How long do you actually sit on a caught drop before you let it point at anything

Genuine question rather than a sales one, and I'll say up front that I have an interest here because people buy drops from me and then ask me this.

The advice I've been giving for two years is: rebuild something on it, leave it eight to twelve weeks, then link. That advice is inherited. I did not test it, the man I got it from did not test it, and I suspect it dates from about 2017.

What I actually see is people doing all sorts. Some point a money link at a caught drop the week they get it and are fine. Some sit on a name for six months, put up thirty articles, and the site never gets a single page indexed.

Depends what you're running, obviously. But I'd like to know whether the waiting period is doing anything measurable, or whether the thing that matters is entirely what you put on it and the calendar is superstition.

The waiting period is a fossil and I can tell you roughly what it's a fossil of.

Around 2013 to 2015, a dropped domain that came back up with different content kept a good chunk of its historic signal for a while and then lost it in a lump. The people who did well were the ones who got the link live before the reassessment, not after. So the advice back then was actually hurry. Somewhere between 2016 and 2018 that inverted in the folklore into "wait", and I have never seen anyone explain why.

Also worth saying: we're two months out from a helpful content update that everyone expects to be aggressive about thin rebuilds. I'd be careful about any advice given in July that isn't retested in October.

the wait exists so the seller has an excuse when it doesn't work

I have run something close to this deliberately, so let me lay out what I did and where the holes are, because the holes are large.

The setup

Between November 2022 and May 2023 I caught 24 expired domains for a client network. All had prior indexed history, all had at least one link from something with real traffic, none had a visible manual action in the archived record. I split them into three groups of eight by nothing more scientific than the order I registered them.

• Group A: rebuild live within a week, outbound link to the money site at week two.
• Group B: rebuild live within a week, outbound link at week ten.
• Group C: parked with a holding page for six weeks, then rebuild, then outbound link four weeks after that.

What I measured

Time to first indexed page on the rebuild, proportion of rebuild pages indexed at day 90, and whether the money page moved in the four weeks after the outbound link went live.

What came back

Group	Median days to first index	Pages indexed at d90	Money page moved
A	9	61%	3 of 8
B	11	58%	3 of 8
C	34	22%	1 of 8

The A/B difference is nothing. Eight domains per arm cannot distinguish 61% from 58% and I would not claim it does.

The C difference is large enough that I do believe it, and it is not evidence for waiting — it is evidence against parking. Six weeks of a holding page appears to cost you something real, and my guess is that it is the crawler learning there is nothing here and reducing how often it comes back. That is the opposite of the folk advice and I want to be clear that it surprised me.

Where this falls apart

1. No control for domain quality across groups. I assigned by registration order and that correlates with which drop lists I was watching that month.
2. "Money page moved" is four different money sites in four verticals. Useless as a metric, included because people always ask.
3. Everything above happened before the September update that is coming. If that update does what people expect to thin rebuilds, group A's advantage may become group A's exposure.

What I'd actually tell someone

Don't park. Put something real on it immediately. Whether you link at week two or week ten I genuinely cannot tell you, and on this evidence neither can anyone else.

I have not tested this on domains with more than about 40 referring domains, and I would not assume it holds at the top of the market.

Twenty-four domains to discover that a blank page teaches Google there is a blank page.

so the folk advice was backwards the whole time. lovely

Group C confounds parking with a later rebuild date. You've got two variables in one arm and you're reading the result as one.

That's correct and it's the biggest of the three holes. I should have had a fourth arm that parked for six weeks and then linked at the same absolute date as B.

What I can say is that group C's rebuilds were slower to index than A and B's from the day the rebuild went live, not just in absolute terms. That's a within-arm observation rather than a designed comparison, but it points at parking rather than at the calendar.

I'd want someone else to run the fourth arm before anyone treats this as settled.

IME the crawl-decay explanation is testable directly if you own the box, which for a rebuild you do.

Pull googlebot hits per day from the access log on a parked name for the six weeks it's parked. If the curve drops off a cliff at week two and never recovers until well after the rebuild, hexforge's guess is right and the mechanism is boring. If crawl stays flat through the parking and the rebuild still indexes slowly, it's something else and worth knowing.

Whoever parks a domain next, just keep the logs. Costs nothing.

parking a domain for six weeks to be safe. i've done that maybe nine times

Changing what I tell people, which is the outcome I wanted from this.

New version: get a real site on it inside a fortnight, don't park, and stop treating the link date as the important decision. If someone wants a number from me I'll say six weeks and be honest that the number is a comfort blanket rather than a finding.

The uncomfortable bit for me commercially is that this means the domain is doing less of the work than I've been implying and the rebuild is doing more. Which is true, and I'd rather say it than have someone spend eight hundred on a name and put four spun articles on it.