Updated Disavow File by Mistake-Traffic Soars!

[McMohan]

22nd Nov - Disavow file overwritten by mistake. Some domains previously disavowed, re-avowed as a result. The error went unnoticed until over 10 days.
29th Nov - Organic traffic from Google increases by over 100 percent and currently has highest ever traffic.

Questions I am asking -
1. Should I again update the disavow file to include the domains that were removed as a result of the error? Those domains were not obvious spam, but a few links from some high PR directories, which were purchased.
2. Is the increase in traffic a result of re-avowing those domains. Isn't a week's time too small for it? Or is traffic increased due to some factor totally unrelated?
3. Should I leave those domains re-avowed, since disavowing-reavowing-disavowing frequently may send wrong signals to Google?

Thanks!

[Jez123]

Hmm, tricky! It's hard to say as you may well be fine and not have any issues - or the next Penguin update might have issues with the recently disavowed links. I suppose it depends WHY they were disavowed in the first place. Google do say that most people do not and should not disavow.

[McMohan]

Jez123, Those links were disavowed around April this year, when ranks suddenly went south. There were a couple of directory listings that listed the URL on their homepage. They may or may not have been the cause for rank drop, which didn't coincide with any Penguin update. But chances are they might be mistaken for paid text links. Ranks improved in August, better than previous levels, but a lot of activities were done that may explain rank improvement.

[Jez123]

I think the links sound OK. I would be tempted to leave it as it is. I have a feeling that nothing is to do with the disavow. I think that disavow, re-avow, disavow won't work well in your favour. Just a feeling... You can always disavow at a later date but I think I'd leave it for now.

[dipper]

are you 100% sure that your problem of "when ranks suddenly went south" was related to the inbound links to your website? .. if not then you possibly disavowed something which was not the cause of the problem! how did you confirm that it was link based problem?

[lucy24]

are you 100% sure that your problem of "when ranks suddenly went south" was related to the inbound links to your website?

He isn't:

They may or may not have been the cause for rank drop

And, by obvious corollary, the de-disavowed links may or may not have had anything to do with the rank jump.

[dipper]

lucy24 - quick to criticise someone trying to help. Why not enlighten us with your wisdom and answer the OP question.

[nakkers]

I wouldn't touch it and just see what happens.

[tangor]

@McMohan...

Time frame is too short to say this change is based on the file rewrite. I'd do nothing for at least 30 days. Something else might have changed... heck, you might have found new love in the algo!

[lucy24]

quick to criticise

Huh? I was answering a question.

[Robert Charlton]

Mod's note: Just to keep this from getting personal... I think that lucy24's answer was precise and logical, and not criticizing anybody.

[McMohan]

dipper, No, I am not sure if those links were the reason why the ranks tanked. I was only taking all possible precautions leaving nothing to chance. But, even if I did disavow useful links, could ranks go up within just a week after re-avowing them?

tangor, that is what I planning to do. But, is there any evidence to believe you shouldn't re-avow and disavow links in quick succession?

[Wilburforce]

is there any evidence to believe you shouldn't re-avow and disavow links in quick succession?

I don't know of any evidence to that effect (and one would hope that Google might allow you to retract a mistake), but I wouldn't personally make backwards-and-forwards changes any more often than I could help it.

When the disavow tool was first launched the timescale of its effects was discussed on a couple of Google forums (I can't now locate them), and as they depend on recrawl you shouldn't expect (and I wouldn't attribute) any rapid change to have come about from using it. However, high-status sites are likely to be recrawled more frequently, so it is possible.

On the effect of reavowing (i.e. removing the sites/pages from the disavow file) see [searchenginewatch.com ]. Here, additional time-lag is deliberate.

[McMohan]

On the effect of reavowing (i.e. removing the sites/pages from the disavow file) see [searchenginewatch.com ]. Here, additional time-lag is deliberate.

Thanks for the link. I am pasting a few sentences if it helps the thread here -

Cutts explained that reavowing a link can "take a lot longer," though no one knows how long that is. Google wants to be really certain that spammers are not going to try to figure out which links are helping or hurting them by doing disavow and reavow experiments

But Matt says that the link may not be given the same weight it once had. Again, this is likely a measure to stop people from gaming the system.

