Did you find this page useful? If so, please consider linking to it
- or use the bookmark facility at bottom right. Thank you.
author: Chris Price
originally published: 2007-08-12
last edited: 2008-04-28
Website Bans / Search Penalties
Part 1: [this page] Introduction to search engine bans/ search penalties. Effects of penalties and bans. The two main classes of ban. Part 2: Search penalties continued. Malware warnings, malware bans. Duplicate content filter. Search engine guidelines. Further resources. Part 3: Search penalty resolution by A3webtech. Advice on avoiding penalties. Part 4: Search penalties: an overview
Part 1
Website bans - Introduction
Terminology
Search Penalty definitions
Search engine aliases
Penalties from the Big 3
Why do search engines ban websites?
Effects of a search penalty
Effect of a Google ban
The two types of search engine ban
SP (search penalty) levels
Algo/ manual penalties
The 'post-mortem' penalty
Problem areas
How to avoid website bans
Introduction: website banned by search engines
At
one time or another, many sites get largely or partially banned by one
or more search engines - although these events are much more accurately
referred to as penalties. Because this is such a common phenomenon now,
we think it probably happens to around 5% of websites at some stage, in
varying measures of severity. On the other hand, what might be referred
to as a 'total' ban is extremely rare (the various definitions of bans
are examined later on). It seems, then, that this area is becoming
increasingly important in search engine optimising, and more and more
often we are asked to clear a banned website (and sometimes we have to
refuse, as our business objectives do not coincide).
This page is designed to present the fullest possible resource on search engine ban and penalty types and their effects.
Note however that, like the search engines, we do not discuss causes or
possible cures in detail, for the reasons stated at the end. In our
case, these boil down to the fact that if you don't know enough to keep
from being banned, you probably don't know enough to get your website
cleared safely, and how to run clean and stay safe in the long term,
without assistance.
The
vast majority of websites need to put in place much higher standards of
housekeeping; this can hardly be achieved at their current knowledge
levels if they have received a search penalty. There is another aspect,
which is attitude: website owners in general need to show more
willingness to read, understand and comply with search engines TOS or
guidelines. Terms of Service may vary slightly between the search
suppliers; please see the section on this lower down the page.
This sort of resource can never be described as complete, for many reasons. It will be edited and added to whenever possible.
Website ban terminology
Because
this is a fairly new subject for public debate, and also because the
search engines themselves do not communicate widely or openly in this
area, we must use untested terminology. In fact it is necessary to
develop our own if we wish to discuss the subject fully. Therefore,
please be aware that many of the terms used here are new, and not
necessarily accepted by all. When a consensus on terminology is arrived
at, we can adjust the terms to suit.
Strictly
speaking, in terms of the accuracy of definition of the word, and the
fairly moderate level of action in most cases by the search engines in
response to non-compliant actions, we should not be using the term
'ban'. A more accurate term would be something like 'search results penalty', 'search results restriction' or 'search position demotion'. This would be a better way to describe a search engine's graduated response to non-compliance, and in fact their response is generally at the tolerant end of the scale rather than the strict. However, the power of modern parlance being what it is, ban it is, and ban it will no doubt stay.Alternatives might include blocked, blacklisted, deleted.
It therefore seems more accurate in most cases to refer to these events as a Search Results Penalty or Search Penalty - or
SP - and this is the term we will use whenever convenient. There are
indications that some search engine staff use similar terms. (However,
a malware ban or penalty is a different type of reaction.) In fact we
would hardly use the term 'ban' at all, since it has alarmist
connotations, except for the fact that this page needs to be found by
people searching for information on the subject - so the word has to be
repeated frequently...
