B2B Thought Leadership: What the Edelman-LinkedIn Study Says Buyers Read

In the Edelman-LinkedIn survey fielded in late 2019, only 15% of business decision-makers rated most of the thought leadership they read as very good or excellent. Six years and three editions later, they read more of it, trust it more than your product sheets, and still find most of it forgettable. B2B thought leadership works, by the study's own numbers, and most of what gets published under the label does not.

The short answer. The Edelman-LinkedIn study finds that B2B buyers spend real time on thought leadership, trust it more than marketing collateral, and use it to build shortlists. They reward content that tells them something they did not know or challenges their view. They punish content that is generic, and that punishment can cost you the deal.

What the Edelman-LinkedIn thought leadership study actually measures

The Edelman-LinkedIn B2B Thought Leadership Impact Report is an annual survey run by Edelman's business marketing practice with LinkedIn. The 2025 edition was described in coverage as the seventh annual collaboration. It asks business decision-makers how they consume thought leadership and what it does to their buying behaviour.

Thought leadership, in the study's sense, is content that offers expertise or a point of view on a topic, rather than describing a product. That covers essays, research reports, webinars and executive posts. Your product page is not thought leadership. Your analysis of why a category is shifting is.

Who gets surveyed, and how it changed

The respondents are LinkedIn members at management level. That matters, because every finding here is a statement about people who already use LinkedIn professionally. The sample and the framing have both moved between editions, which is the first thing to know before quoting any trend.

2020 figures are from the study deck Edelman published. 2024 and 2025 figures are from trade press reporting on the reports, because the primary report pages could not be retrieved for this post. Respondent frames differ by edition, so read across rows with care.

What the study does not measure

It measures stated behaviour. A respondent saying a piece of content led them to research a vendor is a recollection, not a tracked event. Nobody in this study was observed clicking, reading or buying. That limits what you can claim from it, and it is the reason this post treats every percentage as a direction rather than a conversion rate.

B2B thought leadership attention is rising, and quality is not keeping up

The time decision-makers give to thought leadership has gone up across the three editions examined here. The quality rating, the one number that should worry every content team, has no published improvement to show.

In the 2019 fieldwork, 48% of decision-makers spent an hour or more a week on thought leadership, with 17% spending four hours or more. The 2024 report put it at 52% of decision-makers and 54% of C-level executives, per O'Dwyer's coverage. The 2025 report reached 64% of target buyers and 63% of hidden buyers.

An hour a week is not much per person. Across a buying group of several people, it is a meaningful reading budget. Your content is competing for that hour against every other vendor and analyst in the category.

The 2019 fieldwork asked the readers who spend that hour to rate what they read. The result is the most quoted number in the series, and for good reason.

Per the study deck, 15% rated most of it very good or excellent, 57% called it good and 28% called it mediocre or worse. Coverage of the 2024 report repeats a 15% figure for the same question, but only in agency write-ups, so this post relies on the 2019 number it could verify in Edelman's own deck.

The dissatisfaction is not new and did not start with generative AI. A 2021 t3n report on that year's edition said 71% of respondents felt most of the content they consumed gave them little of substance. The quality problem predates ChatGPT by years.

Where thought leadership moves B2B deals, by the study's own numbers

This is the section most marketing teams quote, and it deserves the most care. The deal effects are real in the data. They are also stated as frequencies, such as "sometimes or more often", not as rates.

Shortlists, invitations and new research

In the 2019 fieldwork, 42% of decision-makers said thought leadership had at least sometimes led them to invite a firm to bid that was not in their original consideration set. And 48% said it had led them to award business. In the 2024 report, 75% of decision-makers and C-suite executives said a piece of thought leadership had led them to research a product they were not previously considering.

By 2025 the question had moved to the shortlist itself. 56% of target buyers and 55% of hidden buyers said they use thought leadership to evaluate potential vendors, per David Dodd's analysis of the 2025 report. That matches what we found on how the B2B shortlist now forms before sales is ever called.

The disqualification effect most teams ignore

The same 2019 deck carries a number that rarely makes it into agency pitches. 38% of decision-makers said reading a firm's thought leadership had sometimes led them not to award it business.

The 2024 edition found the same mechanism pointed at incumbents. 70% of C-suite leaders said thought leadership had at least occasionally led them to question an existing supplier. More than half said weak content could push them toward suppliers who understood their challenges better.

This is the finding we would put on the wall of every content team. Mediocre thought leadership is not neutral. It is a public sample of how your firm thinks, and buyers grade it.

The 2025 edition: writing B2B thought leadership for the hidden buyer

The 2025 report, published in mid 2025, shifted the frame from the person who owns the purchase to everyone else who can veto it. It opened with a claim that more than 40% of B2B deals stall because of misalignment inside the buying group.

Who the hidden buyers are

The report splits final decision-makers into two groups. Target buyers are experts in the product being bought. Hidden buyers sit on the decision as representatives of other functions, such as finance, operations, legal, compliance or procurement, per the definitions quoted by CustomerThink.

