If you search for how long guest posting takes, you get answers ranging from "a few days" to "six to eight weeks", none of them measured. We ordered ten guest post backlinks for our own site on Friday 28 August as phase one of a pre-registered link building pilot, and logged every timestamp. The last of the ten went live on the morning of Tuesday 1 September. That span includes a weekend and the UK bank holiday on the Monday, so it is a fair picture of guest post turnaround time in a slow week, not a best case.
The short answer: the median placement was live 5.9 hours after ordering, seven of the ten within nine hours, and the slowest took three days because the publisher went quiet over the weekend and needed a nudge. Three of our first-choice publishers never delivered at all and were replaced. Those three are the most useful part of this post.
What we ordered
Ten guest posts on sites graded 3 to 4 out of 5 by our own review, DR 40 to 72, across nine niches, each carrying one dofollow link to a rank.fast page with a branded or descriptive anchor. Five were assigned to a treatment arm, where we submit the host page through our indexing service the moment it goes live, and five to a control arm that we leave alone. The split, the anchors and the target pages were all fixed before the first order, per the protocol. At our draft rates the campaign would cost a client $2,740. Wholesale figures stay private, as they do for any SEO agency.
Guest post turnaround, placement by placement
Placements are anonymised, matching the public pilot log, so the host sites and publishers are not identifiable. Hours are order to verified live.
| Placement | DR band | Niche | Ordered | Live after | Arm |
|---|---|---|---|---|---|
| P08 | 50-59 | local news and community | Thu 28 Aug 15:05 | 7 minutes | treatment |
| P10 | 70-79 | psychology and general interest | Mon 31 Aug 13:02 | 1.4 hours | control |
| P07 | 40-49 | education technology | Thu 28 Aug 12:20 | 2.4 hours | control |
| P11 | 60-69 | developer tools | Mon 31 Aug 13:02 | 2.7 hours | treatment |
| P03 | 40-49 | local news and entertainment | Fri 29 Aug 07:47 | 5.0 hours | treatment |
| P02 | 50-59 | construction trade media | Thu 28 Aug 07:57 | 6.8 hours | treatment |
| P06 | 40-49 | entertainment media | Thu 28 Aug 12:20 | 8.9 hours | control |
| P01 | 70-79 | personal development | Mon 31 Aug 13:02 | 20 hours | treatment |
| P05 | 40-49 | fashion industry media | Thu 28 Aug 07:19 | 24 hours | control |
| P04 | 40-49 | education | Thu 28 Aug 07:57 | 71 hours | control |
The first five were ordered on Thursday 28 August, not Friday; the first order went in at 07:19 UTC. P08, P03, P01 and P10 are substitutes ordered after a first choice failed, so their order times are later than the batch they belong to. All times UTC.
Three publishers failed. That is normal, and it is the job.
Buying guest posts from a marketplace looks like shopping. It is not. Every listing is a promise by a publisher who may or may not still be paying attention, and the failure modes are quiet:
- Two publishers never responded. Active listings, current prices, no reply to the order or to two messages. One had been inactive on the marketplace for about three months, which we now check before ordering. Both orders were revoked after a four-day window that allowed for the weekend, and replaced.
- One publisher rejected the order the same day because they no longer accept guest posts, despite a live listing saying they do. A substitute of the same grade accepted within the hour and delivered about three hours later.
- One delivered the anchor one word short of the brief. The link works and sits in context, so it stands, but it is recorded as a deviation.
- One editor localised the article to Canadian English and told us the publication time in advance, then hit it. That is the kind of editorial care our grading tries to detect from the outside, and it is worth recording when it shows up.
The substitutions are why the campaign finished in under five days rather than stalling. A link building service earns its fee in exactly these moments: noticing that a listing is dead, having a vetted alternative ready, and not billing the client for the one that never happened.
Interim: are the guest post pages indexed yet?
A guest post backlink only counts once Google has indexed the page it sits on, which is the failure mode our marketplace study found wholesalers quietly admit to. The pilot's core measurement is whether submitting the host page changes that. These are the numbers as of publication, and they are interim. The protocol's verdict comes at the 30 and 90 day checks.
| Arm | Placements | Indexed so far | Time from live to indexed |
|---|---|---|---|
| Treatment (host page submitted on delivery) | 5 | 5 | 3.0 to 6.7 hours for four of them; see the P01 note |
| Control (left to the publisher) | 5 | 1 | 5.7 hours for the one; the other four not yet indexed after one to four days |
The P01 note. P01 was delivered in the morning and, through our own delay, was not submitted until the evening. Our check found it in Google one minute after that submission, so the publisher's site had almost certainly indexed it on its own during the day. It is counted as indexed but it is not evidence for the treatment. The clean treatment evidence is four placements, and the lesson for our own procedure is to submit the moment delivery is confirmed.
What this proves, and what it does not
- It proves that ten vetted guest post placements can be ordered and live within five calendar days, across a weekend and a bank holiday, provided someone is watching for failures and has substitutes ready.
- It proves that roughly a third of first-choice marketplace listings did not deliver in this batch. One batch of ten is not a rate; it is a warning to check publisher activity before you pay.
- It does not prove anything about rankings. Indexed is not ranked, and no guest post backlink comes with a ranking promise from us.
- The indexing comparison is interim. Four indexed treatment placements against one indexed control is suggestive and small. The 30 and 90 day checks, pre-registered in the protocol, are the measurement that counts, and they will be published whichever way they go.
Phase two, on a second supplier, follows when it is funded. The pilot log updates as each check runs, and the protocol post carries the changelog.