So, the spike in traffic in a weak after reavowing a few domains might only be a coincidence, if we take what Matt says at a face value. If that argument holds, then the case is stronger for again disavowing those domains to play it safe than sorry, contrary to what many suggest here, i.e. to let those domains remain reavowed for sometime to see what happens.

[Andy Langton]

This seems to coincide with an update at Google (occurred around 18th/19th in the UK, affecting numerous sites), so it may not be related to the disavow. Causation is always tricky with this sort of thing!

[Storiale]

I think all of us have been doing this long enough to know that there is no invisible lever puller in the sky altering things.

We look for 3 things:

1. Did my site change (including inbound links)
2. Did the Algo Change (for major changes, it would be all over the internet)
3. The reason SERPS changed was because of either 1 or 2. .

So, if you made a change like adjusting a disavow file and you noticed major changes in the SERPS and there is not evidence that the Algo changed, it is likely because you adjusted your disavow file. Pay attention after 30 days to ensure there are not negative effects after Google notices non-conforming inbound links, but either the site changes or the algo changes, that is all.

The fact that Google doesn't like it when people disavow, then re-avow to find the SERPS that are helping them, clearly indicates that re-avowing links could be a powerful act. You may have chosen the right links to re-avow. Keep a close eye and enjoy the rewards.

[netmeg]

I'd leave things alone until I determined whether or not the movement was stable. People tend to use link disavowal like a sledgehammer instead of a scalpel.

[glakes]

Has nothing to do with the disavow file. All serp reporting tools show major changes on the same day your traffic soared.

[Robert Charlton]

All serp reporting tools show major changes on the same day your traffic soared.

glakes, thanks. I feel that this was an extremely helpful observation, but it needed some digging out the details.

In this thread, that Nov 29 date of McMohan's report was a close-enough match for me to the Nov 19 and Nov 28 dates of Glenn Gabe's study, to suggest a possible connection. Not necessarily cause and effect yet... but I'm hoping this might prompt McMohan to take a really careful look at the factors suggested in the Glenn Gabe article to see if he can match or exclude them from what he sees in his own site, good or bad, and give us some feedback about whether there might be a connection between his Nov 29 move up and Glenn's observations.

Nov 2015 Phantom updates: user-engagement factors + Panda
Nov 19 and Nov 28
https://www.webmasterworld.com/google/4780955.htm [webmasterworld.com]

McMohan's comments earlier in this thread...

29th Nov - Organic traffic from Google increases by over 100 percent and currently has highest ever traffic.

[McMohan]

give us some feedback about whether there might be a connection between his Nov 29 move up and Glenn's observations

Thanks for pointing me in the direction on Glenn's article. I don't see any major changes to traffic on the 19th, but 28/29 yes. Per Glenn's article, 28th update was a reinforcing of Phantom update of 19th. I don't what to make of a scenario where noticeable change was only on 28/29.

[MarieHaynes]

Barry wrote about this post this morning, so I thought I'd jump in and give my thoughts.

You mentioned that you disavowed a couple of bad directory links in April and that rankings improved in August. These dates don't align at all with Penguin. I am guessing that the improvement you saw in August was not due to the disavow/reavow work that you did. There are other algorithms aside from Penguin that use links in their calculations, but there really wasn't any widespread link related activity at that time.

When you filed your disavow in April, did you see a further drop in rankings in the week that followed? Here is why I'm asking...Let's say that you disavowed a couple of high pagerank links and for whatever reason Google hadn't already algorithmically determined that these were self made directory links. There's a possibility that if this happened then reavowing the link could actually have a benefit. But, this would only be true if your rankings dropped after you disavowed them. In other words, if Google was counting these as good links, you'd disavow and rankings would drop, then when you reavowed, you could possibly see an improvement.

But, I think that this is unlikely because reavowing is supposed to have a lag built in. I have seen a lot of sites reavow links (usually because a previous SEO company had disavowed some good links) and I can't recall seeing any that have seen an improvement that seemed to immediately corresponded to the reavow.

I would definitely reinstate the disavow. Here are my reasons as to why:

1) If the ranking improvement was due to something else (i.e. perhaps an uptick due to Phantom/Quality) then, having these links active could harm you when Penguin finally updates.

