Link building statistics from 1,873 real outreach emails

Now three months of tracked outreach, though September is only four days old at the time of this refresh. The per-email reply rate moved from 4.4% in July to 5.6% in August — real volume growth, from 273 to 1,265 sends, without the rate falling apart, which is the more meaningful read than any single month's number. September's 6.3% is provisional: a four-day slice of a month will always move around, and shouldn't yet be read as a rate.

All four known-DR tiers clear the 20-prospect minimum. DR 60+ now leads at 14.3%, ahead of DR 20-40 at 10.4%, DR 40-60 at 8.9%, and DR 0-20 at 7.4% — a rough upward slope with Domain Rating that wasn't visible in the earlier five-week dataset, though DR 40-60 sitting below DR 20-40 keeps it from being a clean monotonic trend. With 11-41 replies per tier on 105-460 sends, this is a stronger signal than before but still short of "authority reliably predicts replies." A separate group of 12 prospects has no recorded Domain Rating and is too small to rate.

Mentiohunt scores how well a specific article fits a specific site before any outreach goes out, and that filter is still doing its job almost too well to compare against: 1,023 of 1,042 contacted prospects (98%) score 0.9+ on fit. The three middle tiers remain at zero contacted prospects, and "very low fit" has grown from 5 to 16 but is still under the 20-prospect minimum. This isn't a sample-size problem more time will fix — it's the intended effect of screening for fit before sending. There's still no genuine low-fit cohort in the platform's outreach to compare the 0.9+ tier against.

62% of replies arrive within 24 hours of the email being sent, and 81% arrive inside three days. Only 4 of 102 replies — the same count as the 7-14 day bucket — have taken longer than two weeks. The shape has held steady as the sample nearly tripled from 33 to 102 replies since the last refresh: fast responses dominate, consistent with typical cold-outreach timing.

Of 743 inbound messages received, 586 were bounces and 22 were automated out-of-office replies — both excluded here, since neither is a real reply. Of the 135 genuine messages left, 68 are logged as needing manual review, 57 as a human reply, 8 as a negative reply, and 2 more split evenly between a wrong-person contact and an unsubscribe request. Mentiohunt's migration to the locked 8-value classification (interested / paid-placement ask / link-exchange ask / free-product ask / declined / other) still hasn't run, so "needs review" remains the largest single bucket — these are the platform's current raw labels, not the final taxonomy.

Measured against everyone contacted, not everyone still in the sequence: the initial email had produced replies from 7.3% of the 1,042 contacted prospects by the time each got a first follow-up, the first follow-up lifted that to 9.2%, and the second to 9.8%. Per step, replies arrived from 7.3% of the 1,042 who got the initial email, 3.5% of the 574 who reached a first follow-up, and 2.3% of the 257 who reached a second — the same declining-but-positive pattern seen in the earlier five-week dataset, now on a sample more than twice the size.