What the skyscraper technique is
The skyscraper technique has three steps as it is normally taught. Find a page in your subject that has attracted a lot of links. Create something better on the same subject. Then contact the sites that linked to the original and tell them yours exists.
The logic is genuinely sound, and it is worth stating why before I spend the rest of the page qualifying it. Existing links are proof of demand: somebody already decided this topic was worth citing, which removes the largest unknown in link building. The people who linked once are demonstrably willing to link to this kind of thing. And competitor link data makes the prospect list mechanical to build rather than speculative.
What has changed is not the logic but the environment. The technique was popularized more than a decade ago, taught in every introductory course since, and executed by an enormous number of people simultaneously. The editors on those prospect lists have now received the same email hundreds of times. Response rates fell accordingly, and they fell hardest for the version that required the least work. I cannot give you a figure — anyone quoting a precise decline is inventing it — but the direction is not disputed by anyone who has run the outreach.
How it works, step by step
- Find the target page properly. Not the highest link count you can find, but a page with links from sites you could plausibly reach, on a topic you have standing to write about. Run a link gap analysis across two or three competitors and look for pages that attract links repeatedly, not once.
- Read the linking pages before you write anything. This is the step that separates the two versions of the tactic. Why did each of those sites link? What sentence sits around the link? If forty of them link because the original has a chart of a specific number, then the chart is the asset, and length is irrelevant.
- Decide whether you can actually do better. Better means one of: newer data, original data, wider scope, an interactive tool, clearer explanation of something the original explains badly, or expertise the original lacks. If your honest answer is that you can only make it longer, stop here.
- Build it. Properly. This is a content project measured in weeks, not an afternoon of expanding a listicle.
- Segment the prospect list. Group linking sites by why they linked. Each group gets a different pitch, because the reason to swap differs.
- Send individually. Reference the specific page, the specific link, and the specific reason yours serves their reader better. Name what is new.
- Accept a low conversion rate and use the asset elsewhere: digital PR, journalist requests, internal linking, ongoing citation. The outreach is one distribution channel, not the only one.
What it costs in time and effort
Two costs, and people budget for the wrong one. The visible cost is the asset: research, writing, design, sometimes development. A genuinely superior resource on a competitive topic is a multi-week project, and if it involves original data collection it is longer still.
The hidden cost is the outreach, and it is larger than the asset for most teams. A prospect list of two hundred linking pages, each read, categorized and pitched individually, is days of work. The mass-emailed version exists precisely because that step is expensive — and removing it is what destroyed the technique's response rate.
There is also an opportunity cost worth naming. The same weeks spent producing a marginally better version of a page that already exists could produce something that does not exist at all. Original research, a tool, a dataset. Those earn links without needing anyone to be persuaded to swap. The skyscraper approach constrains you to a topic where the ceiling has already been established by somebody else.
When it works and when it does not
The version that still works looks like this. The topic has a real information gap — the well-linked original is outdated, thin on a specific point, paywalled, missing data, or simply wrong now. You have genuine subject expertise or genuine data. You build something that a reasonable editor would agree is better on a stated dimension. The prospect list is fifty to two hundred pages, hand-read. The pitch names the page, the link, and what changed. Conversion is in the low single digits, and that is a success.
The version that does not work is the one taught in most courses. Find the page ranking first for a keyword. Write a longer article covering the same points in the same order. Export every referring domain to a spreadsheet. Send a merged template to two thousand contacts saying you noticed they linked to an article and you have written something more comprehensive. That email is instantly recognizable to anyone who has run a website for more than a year, it is often the third one they received that week, and it converts at approximately nothing.
There is a further problem with the mass version. It builds no relationship, damages your sending reputation, and burns the same prospect list you will need for every future campaign. You do not get to pitch those editors twice.
Common mistakes
- Confusing longer with better. Word count is not a quality dimension. Editors do not link to things because they are long; they link because the thing answers something.
- Not reading why the links exist. If the original earns links for one statistic, replicate the reason, not the format.
- Pitching everyone identically. A university library page, a personal blog and a trade publication linked for different reasons and need different arguments.