2) Even if the reavow actually was the cause of the ranking jump, if these are paid directory links in directories that only existed for SEO purposes, then it's possible that the current version of Penguin is not catching these as unnatural links but a new iteration might. *If* your rankings are currently being boosted by these links, it likely won't be for long. And, with a new real-time Penguin running soon, if you see a drop in the future because Penguin figured out that these are unnatural links, you are not likely to know that this was a Penguin hit. As such, your site may be forever suppressed because of these unnatural links.

[McMohan]

When you filed your disavow in April, did you see a further drop in rankings in the week that followed?

No. There wasn't any drop in ranks after filling disavow file back in April. However I wouldn't count on it since it isn't likely that disavow file will be processed and ranks adjusted before the next Penguin update?

I would definitely reinstate the disavow. Here are my reasons as to why

My thoughts exactly. If these links ever prove to be bad, it will be a pain recovering the site from there. I think I will play it safe than be sorry.

[MarieHaynes]

However I wouldn't count on it since it isn't likely that disavow file will be processed and ranks adjusted before the next Penguin update?

The idea that the disavow file is not processed until Penguin runs was a theory that was circulating but John Mueller from Google has said a few times that as soon as you file your disavow, Google uses it as they crawl the web. So, if you had domain:example.com in your disavow file, the next time they crawl example.com they would stop counting the links pointing to your site and treat them as nofollowed links.

If these links that you disavowed were highly authoritative ones then they likely would have been recrawled within a week or so and you should have noticed a loss of link equity from them.

But let's say for argument's sake that there was a Penguin update in April (There wasn't but let's pretend there was) and that that Penguin update determined that those directory links were bad ones. If there were enough unnatural links in the profile that Penguin did not like, then Penguin would act like an anchor pulling your site down and keeping you from ranking at your full potential. If this happened in April, then yeah, disavowing those links after the drop would not have caused a further drop because you'd already be suppressed by Penguin.

But, if this theory were true, then reavowing those links would not cause an improvement in November as there was no Penguin update. You would have needed a Penguin update to happen in order to get the Penguin anchor that is suppressing your rankings off.

So, the only way that the drops and rises in traffic could be due to disavowing/reavowing would be if this were something outside of Penguin. For example let's say that hypothetically I had great links from say, the BBC and CNN and a few other highly authoritative sites and I disavowed them. When I disavow them, most likely the next time Google visits those sites they'll stop flowing PageRank from those sites to mine. And, within a short time period my rankings would probably drop. Theoretically if I removed those domains from my disavow file then at some point I'd regain the link equity and I'd start to rank better. I say, theoretically though because I don't think I've seen a case where that has happened. I feel like it really is true that when a link is reavowed it has much less link equity than it did in the past.

My point is that the only way for us to see that reavowing a handful of links could make an improvement like this outside of a Penguin run would be if disavowing them had caused a drop.

[keyplyr]

I posted about removing almost all the disavow sites/pages a while ago. I gained back ranking almost immediately, within a week. I did leave a half-dozen obvious spammy links.

I, like others, reacted with a knee-jerk response to Penguin fueled by posts here at WW and other webmaster resources by amassing a significant list to disavow. A few days after uploading, I saw a mild recovery from the 20 SERP position loss by Penguin. But since I also reduced the number of keywords/keyphrases stuffed in page content and internal links and a few other things, I really couldn't tell exactly what was responsible for the mild recovery.

Then approx 6 months later I read about someone removing most of their disavow file and regaining ranking, so I tried it also. Yup, worked for me as well. I can conclude one or two possible reasons. Google removed the algo filters or lessened the impact - or - the mild recovery I saw after first uploading my disavow file was due to my other actions and not entirely the result of the disavow file. By later removing most of the file from GWT, I put back a lot of the link-juice.

[McMohan]

My point is that the only way for us to see that reavowing a handful of links could make an improvement like this outside of a Penguin run would be if disavowing them had caused a drop

You have put it succinctly MarieHaynes. I have added those domains back in disavow file. There is no way for me to know for sure if the rank improvement was due to reavowing them, but I would rather live with reduced traffic at the expense of those disavowed domains than having to go through a Penguin filter.

I gained back ranking almost immediately, within a week

keyplyr, did it by any chance coincide with any algo update?

