The Death of the Gated Ebook: Why the Form Costs More Than the List It Builds

Forrester measured the process your gated ebook feeds. Fewer than 1 in 100 inquiries becomes a closed deal. Gated content is not failing at the margin, it is failing by design. The form was built for a buyer who no longer exists, and it now hides the page from the systems doing the discovery.

What the gate actually charges you

A gated asset is a trade. You hand over a document. You take a name, an email address and a company in return.

That trade was sound when the document was scarce and the identity was hard to obtain. Neither condition holds in 2026. Most gated ebooks restate material that is already public, and a buyer can assemble an equivalent answer from a model in under a minute. Meanwhile enrichment, intent data and public profiles have made the identity half of the trade close to free.

So the form now charges a real price for something you could get another way. That is the part marketing teams underweight. The cost is not the form itself. The cost is everything the form prevents.

Three costs, and only one shows up in a dashboard

The first cost is drop off. Some share of readers see the field list and leave, and you never learn who they were.

The second is quality. The people who do complete it include competitors, students, job seekers and anyone who has worked out that a disposable address still returns the PDF.

The third is invisibility, and this is the one no dashboard can show you. A page behind a form is not in the search index, not in the retrieval corpus, and not available to be quoted. It cannot lose a ranking it was never eligible for.

Only the first cost appears in a report, as a bounce rate on a landing page. The other two are absences. Absences do not show up in analytics, which is exactly why the gate survives review after review.

The conversion math nobody puts in the board deck

Forrester retired the marketing qualified lead as a planning unit in 2022. The reason was arithmetic, not fashion.

Under 1% is the benchmark, not the worst case

In April 2022 Terry Flaherty, a VP and principal analyst at Forrester, published the firm's benchmark for the lead-centric process. The typical conversion rate from inquiry to closed won is less than 1%. His summary was blunter than the number. He wrote that the process "fails more than 99% of the time; that's tragic and expensive."

That figure is a benchmark across Forrester's waterfall data. It is not one bad quarter at one company. It is the normal output of the design.

The firm's replacement, the B2B Revenue Waterfall, counts buying groups and opportunities instead of individual leads. The justification sits in the same research. Over 80% of buying decisions are made by a group of more than three people. A form captures one member of that group. It then reports that person as progress.

What 1,000 inquiries actually turns into

Take the benchmark literally for a moment. A thousand gated inquiries at the top produce fewer than ten closed deals at the bottom.

Now price it against your own numbers. Take your blended cost per gated inquiry, whatever that figure is in your system, multiply by 1,000, and divide by the deals that channel actually closed. That is your real cost per deal from the gated channel. Most teams have never written it down, which is a choice.

I disagree with the standard response to this benchmark. The usual move is to improve lead scoring. I think the number is telling you the unit is wrong, not that the scoring is bad.

The buyer moved, and the form did not

6sense surveyed nearly 4,000 B2B buyers for its 2025 Buyer Experience Report, across North America, Europe and Asia Pacific. Two findings matter for anyone still defending a form.

First, the point of first contact with a seller now sits at 61% of the buying journey. In the 2024 edition it was 69%. The journey itself also compressed, with the average cycle falling from 11.3 months to 10.1 months.

Second, and this is the finding that should settle the argument, 94% of buying groups put their shortlist in order of preference before they engage any seller. The vendor they favoured before contact went on to win 95% of the time in the 2025 data.

The shortlist is ranked before you know the account exists

Read those two facts together. Roughly 60% of the decision happens somewhere you have no form, no cookie and no record. By the time somebody completes your form, the ranking already exists, and you are usually confirming it rather than creating it. What that pre-contact ranking is built from is covered in the piece on how AI assistants assemble the B2B shortlist.

The same 6sense report found 94% of buyers using large language models during the buying process. That is where the gate does its most expensive work. You stay absent from the surface that builds the ranking, so that you can be present in a database that records the ranking after the fact.

A form is the one wall retrieval cannot climb

Gating used to cost you a share of human readers. It now costs you eligibility with the machines that summarise the category.

What Google's own documentation says

Google publishes explicit guidance for content behind a paywall or a registration wall. The instruction is not ambiguous. If you want Google to crawl and index your content, including the gated sections, you must make sure Googlebot can access the page. Google also asks for structured data marking the content as not free, because without it a crawler cannot distinguish deliberate gating from cloaking, which is a spam policy violation.

Almost no B2B marketing team does either thing. The typical gated ebook is a PDF sitting behind a form submission, with no crawler access and no markup. It is not being penalised. It was never a candidate in the first place.

Gartner forecast in February 2024 that search engine volume would fall 25% by 2026 as chatbots and virtual agents take share from search marketing. Treat that as a forecast rather than a measurement. Even at half the predicted size, it moves a meaningful slice of discovery to a surface where a form returns nothing at all.

What the crawl data says about the trade

Cloudflare measures this at network scale, and the numbers cut both ways. In August 2025 it published crawl to refer ratios for July 2025: Anthropic at 38,066 to 1, OpenAI at 1,091 to 1, Perplexity at 194.8 to 1, and Google at 5.4 to 1.