More search penalty terms definedwebsite ban used as a generic term for any kind of penalty or warning
SE, SEs search engine/s
SERP, SERPs search engine result page/s
SP search penalty
penalty an automated demotion of search position
manual adjustment a demotion allocated by human intervention
deletion a site removed from the index by human intervention
algo the algorithm that controls a search engine's index and results PR PageRank
TBPR Toolbar PageRank TOS, ToS Terms of Service - search engine guidelines SE spam a catch-all term meaning junk webpages seen in the search results
Search engine aliases
Search engines, like any business, change their main business name occasionally; or use
different names for distinct divisions of their enterprise; or have
connections to other similar businesses. Here are some connected names
and variants:
Yahoo! = YPN, Overture. Linked to: AllTheWeb, AltaVista.
MSN = Live Search, Live.com, Microsoft Publisher Program.
Ask.com = Ask Jeeves. Linked to: Lycos, Hotbot.
Google, Yahoo, and MSN search penalties
From
a financial perspective it only makes sense to consider penalties
applied by the Big 3 search engines. This is because other SEs may only
affect 1 or 2% or so of site traffic; and their reaction to
infringements of guidelines is likely to be the same as or less severe
than that of the Big 3.
Further, it makes sense to consider Google's reaction in particular for
three reasons: they are more sensitive to infringements than the
others; it is easier to measure problems and recovery; and they tend to
communicate better. This last is relative, of course, since in
practical terms they communicate about as well as the Masons. A minor
point is that, in addition, their terminology is more complete.
A
search penalty generally results in some form of SERPs demotion, and
this is closely linked to page rank. Although this measure is used by
other SEs in one form or another, in practical terms, when talking of
page rank, we are usually referring to PageRank - Google's registered
version of the term.
Why do search engines ban websites?
Simple:
because of attempts to manipulate page rank or page relevancy - or
actions perceived as attempts to do so. Page rank (or in the case of
Google, PageRank) and relevancy are everything to search results. These
factors determine both a web page's position in the SERPs, and,
obviously, its usefulness to the end-user.
A
page that has an unjustifiably high position in the search results does
no one any favours. At one end of the scale, such pages are simply less
useful than they should be; at the other end, they are completely
irrelevant. At some point on the graph they become spam.
There
are other contributory factors to search position, such as the strength
of the website the page is on - but the two core factors are relevancy
and PR. Therefore, if either is artificially enhanced, a page will be
given a position in the results to which it is not entitled. That would
mean the search results would be useless, so of course the SEs cannot
tolerate that, and must react.
To
simplify these factors, page rank is primarily based on links (their
number and importance), and relevancy is primarily based on names and
naming (which includes keywords and their number, position and type,
among other things).
As brief examples of such non-compliant actions, from a non-technical angle, consider:
1) Buying links - an attempt to artificially increase PR.
2) Selling links - complicity and assistance with such attempts.
3) Keyword spam - stuffing keywords everywhere possible to make a page (or site) appear more relevant.
4) Redirecting either searchbots or visitors - and sending them to different pages according to who they are.
Such
actions are violations of search engines' guidelines or TOS - terms of
service. Anything done to artificially 'improve' a page's ranking is
non-compliant. Adding valuable content, and other important resources
linking to you naturally, are beneficial. Improving the accessibility
and usability of a site are beneficial. Improving a site by making it
simpler, faster, and better-organised is beneficial. Improving the
marketing on a site is neutral, but improves earnings. Trying to fool
somebody somewhere is negative, and the wrong direction.
Effects of a Search Penalty
You
soon know about severe search penalties because sales and traffic fall
through the floor. A partial ban is harder to recognise, because
normally it only means that your site doesn't have the page rank it
deserves. In some cases you may be as much as three points below your
true rank; in others the site may only be one point down, and this may
seem hard to quantify as a ban.
Of
course at some position on this graph, every badly-performing website
is experiencing a measure of ban, simply because they are not complying
with search engine requirements. This means in effect that basic SEO
will fix it for many sites that are just not achieving their potential.
Google website ban
It
is important to realise that a Google ban (or that of other SEs) can
have various levels. In fact it is rare for a site to disappear
completely from a search engine's results, and not be spidered at all.
