case.

[JD_Toims]

if it are not user metrics what else could it be?

A number of things, including the relative level of co-occurring phrases used in the wording of the page.

What I can say is that if you have terrible user metrics (from search engine traffic) you don't get much visitors. The other way around is not always the case.

I would definitely keep in mind correlation is not necessarily causation -- "Everyone" says "if you increase user engagement" [manage the numbers] rankings will increase, but the numbers aren't, IME [In My Experience], the cause/effect. What's more important, again, IME, is giving people what they are looking for in a way they can find easily.

One thing to keep in mind when reviewing numbers is: Google has stated it's interpretation of it's data indicates people return to the results and keep searching even after they have found the answer to their question. Granted, Google has more data than I do, but I have yet to find a single person who actually does the preceding, even though that's what Google interprets the data it has as indicating -- The people I have spoken with stop searching when they find the answer they are looking for.

The "take away" from Google's interpretation of the data is: Search, click, return/search again becomes much more "noisy" as a signal and much more difficult to use as a direct ranking signal due to the way they interpret the data they have access to.

[Saffron]

I have around 500 articles on my website and have edited around 30 since August.

I've decided to re-focus on adding new content & not improving the weaker ones for now and will monitor their progression over the next few months.