Those ratios are the standard argument for putting up a wall. The models take a great deal and send back very little. Cloudflare's earlier post on the same metric is careful to say the figures may be overstated, because referrals from Claude's native app carry no referer header.

The ratio still is not the decision. The decision is whether you appear in the answer. Cloudflare also found that 80% of AI crawling in the prior 12 months was for training, 18% for search and 2% for user actions, with the search share falling from 26% to 17% year over year. The slice that can produce a citation is small and getting smaller. Sitting outside it entirely is not a defensive position, it is a forfeit. What actually earns a citation is examined in the review of the GEO evidence against ordinary SEO.

Publishing more open pages is not automatically the fix either. Volume without a distinct number attached runs into the problem described in the work on content saturation and visibility.

Where gating still works

I would keep the gate on exactly one class of asset: anything that costs you money to deliver per person. A seat at a roundtable, an hour of an engineer's time, a bespoke assessment, a physical item. In those cases the form is a booking, not a toll.

Everything else deserves a harder look, and most of it deserves a different treatment rather than a straight deletion.

The half gate is usually the right answer

Publish the finding as a full HTML page, with the numbers in the visible text. Put the underlying dataset, the model, or the working file behind a request. The page earns the citation. The file earns the conversation. You lose nothing on either side, and you stop asking a stranger to pay before they know whether the thing is any good.

The last row is the concession. If your entire gated library looks like that row, this post does not apply to you and you should ignore it.

Where this argument is weakest

Every post arguing against gating skips this part. Here it is.

The strongest case for keeping the gate

The best counter argument is not about content at all. It is about the reporting line. A demand gen team with a lead target and a form can produce a number every Monday morning. Remove the form and that number disappears, while no replacement metric arrives for two quarters or more.

Careers end in that gap. Anyone recommending ungating without a plan for the reporting gap is giving advice that is both correct and unusable. That is why the last section of this post is a sequence rather than a slogan.

There is a real case in narrow markets too. If your addressable market is 400 accounts, list precision matters more than reach, and a form is a cheap qualifier. I would still publish the substance and gate only the extras, but the trade is genuinely closer at that size.

The numbers on my side are not clean either

The Forrester benchmark dates from 2022. The 6sense report is vendor research, published by a company that sells the alternative to forms. Neither fact makes them wrong. Both mean you should treat the direction as firmer than the decimals.

Cloudflare's ratios are a proxy measured on one network, with an overstatement risk the company states itself. And I have no controlled experiment to offer. Nobody has published a clean test of gated against ungated at scale, with the same asset, the same audience, and pipeline as the outcome. That study does not exist. If it did, this post would be much shorter.

The statistic everyone quotes and nobody dates

You will meet a claim that B2B buyers spend only 17% of their buying time with suppliers. It appears in dozens of posts, almost always undated. It traces back to a Gartner survey of 750 buyers conducted in 2017.

I do not think that figure should be quoted any more without the date attached. It is a clean example of the failure mode running through this whole debate. A striking number turns up in six places, which feels like corroboration, and all six are restating one study from a different era of buying.

What replaces the form

The honest answer is that nothing replaces it one for one. You are swapping a countable proxy for an uncountable reality, and that is a real loss on the reporting side. Anyone selling you a clean substitute is selling you a dashboard.

What does work is moving the ask after the value rather than before it. Keep a form on the page, below the content, optional. People who want the spreadsheet, the vendor questionnaire or the sample contract will give you an address once they have seen that the thing is good. The list gets smaller and the intent gets much higher.

Then move the qualification work to the pages a shortlist is actually built from. Comparison pages, pricing pages and case studies with checkable outcomes carry more weight than any ebook, and the shape of them is set out in the template for B2B comparison pages and the standard for verified case study outcomes.

The one format a model cannot obtain anywhere else is a number that only you hold. Publish it with its sample size, its time window and its exclusions in the visible text. That argument is made in full in the case for original research as a content moat.

For measurement, add one question to your demand form and your closed deal records: how did you hear about us. Self-reported attribution is unreliable at the individual record level and useful in aggregate, which is the opposite of what most attribution tools promise. The practical version of that trade is in the piece on measuring dark discovery.

Running the swap without losing your number

Do not ungate the whole library in one week. Pick five assets, hold the rest constant, and give yourself a comparison you can defend when somebody asks what happened to lead volume.

Twelve weeks is not long enough to prove a pipeline effect on a 10 month buying cycle. It is long enough to prove the visibility effect, which is the part the gate was hiding.

Frequently asked questions

Where to start

Start with an audit, not a deletion. Open your five highest-volume gated assets and find out how many closed deals each one is credited with over the past 12 months. If that number cannot be produced, you have learned the most important thing on offer here, and it took an afternoon.

Then take the single asset with the best content and the worst conversion, publish it in full as an HTML page this week, and record the date. In six weeks you will know whether anything cites it. That is a smaller commitment than the argument usually demands, and it produces evidence from your own market rather than from someone else's benchmark.

Related reading