- Pitching a commercial page. Nobody swaps an editorial reference for a page with a pricing table on it.
- Choosing topics you have no standing in. An ecommerce site with no medical expertise producing the definitive medical guide fools no one.
- Asking for the swap in the first sentence. The request is the last line, not the subject line.
- Treating the asset as disposable. If the outreach fails, you still own a good page. Most people abandon it instead of promoting it another way.
A worked example
A software company selling to logistics operators finds that the most linked page in its space is a five-year-old guide to a regulatory compliance process, published by a trade site, cited by hundreds of pages. The regulations changed twice since publication, and the guide has not been updated.
The wrong response is a longer guide to the same process. The right response starts with reading the linking pages. It turns out most of them link to one section — a table mapping requirements to documents — because that table is the only place the mapping exists in one view. That is the asset. Everything else on the page is filler around it.
So the company builds a current version of the table, verified against the present regulations, with a downloadable version and a short changelog explaining what moved and when. Then it segments: consultants who linked from advice pages, associations who linked from member resources, forum threads, and a handful of educational sites. Each segment gets a pitch that leads with the specific thing that is now out of date in what they currently recommend, because that is the reader's problem, not the company's.
A minority respond. Some add the new resource rather than swapping. The page continues earning links for years, because it is now the current version of a thing people were already citing. The outreach started the process; being right kept it going.
How to measure it
Measure the two halves separately, because conflating them hides which one failed.
- Outreach metrics. Prospects contacted, replies, links placed. Track reply rate as well as link rate; a decent reply rate with no links means your asset is not better, while near-zero replies mean your pitch or your list is wrong.
- Passive acquisition. Links to the asset from sites you never contacted. This is the number that tells you whether you built something genuinely good, and on successful campaigns it eventually exceeds the outreach total.
- Cost per referring domain, including build time. Compare it honestly against what the same weeks spent on original research would have produced.
- Durability. Assets that are the current version of something people cite hold links far longer than assets that were merely comprehensive. Check at twelve and twenty-four months.
The verdict
Situational, and the situation is whether you are prepared to do the expensive half. The underlying insight — that pages which already attract links reveal proven demand, and that the people who linked once will link again — remains correct and is worth building on. What has collapsed is the templated execution, and it collapsed because the technique was taught to everybody at once and the recipients learned to recognize it.
If you can genuinely improve on a well-cited resource, and you are willing to read every prospect and write every pitch, it still earns links and the asset keeps earning after the campaign ends. If your improvement is a longer version of the same article and your outreach is a mail merge, you will spend weeks to earn nothing and burn a prospect list you will want later.
Given the choice, I would spend those same weeks building something that does not already exist. Original data does not need anyone to be persuaded to swap.
Questions
Is the skyscraper technique dead?
No, but the easy version is. The idea that proven link demand is a good prospecting signal remains sound, and a genuinely better resource still earns links. What stopped working is the templated outreach: editors have received that email many times over and recognize it immediately. If you are not willing to build something meaningfully better and pitch individually, pick another tactic.
How much better does my version have to be?
Better on a dimension you can state in one sentence to a stranger. Newer data, original data, broader coverage, an interactive version, or expertise the original lacks. If the only sentence you can write is that yours is more comprehensive, that is not a dimension, that is a word count, and it will not persuade anybody to edit a page.
How many prospects should I contact?
As many as you can read properly, which for most people is fifty to two hundred. The list size is limited by the quality of the pitch, not the other way round. A hundred researched, individually written emails will outperform two thousand merged ones, and will not damage your sending reputation or exhaust a list you need for future campaigns.
What is a realistic conversion rate?
Low single digits for well-executed outreach on a genuinely better asset, and effectively zero for templated sends. I would be skeptical of anyone quoting a specific published figure, because it varies enormously by market, by asset and by who is sending. Plan the campaign so that it is still worthwhile at the low end.
Should I ask people to replace the existing link or add mine?
Asking for a replacement puts the editor in the position of judging another publisher, which many will not do. Asking them to add yours alongside is easier to agree to and often gets accepted. If the original is genuinely out of date and factually wrong now, say so plainly and let them decide what to do about it.