Hidden buyers read almost exactly as much as target buyers, by these figures. They rarely take a sales call. Content is often the only channel through which your argument reaches them before the decision meeting.

53% of both hidden and target buyers agreed that a vendor's fame matters less when its thought leadership is high quality. Edelman framed this as content levelling the field for less familiar vendors. CustomerThink notes 48% of respondents worked at companies with 200 or fewer employees, which makes this one of the few findings that speaks directly to small firms.

This is where we would put the budget if we were a challenger. You cannot outspend an incumbent on reach. You can out-think them in front of the finance director who will never sit through a demo.

Which executive content formats earn attention, and which only add volume

The survey does not rank individual formats like webinars against white papers. It does tell you which properties buyers reward, and those properties sort content into two piles quite cleanly. On the webinar question specifically, we looked at what replaced the webinar as a B2B reading habit.

The last row is a concession. Volume content can still earn search traffic. It does not earn the buyer behaviours this survey measures.

The two largest numbers in the 2025 edition both point the same way. Buyers want to learn something. 91% of hidden buyers and 81% of target buyers said good thought leadership helps them uncover needs or opportunities they had not recognised, per O'Dwyer's. That is not a request for more opinions. It is a request for analysis with a finding in it.

Our view is blunt. If your piece could be summarised as "AI is changing the industry and leaders must adapt", it belongs in the 57% that readers call good and then forget.

Short beats academic, but short is not shallow

57% of hidden buyers favour quick takeaways over deep, academic-style content. Read that alongside the 91% and the instruction is specific: put the surprising finding in the first two sentences, then show the evidence. Earlier editions pointed the same way on evidence, and the reason original research is the content moat that survives is that it makes the first sentence impossible to copy.

Evidence also has to hold up when a buyer checks it. A case study with a vague multiplier fails the same test, which is why case studies with verified outcomes outperform the testimonial format.

What AI-generated content has done to B2B thought leadership

The Edelman-LinkedIn coverage reviewed for this post does not report a question about AI-generated content. So this section joins the survey to separate evidence, and the join is ours.

The volume side of the ledger

Supply has exploded. An analysis by AI detection firm Originality.ai, first reported by Wired and covered by ReadWrite in November 2024, flagged 54% of 8,795 long-form English LinkedIn posts as likely AI-generated. That definition counts posts edited by AI, not only fully generated ones, and detector output carries error.

We examined what those posts earn on the platform in the analysis of AI-written LinkedIn posts and their engagement. The economics behind the flood, where the cost of a post fell far faster than its reach, are in the breakdown of content volume economics.

Put the two together and the conclusion is uncomfortable for anyone running a high-volume programme. Buyers in the 2025 edition said they want risks they had missed and views that challenge their own. A language model generating from a prompt tends to reproduce the consensus of its training data. That is the opposite of what 86% of hidden buyers asked for.

We do not think AI-assisted writing is the problem. A team that uses a model to draft around original data, then edits for a stance, can meet every bar in this survey. The problem is AI used to produce the opinion itself, because the opinion is the part buyers are grading.

There is also a second-order effect. When most of a feed is fluent and generic, fluency stops signalling effort. A named author with a specific, checkable claim becomes the scarce signal, which is why the 65% preference for a human tone may matter more in 2026 than when it was measured.

Where this argument is weakest

Most write-ups of this study skip this section. Here it is, because the evidence base has real problems.

The sponsors sell the thing being measured

LinkedIn sells the advertising that distributes thought leadership. Edelman sells the services that produce it. Neither fact makes the data wrong, and the survey has run with consistent methods for years. It does mean the questions are framed by organisations that benefit when thought leadership looks effective, and you should weigh the findings with that in mind.

The respondents are also all LinkedIn members. People who answer a LinkedIn survey about professional reading are likely heavier readers than the average executive. The hour-a-week figures probably overstate typical consumption.

Stated behaviour, shifting frames and a figure we did not use

Every deal figure here is a self-report. "Sometimes led me to award business" is a memory, not a measured conversion. The respondent frames changed between editions too, so the rise from 48% to 64% mixes a real shift with a definitional one.

Some figures do not reconcile. Coverage of the 2025 edition attributes 71% to two different findings. One outlet says 71% of hidden buyers rate thought leadership above sales materials. Another says 71% have little or no contact with sales reps. We have not used that figure. A widely repeated 2024 claim that 86% of decision-makers would invite a strong publisher into an RFP appears only in agency write-ups we could find, so it is also left out. The primary report pages could not be retrieved for this post, and every 2024 and 2025 number here is one step removed from the source.

Frequently asked questions

How to apply the Edelman-LinkedIn findings this week

Pull the last ten pieces your team published. For each, write the one thing a finance or legal reader would learn from it that they could not have guessed. Any piece where that line is empty is part of the 57%, and it should not set the template for the next ten.

Then pick one hidden-buyer function in your typical deal, such as finance, and commission a single piece aimed at their risk, not your product. Put the finding in the first two sentences and name the author. That one piece will tell you more about this survey's claims than any benchmark will.

Related reading