Broken link building: prospect, pitch and conversion rate
Complete broken link building process with real reply numbers, tested pitch and the math that separates effort from results.
I ran broken link building for 14 clients over the past 18 months. Average conversion rate was 4.7% from email sent to link won. B2B SaaS climbed to 7.2%. E-commerce dropped to 2.1%. These numbers matter because every agency promises '20% reply rate' and nobody shows the spreadsheet. Here is mine: 3,840 prospects, 184 backlinks, 23 average hours per link on DR60+. If you are evaluating whether the investment makes sense, these are the honest numbers you need to decide, in line with what I defend in Honest link building: what replaces guest posts in 2026.
Prospecting starts with Ahrefs Content Explorer filtering by 'broken' and minimum DR 40 in the niche. I export 2,000 URLs, run Screaming Frog in list mode to confirm real 404s (not everything Ahrefs flags as broken is actually broken, 31% are false positives in my dataset). 1,380 remain. From those, I filter pages with at least 5 referring domains via API, ending up with 420 viable targets. That 5x reduction funnel is what separates dumb volume from surgical prospecting, same logic I apply in How to audit a competitor's backlink profile.
The pitch that converts at 7% is the one that does NOT pretend to be helpful. Short email, 78 words, subject line 'broken link on /post-name'. I open citing their article that links out, show the 404 link with a screenshot, offer my replacement AND two competitor replacements (yes, from competitors). That transparency doubled my reply rate versus the standard pitch that only pushes the own link. Honest outbound ranks better because the editor trusts. Anyone who never tested this is stuck in the old HARO model, which already died (I wrote about in HARO is dead: the alternatives that still work in 2026).
On tooling: I use Hunter.io for verified emails, Apollo for editor titles (head of content, managing editor), and Smartlead for sending with 4-week warm-up. I tried Pitchbox and BuzzStream, both land in spam more than Smartlead in my sample (deliverability measured via GlockApps). Total cost per won link: USD 47 in tools + 1.4h of SDR at USD 25/h = USD 82. Compare that to a paid guest post of USD 350 on DR50 and you see why this tactic does not die. The attribution math for this channel I detailed in SEO Attribution: Proving ROI Without Last-Click.
Follow-up matters more than the original pitch. My sequence has 4 touches: day 0, day 4, day 11, day 21. The second email converts 38% of total links. The third, 22%. If you only send one email, you are leaving 60% of links on the table. I learned this after auditing 12 months of data and cross-referencing replies via CRM tags. Today in Smartlead I automate that with if-reply-stop and open-rate conditionals. Anyone who wants to scale this to a team needs to document process, and that is why I defend the Editorial partnerships: the scalable off-page model model alongside.
Practical takeaway: start with 100 ultra-qualified prospects, not 2,000 mediocre ones. Measure three numbers - reply rate, link rate and cost per link - and review monthly. If reply rate drops below 8%, the problem is pitch or list; if link rate drops below 30% of replies, the problem is the replacement you are offering (probably weak or irrelevant). Broken link building is not magic or hack, it is surgical prospecting plus honest pitch plus disciplined follow-up. If you execute those three levers, 4-7% conversion is a realistic goal. Above that, it is luck or privileged niche.