In the past, it would probably have been necessary to attack a search
engine in some way to achieve this. However, as the war between the
spammers and paid promoters versus the search engines intensifies, no
doubt manual deletions of sites will increase.
A
website might be banned by Google alone, or by another of the big three
search engines; or by all of the major search engines. Often, if you
are penalised by one, you will be penalised by all - though it is hard
to measure the effects precisely for some of the smaller search engines
unless you have a full set of historical site data.
Bans,
penalties and warnings come in two basic types: search penalties and
malware warnings. The severity of search penalties rises as follows:
- a filter
- a penalty
- a manual adjustment
- a deletion
A filter, at the lowest level of search penalty, means that an on- or
off-site factor has resulted in a negation of any positive value being
attributed to your page, or less commonly, site. Such factors might
include duplicate content. Penalties are a more severe allocation of a
negative value to pages/ sites. Severe penalties perhaps justify the
use of the term ban.
Major
bans are a serious matter that need fixing quickly, otherwise your
revenues will remain poor to almost non-existent - unless you have a
large amount of search engine independent traffic, which is unusual. A
high figure here would be 25% of traffic and/or revenue from
independent sources such as other links, meaning that even if banned
your revenue stays at 25% of the previous level rather than the much
more likely 10%. There are cases where a
site with a very large number of quality inbound links can remain
afloat even if 'hard' banned, with a traffic level of perhaps 35% of
normal.
Search engine bans are of one of two types:
1. An SP (search penalty). This
is the most common type of ban, and in effect is simply a SERPs (search
engine results pages) demotion - it demotes a website's position in the
search results. The majority of bans are of this type, and have various
levels of severity.
Since
around 240,000 websites seem to have had a malware warning ban (this
figure from stopbadware.org - see Resources at foot of page), there
must be at least a million who have been SRP banned. That statement is
made because we see clients at a ratio of about 5 to 1, and that ratio
is also confirmed (indeed amplified if anything) by search results in
general.
2. A malware* ban.
Strictly speaking this type is better termed a malware warning (or
badware warning as some of the major SEs prefer), rather than a penalty
or ban, since the site in question is often not removed or demoted. The
ban does not affect the search positions, or page rank; however, even
though a site receives its normal search result positions (for example
at Google #5), the result cannot be clicked-through to the site as per
usual - there is no live link to the site. In addition, a warning is
placed in the result that says "This site may harm your computer."
Clicking the result takes you to a warning ('interstitial') page, and
not the site. You have to copy & paste the text of the URL - they
want you know the site is risky (and not be guilty of sending you there
unknowing).
*
Malware, aka scumware or badware, is software that makes changes to a
user's system, or leaks information to third parties, without the
user's consent. A popular description might therefore be 'evil'
software. It can comprise simple applications such as trojans that
collect information; or diallers that 'phone home'; or small
applications that enable other larger ones to be installed unknown to
the user. Malware that stays resident on a server can have many
purposes, none of them beneficial to visitors; one function being to
install unwanted software invisibly on visitors' computers.
Search Penalty (SP) levels:
- A total ban - site deleted. The website disappears, and for the search engine concerned, it doesn't exist.
- A 90% SP (or hard ban) - in effect, 90% disappearance, as described in detail in the next section.
- A partial SP (or soft ban) - a 'minor' ban that results in a page rank drop, affecting SERPs position.
- A snap SP
- or 'temporary glitch' ban. An SRP that is applied rapidly and has no
obvious cause. It also resolves itself fairly quickly - at least in
search engine response terms.
- A filter
- a page has any positive search value removed, probably due to a
duplicate content issue. No negative penalty value is allocated, though.
A total ban (100% SRP) means:
- The
site is removed from the index and the search results. No pages are
indexed, no searches present the site name. Searching for 'example.com'
does not return the site.
Even
clients coming to us with a history of multiple non-compliant actions
have not managed this. We would have to regard this as unlikely* except
for active aggressors; and by this, not just excessive auto-querying is
inferred but some form of exploit / scraping their database.
