Parasite is over and half of you are still paying rent on it

Category 2 in my country is just how newspaper works, everything is paid, nobody pretend otherwise. Anyway is interesting the policy assume oversight and payment are different thing.

Your whole framework rests on the policy text meaning what you say it means. Has anyone here had an action lifted by arguing oversight, or is that a reading?

It's a reading, and you're right to say so.

I know of two reconsiderations filed against site reputation actions. One was rejected with boilerplate. One was accepted after the host deleted the entire directory, which tells you nothing about the argument because the evidence was destroyed rather than defended.

So: nobody has tested the oversight framing in the only forum where it could be tested. I've built three categories on a sentence, and the sentence may be describing an intent rather than a rule that gets applied. Take the taxonomy as a way to think about your own exposure, not as a prediction of what a reviewer does.

Re-ran my eleven, three and a half months on, through the March and June core updates.

Host type	n	Live and stable	Fell 50%+	Gone from index
Rented subfolder, bulk	4	0	0	4
Sponsored section, editor-reviewed	3	1	2	0
Single editorial page	4	3	1	0

The fourth bulk one has now gone, which completes that row and surprises nobody.

The middle row has moved the wrong way — two of three are now down more than half, and the second one fell during the June update rather than at a manual-action-shaped moment. That is weak evidence for hexforge's algorithmic prediction arriving, and I want to stress weak.

The single-page row lost one. Worth saying that the one it lost is on a host whose sponsored section also got hit, which is the contamination question hexforge flagged in March. One case. Still one case.

n is far too small for any of this. I'm posting it because nobody else is posting anything at all.

reporting back on mine since i said i would. page held top 2 until the first week of june and then it just isn't in the top 100 anymore. host site's whole /reviews/ folder looks gutted

1400 quid for three months of good rankings. not the worst trade i've made this year honestly but old_growth_seo called it almost to the month

I said September and it was June. I'll take being wrong in that direction.

Updating my side honestly rather than defending March.

Of my four: three still hold, one lost about 60% in June. The one that lost it was the trade publication, which is the closest of the four to hexforge's category 2 — it has a partner section I wasn't paying much attention to when I placed.

So the line I drew in March is holding but it's a narrower line than I drew it. I said "one piece a quarter with a real editor" and I'd now add "on a host that doesn't also run a rental operation somewhere else on the domain", because I think that's what got me.

I'm not signing anything longer than six months from here. Which is effectively conceding the wasting-asset framing, and I'd rather say so than let this thread sit with my March post as my last word.

a broker updating his own position honestly. mark the date

Caveat: I do not run SEO as a business, I buy placements to support offers, so my exposure is different from most of this thread.

1. I stopped renting anything with a monthly invoice in December. Not because I worked out the policy but because I could not answer the question in the opening post — what am I paying for — and I have a rule about invoices I can't explain.

2. What I moved the budget to is boring: the same money into three sites I own outright. Slower, obviously. But there is no month where somebody else's compliance decision deletes my asset.

3. The thing I would say to anyone still deciding is that the rent model has a failure mode nobody in this thread has mentioned. When the host gets actioned, you do not get a refund and you do not get notice. You find out from a rank tracker. Every other business relationship you have gives you thirty days.

I would rather own something slow than rent something fast from someone whose interests stop aligning with mine the moment a reviewer looks at his site.

There's a quieter cost to renting that this thread hasn't touched.

Every placement you buy is a record, held by a broker, connecting your money site to your payment method and usually to a real name on an invoice. Brokers in this space keep everything, secure nothing, and change hands regularly.

So the map of your link buying exists, in order, with dates, in a system you don't control. That's fine if your threat model is only the search engine. It is not fine if it includes a competitor, a disgruntled ex-employee at the broker, or a subpoena in an unrelated matter.

Owning the sites, as Marta_K describes, also removes that. Worth counting.

Had a client go through this last year and the bit that surprised me was how it ended - not with a penalty but with the host just deciding it wasn't worth it.

Big regional publisher, ran a partner section for about three years, took my client's money happily. In February a new editor arrived, looked at the section, and deleted eleven hundred pages in an afternoon. No warning, no refund of the six months prepaid.

No algorithm involved at all. Somebody just didn't like it.

That's the risk nobody prices - are any of you renting from a business that could change its mind on a Tuesday?

reading all this and glad i never had the budget

Removed four broker pitches from this thread since March. If you sell placements, the marketplace section exists.

Coming back to this after the August spam update, because the June data plus August changes what I'd say and I don't want the March post standing as my position.

What I got right: the three categories are real and they are enforced differently.

What I got wrong, or at least too confidently: I said category 3 was "squarely out" of the policy. MetricsOverMagic has a category 3 page down on a contaminated host, byte_broker has one down on a host with a partner section, and both losses happened at update boundaries rather than at manual-action moments. Two cases is not a pattern. But my March claim was a strong claim and two cases against it is enough to weaken it.

The revision I'd make is that the unit being scored may not be the page or the site but the section, and that a host running a rental operation anywhere on the domain is a worse neighbourhood than a host that isn't, regardless of which category your own page falls into. That means the question to ask before placing is not "is this a real editorial page" but "what else is on this domain and who put it there".

That's a much harder question to answer, and it costs an hour per prospect rather than a glance at a metrics tool.

Still unsure about: whether any of this survives the next core update, and whether the June movement was the algorithmic enforcement I predicted or ordinary volatility that we're all reading into because we expected it. I lean the first. I would not bet much on it.

That's the second time you've revised on two cases. At what point does the framework just not have enough behind it?