The skyscraper technique is the most copied link building method of the last decade, which is exactly why the naive version stopped working. Find the top-ranking page, write something longer, email everyone who linked to the original. When a few thousand people run the same play against the same pages, the recipients notice.
The underlying idea is still sound. The execution most people learned is not.
Why "longer" stopped being a reason
The original framing was that better meant more comprehensive, and comprehensive meant longer. That was a reasonable proxy in 2013. It is now the single most common failure: a 4,000-word article that covers the same ground as a 1,500-word one, padded with definitions nobody asked for.
Nobody links to a page because it is long. They link because it answers something the page they were going to link to does not.
The three versions that still work
1. The update
The incumbent ranks well and is materially out of date — screenshots from an old interface, statistics from four years ago, a workflow that has since changed. This is the strongest version because the gap is objective and you can point at it in the pitch.
2. The specificity play
The incumbent is a general guide; you write the version for one audience or one situation. "Link building" versus "link building for a local service business with no content budget." You will not outrank the general page, and you do not need to — you will own the narrower query and earn links from people who needed exactly that.
3. The format change
Same information, genuinely better structured for how it gets used — a reference table where the incumbent has prose, a calculator where it has worked examples, a decision tree where it has a list of considerations. This earns links from people who were tired of citing the original.
What to check before you commit
| Signal | Green light | Don't bother |
|---|---|---|
| Incumbent's last update | Two or more years ago | Refreshed this year |
| Who links to it | Independent sites, varied | Its own network or syndication |
| Why they link | As a citation mid-article | Sitewide or roundup mentions |
| Your genuine edge | Newer, narrower, or better formatted | "Ours is more detailed" |
| Link count | 30–150 referring domains | Under 15, or over 500 |
That last row is worth explaining. Under fifteen referring domains, the outreach list is too small to justify the content cost. Over five hundred, you are looking at a canonical resource whose linkers have no reason to switch.
The outreach is where it goes wrong
The standard skyscraper email — "I noticed you linked to X, I wrote something better, would you link to mine instead?" — asks a stranger to edit a published article to your benefit, and implicitly tells them their existing citation was poor. It converts badly and deserves to.
The version that works does not ask for a swap. It says what changed since they published, states plainly that the resource they cited is now out of date in a specific way, and offers yours as a current alternative without requiring them to remove anything. That is a genuinely useful email; the swap request is not. The structural principles are the same ones in the templates guide.
When to skip it entirely
If your only differentiator is length, skip it. If the incumbent was updated this year, skip it. If the linkers are all syndication, skip it. The content cost of a skyscraper attempt is high enough that a marginal case is usually worse value than broken link building or a competitor gap list, both of which start from an already-qualified prospect set.
Where it does fit in the wider plan is covered in the methods overview and the strategy guide.
Frequently asked questions
Does the skyscraper technique still work in 2026?
The version based on writing a longer article does not, because it has been run against the same pages thousands of times and the recipients recognise the email immediately. What still works is a genuine improvement — an update to something materially out of date, a version written for a narrower audience, or a better format for how the information gets used.
How much longer should skyscraper content be?
Length is the wrong variable. A page that is 40% shorter but current, correctly scoped and better structured will earn links that a longer one will not. If the only description of your improvement is "more comprehensive," you do not yet have a reason for anyone to link.
How do I find pages to target?
Look for well-linked pages that have not been updated in two or more years, whose links come from varied independent sites citing them mid-article rather than from syndication. Thirty to a hundred and fifty referring domains is the useful range — fewer and the outreach list is too small to repay the content cost.
What should the skyscraper outreach email say?
Not "mine is better, please swap." Say specifically what has changed since they published — the interface, the data, the recommendation — and offer your current version as an alternative without asking them to remove their existing citation. Asking someone to edit a published article to your benefit is a large request with nothing in it for them.