This
action can only be taken manually - in other words the AI cannot make
this decision. For this reason (along with similar manual penalties) it
has the slang name of a 'hand-job'.
It
is reported that Google has taken this route with some directories,
which were essentially paid link farms, and were causing a disruption
of search results. If that were indeed the case, it could be argued
that the standard search penalty might have been more appropriate. On
the other hand, it does send a message very effectively.
* In February 2006, BMW motors of Germany were reportedly blacklisted (implying a full ban) by Google ( source: BBC).
This was apparently for cloaking and doorway paging, both of which BMW
seems to have admitted. These are usually regarded as 'black hat SEO'
techniques designed to fool the search engines, though BMW argued that
they simply resulted in better search information for customers. Google
didn't agree.
Our
research, however, seems to show that this 'ban' lasted from the 4th to
the 9th of February. If that is the case, it hardly qualifies for the
term even if the site was deleted; while surely only the most
embittered cynic would suggest that a small business might find the
process just a teensy bit longer...
There
are numerous cases of bans where a website is said to have
'disappeared' or 'been removed'. On closer inspection one finds they
have receded to #90 or thereabouts in the SERPs; which probably amounts
to the same thing. A #90 search position is equivalent to deletion in
practical terms, as there certainly won't be any traffic from search
results like that.
A 90% SP typically results in:
-
The site falls to #90 or so for all its keywords, in any search, perhaps from previous page 1 or 2 positions;
-
Pages rapidly drop out of the search engine's index;
-
Very few pages end up being indexed;
-
All pages register zero PageRank, even if previously at PR6 etc.;
- All pages except the index are greyed-out on the Google PR browser toolbar;
-
The
result is commonly a 75 - 90% traffic drop, if all the top search
engines are involved. Some sites with quality links do not suffer as
badly.
-
But: the site is still spidered regularly by the search engine/s.
This
might be referred to as a 'major' ban for example, and happens to those
with website errors of a serious nature; or to those who have been
somewhat over-adventurous with their website management; and especially
for a combination of the two.
Note
the last, seemingly anomalous, point: the site is still spidered. This
scenario mainly refers to a Google ban, since it is the easiest to
quantify. The various factors taken together tell you that you've got
problems, but they can be fixed.
Again,
it is necessary to state that there is a lot of variation in how bans
appear to affect websites - there is absolutely no fixed set of
'symptoms' even for one class of ban (or more correctly Search
Penalty).
For
example: a very large site we saw experienced a severe penalty. Most of
their pages dropped out of the index, PR went to zero for all pages,
and no search results were returned except for very minor terms that
resulted in few visits per day. All other terms, e.g. 'larger' ones,
were demoted to about #90 in the results. Only around 1% of their pages
remained in the index.
Situation
serious... After some repairs were made, pages indexed rose from a
couple of thousand to 250k, which was probably most of the
currently-indexed pages (again, this was a large site). But even though there were 250,000 pages in the index, PR and therefore search results were still zero. zxc4vbn There is no fixed set of rules that apply to ban effects - perhaps especially so in the big and complex cases, since gfd3saq there may well be a mix of human and AI (algo) factors involved with the search penalty levels.
It
is probable that the majority of total bans/ blacklist/ deletion/
'removed-from' cases are in fact a 90% SP; the effects are more or less
the same, certainly as regards traffic.
A partial SP results in:
-
A two or three-point drop in PageRank
-
A keyword search position drop - 30 places seems quite common
This
type of ban can result from a simple single error, or group of errors,
that when fixed results in the site jumping back up, for example, from
PR 0 to PR 2 or 3. The partial SP is resolved much more quickly than
the 90% SP.
It is probable that more sites are suffering one of
these partial bans than any other type. This is because the effects can
be insidious: there is no obvious complete lack of search results, they
are just poor. Together with the fact that the majority of website
owners have no idea at all of their page rank or even comparative
search performance, a partial ban simply means their sites don't
perform well, and the owners don't appreciate why. The cure is often
basic SEO, since this can fix the more serious non-optimal site factors
that caused the SP in the first place.
