Skyscraper Technique Case Study

The skyscraper technique — find link-magnet content, build something clearly better, pitch everyone linking the original — is the most famous playbook in content-led link building, and the most quietly abandoned: everyone's read it, few report their numbers. Here are ours: one asset, one niche, the playbook run faithfully, tracked for six months. It worked — at roughly half the glamour and twice the timeline the legend implies, with the campaign's real lesson hiding in which outreach converted.

The build

Target selection per the playbook: a statistics-roundup page in our niche with 140+ referring domains, visibly stale (dead sources, old numbers) — the classic vulnerable incumbent. Our version's "better" was concrete, not cosmetic: every statistic re-verified against primary sources (a third of the incumbent's had rotted), 40% more data points organised by use-case, sourced-and-dated methodology notes per line (the credibility layer journalists check), and clean design with copy-ready citations. Build cost: ~35 hours — the technique's underquoted first invoice. The asset was, we'd argue and readers agreed, unambiguously the best page on the topic. That turned out to be necessary and radically insufficient.

The outreach, and the number everyone hides

1. The incumbent's linkers, harvested and screened: 140 domains → 96 live, real, contactable prospects (the usual third lost to dead pages and junk, per the screen).
2. Pitched (the classic "you link to X; ours is better because..." format): 96, plus follow-up.
3. Replies: 14. Links from the classic pitch: 6 (6.3%). There it is — the number the legend omits. People who linked something years ago mostly don't maintain those pages; "better exists" is a weak reason to edit an old post. The six who switched were the actively-maintained sites — resource curators and recent linkers.
4. The second wave — where it actually worked: 31 more links over six months from (a) forward pitching: writers currently covering the topic (fresh articles, roundups, newsletters) offered the asset as a source — converting at ~15%, because linking decisions being made now are influenceable; and (b) passive accumulation once the asset started ranking for its own "[topic] statistics" queries — the self-perpetuating loop where the page that ranks gets cited by everyone researching, which is the technique's true endgame.

The revised playbook this teaches

The skyscraper is an asset strategy wearing an outreach costume. The backward-looking pitch (the famous part) delivered 6 links for ~25 outreach hours; the forward-looking motion and the rank-then-accumulate loop delivered 31 for less effort — so build the asset for the loop: target a "[topic] + statistics/data/reference" query the asset can own, launch with the full promotion pass, pitch the people writing now, and treat the incumbent's linker list as one prospect source among several rather than the strategy. And maintain it — our quarterly re-verification refreshes are why it kept its rankings and kept compounding, per the maintenance contract; the incumbent we displaced lost its position exactly by skipping that.

Frequently asked questions

Six months in — total scoreboard?

37 referring domains to the asset, page-one ownership of its target statistics query, and the lagged lifts on the money pages it feeds via internal links. Cost: ~35 build + ~40 outreach/promotion hours — ~2 hours per link and improving monthly as the loop runs, versus the guest-post campaign's flat 7.8.

Why do most skyscraper attempts fail?

Three autopsy findings recur: "better" that's just longer (10% improvements don't move maintainers — ours won on verified accuracy, a differentiator with teeth); all budget in the build, none in promotion (the series pattern: unpitched assets don't get found); and quitting at the classic pitch's honest 5–8% instead of running the forward motion where the yield lives.

Does this still work in niches with strong incumbents?

The vulnerable-incumbent read is the go/no-go: stale, unmaintained leaders are displaceable; actively-maintained ones mostly aren't worth the frontal assault — pick a sharper sub-topic instead, per the difficulty read. And the asset-plus-distribution engine this teaches is exactly what pairs with placement muscle when you want the loop started faster (that's us).