[fathom]

You have put it succinctly MarieHaynes. I have added those domains back in disavow file. There is no way for me to know for sure if the rank improvement was due to reavowing them, but I would rather live with reduced traffic at the expense of those disavowed domains than having to go through a Penguin filter.

Who created the links? That would be the first question.

Why do you believe they deserve to be disavowed? Would be the second question.

The third question is why didn't you just ask for the nofollow micro-expression to be added to the link.

The cause & effect of disavowing links prevents PageRank from flowing.

Ignoring PENGUIN for the moment since it did not run... You prevented PageRank from flowing, then you allowed it to flow and now turning it back off again.

The quote "I'd rather be safe than sorry," is a joke. Are the links indeed UNNATURAL and how do you define what an UNNATURAL LINK is?

[keyplyr]

keyplyr, did it by any chance coincide with any algo update?

No, not to my recollection although it is my understanding that the algo is always in flux and that it is only an event to us as webmasters when there appears to be a significant noticeable change.

[fathom]

Followup

Using Matt Cutts own advice for making an UNNATURAL LINK more NATURAL

[moz.com...]

"The factors that make things safer are the commonsense things you'd expect, e.g. adding a nofollow will eliminate the linking issue completely. Short of that, keyword rich anchortext is higher risk than navigational anchortext like a person or site's name, and so on."

KEYWORDY LINK ANCHORS makes things risky while a username, site name, actual URL, etc. are very low risk.

The reputation of the domain is commonly an easy thing for the Webspam Team to gauge put PENGUIN (to date) has had difficulty to determine (which is commonly why PBNs tend the get Pure Spam Manual Reviews).

Therefore if you have any power to edit the link anchors (especially before PENGUIN ReRUNs) you can likely keep the PageRank (and probably the results) with only minor internal changes.

But it all depends on how you define an UNNATURAL LINK & the schemes you may have participated in.

[Robert Charlton]

I don't what to make of a scenario where noticeable change was only on 28/29.

The Glenn Gabe article isn't detailed enough to know whether the Phantom ranking changes on Nov 28/29 occurred only with sites that had also changed in earlier Phantom updates, and circumstances of individual sites include many variables.

It's quite possible, though, that the Nov 28/29 Phantom adjustment, even by itself, could have affected sites which had been making onsite usability improvements of the kind that Google is rewarding. If Google didn't catch some intended things on Nov 19, it makes sense that the Nov 28/29 tremor was a bit of tuning designed to fix this. That's completely consistent with how Google's trial and error (aka "heuristic") approach on Panda works.

This is why I suggested you look at the Phantom article and see what factors you could "match or exclude" with regard to your own site. Take a look also at the searchmetrics article now cited in the Phantom thread, which takes a deep look at some specific content changes that were rewarded.

It sounds like you may have been working to improve onsite factors since your disavow in April, and it's very possible that Google might have noticed the changes and given you a boost with the Nov 27/28 Phantom adjustment. I'm curious to know what kinds of improvements you have been making, as I think we're all curious to know what might have produced a 100% traffic boost. Conceivably, this traffic boost could be due to no change at all... simply algo tuning.

Marie has made some excellent arguments why the ranking changes are probably not connected with the disavow. In all possibilities there are temporal connections that might be coincidental, so we're not sure about cause and effect. I think that the largest clue might have to do with what kind of work you've been doing on your site... and I tend to think that sites which are showing ongoing improvement are most likely to get rewarded.

PS: All that said, good luck with reinstating the disavow.

[McMohan]

Who created the links? That would be the first question.
Why do you believe they deserve to be disavowed? Would be the second question.
The third question is why didn't you just ask for the nofollow micro-expression to be added to the link.

1. The links came as a legacy.
2. The links that were disavowed were largely directories and a couple on the homepages, directories made for SEO. Ranks went south a few weeks after the links appeared on the homepage. May not be cause and effect, but didn't see any value keeping those links.
3. There was no manual action on the site. I am not sure for an algorithmic filter such as Penguin, you have to prove to Google that you have tried to take the links down or add nofollow. Easier will be to add the domains in the disavow file IMHO.

Are the links indeed UNNATURAL and how do you define what an UNNATURAL LINK is?

I will not venture a definition, but those directory links are hardly natural, allowing only the sites that submit free or paid, unedited.