The
SEO consultant's job here is to tune out these handicaps and clean up
the site's operation. This is a major improvement for everyone, because
the huge numbers of sites currently experiencing a partial SP cannot be
good for anyone - the owners, the customers, and even the search
engines themselves will surely benefit. This is yet another good reason
why the apparent position of some search engines on the desirability of
site improvement consultants makes no sense at all.
Of course, there are many types of improvements, and many types of consultants.
A snap SP shows:
These
bans are applied rapidly, are inexplicable, are severe, and have no
logical cause. Their effect is usually identical to a 90% SRP, a
stringent ban. They can apparently occur even if no pages or anything
else has changed on the site for weeks (or so it is reported). The
affected website finds that traffic has dropped right off due to low
SERPs results. As no changes are obvious to the owner, resolving the
issue is difficult.
The
effect is just as if the robots.txt has been changed and all bots have
been blocked, and/or as if page metatags have had a 'noindex,nofollow'
attribute added. Indeed, it has been reported that search engine data
even showed some URLs blocked by robots.txt, although that was not
found to be the case.
These
cases generally resolve themselves over a short time period (one month)
and the page rank returns, pages return to the index, and less pages
are reported as being 'blocked by robots.txt'. It is tempting to say
that it may have been a data glitch at the SE concerned, which resolved
itself fairly quickly; or it may in a very few cases be a result of
security exploits at the website. There is no recourse since the
problem clears within a month or so, though it takes longer to fully
recover.
However:
when one of these snap bans was investigated, it was found that in fact
the server had been changed over at that time. The suspicion, then, is
that a network error resulted, which had the appearance of making the
site seem to be involved in some sort of nefarious activity (or no
longer existed). One therefore wonders if a proportion of these
'glitch' bans are in fact attributable to similar causes.
The
snap SP has the distinction of being the fastest applied, and the
quickest (of the severe bans) to be resolved. It seems that it can be
applied almost instantly. It resolves itself in a comparatively short
time, by search engine damage repair standards. No action is needed by
the site owner, and indeed this may only confuse the issue - it is
unlikely that the matter could be resolved any quicker.
The snap search penalty can be distinguished from a 90% SP by the facts that:
-
There is absolutely no clue to its cause, and the site itself may have high standards of housekeeping.
-
Improvements start to be seen at 3 or 4 weeks, with no work input.
-
A search engine may report inexplicable data such as 'URLs blocked by robots.txt', when this is clearly not the case.
Algo / manual penalties
In
some complex cases, or those where the severest penalties have been
applied, there is a likelihood that a manual adjustment has been made
and that the penalty is not simply a result of the algorithm's effects.
This can be seen where there is a difference between the search
positions for 'big' terms and those for 'small' terms.
For
instance, a penalised site may have disappeared down to #90 or
thereabouts for its big terms - let's say in the insurance area, for
'house insurance'; but still be ranked at number 3 or so for 'Butte
City home insurance', which brings in about two visitors a day.
It
would be hard to see this as purely an algo-influenced result, since
the site should be equally poorly-ranked for all terms. This is
sometimes seen where there are multiple problems to be resolved.
In
other cases, the website's most important terms may have been totally
deleted. This of course is a manual adjustment to the search results.
The 'biggest' terms that the site competes for have thus been
eliminated, in order to ensure the website cannot gain unjust search
positions for those terms; and instead of appearing at perhaps #1 to #3
for such terms, the site does not appear for them at all (at least, not
in the top 1,000, which is as far down as we are allowed to measure).
The 'post-mortem' penalty
Some
violations of search engine guidelines will raise flags that quickly
attract attention. Even so, the site owner may fix these issues before
search engineers get to look at the site. If this is the case, but the
engineers feel that the non-compliant action was of sufficient severity
to warrant it, they may still apply a penalty. For less severe cases,
or where the site appears to be run honestly apart from the single
transgression, they appear to give you the benefit of the doubt and
allow you some leeway for the one occasion.
This
type of 'after the act' penalty might be avoided, then, if you fix a
problem swiftly, and normally run clean. Good metrics are the key to
both.
Problem areas
A website and its operation can be broken down into four areas:
- The network, the host's facility, and the server
- The site and page code
- The site content
- Offsite policy
Problems
are equally likely to occur in all three of the site areas, and also in
the matter of offsite management; it would be unwise to emphasise any
one area as the main source of errors. Offsite, there is one major area
of problems: paid links, which should be avoided.
The
site owner has limited or no access to some of these areas; in others
they are totally responsible. Or, in cases where the owner has no
personal interaction with the site, his agents are responsible.
It
is advisable, as a good operational principle, to maintain as much data
on the site as possible, from as many different sources as possible,
both internal and external. In many cases where an SP needs
investigating, historical data is extremely valuable. It is also
possible that, had those metrics been available and studied in the
first place, a penalty would not have been applied to the site at all.
As
far as paid links go, you need to avoid them. Don't buy them, and don't
sell them. It is true to say that if your site is small and has a low
profile, then you will most likely get away with things that bigger and
louder offenders won't. If you stick your head above the parapet,
expect to get it shot off.
How to avoid website bans
Avoiding
a search penalty is done by running clean, operating in honest mode,
and trying to avoid technical choices that might conflict with SE
quality guidelines. For larger sites, and those with complex management
routines, the latter requirement might not always be so easy; but all
should take careful note of the first two points. Therefore, you should:
1.
Read the Terms of Service of all those search engines whose indexes you
wish to appear in. Because these guidelines are a little sketchy, you
should read as many of them as you can find, and then you will
appreciate the requirements better.
2.
Monitor your website's operation. This means you should obtain data,
and check a variety of metrics, on a regular basis. It is not someone
else's job - it is yours. Get a proper website statistics package - or
two - that allows you to fully monitor your website data and especially
errors. If you don't know the precise page errors occurred on, what
file was requested, which page was requested, and the requesting IP -
then clearly you are not complying with this very important
requirement. How does your server perform? What DNS checks have you
carried out? What have your hosts done about redirects? How is your
mailserver set up? Exactly what is the state of your domains, in your
domain registrar's account? And please don't tell me you haven't got
one and you use a host to hold them. What does your network monitoring
service say about your server? What does your security audit say? These
questions - and many, many more like them - determine if your site runs
safely or not. If you don't know the precise answers here, you are
wasting your time worrying about marketing or other issues - your
website housekeeping is abysmal and you are inviting a ban.
3.
Don't use the cheapest hosting deal you can find; and don't stick with
a host you have been using for years, just out of habit - unless they
are proven reliable and of good value. Each of those options is risky;
one because you only get what you pay for, and the other because in our
experience some of the longest-established hosts are the worst. Get
your hosting audited by someone who knows what they are looking at.
Read our SEO Hosting section for an idea of modern requirements. Just
as SEO requirements change drastically every 18 months, so do hosting
requirements, though admittedly over a longer term. A host whose
operation is essentially the same as it was in the 90s - and there are
plenty of those, especially among the big names - is a bad bet now.
Continued in Part 2 - see main menu item:
>> Banned Websites - Part 2
The weird and misspelt postscript:-
- with the strange terms this page gets found for: unban site - how to
open banned sites - how to know banned websites - how to open banned
website - how is open baned website - baned websites (nice). And hardly
any of the obvious ones you would think of, since apparently no one is
searching (yet) for 'search penalty' or 'search penalties'. No point in
being subtle then...
It
seems there is some confusion here between websites under search
penalty, and websites banned from being viewed in educational,
military, government or commercial establishments and so on - or even
whole countries. Unfortunately we cannot help with the latter, modern
life requires us to specialise.
|