From Access to Transcripts to Outcomes: The Three Eras of Primary Research

✦

The expert network industry was born from a simple observation: investors would pay to talk to people who actually knew things. For more than two decades, that model barely changed. What started as a phone directory for hedge funds grew into a $2.5 billion global industry, but the core product remained stubbornly the same. An analyst asks for an expert. Someone finds one. A call gets scheduled. What happens on that call is the analyst's problem.

That era is ending. And what replaces it will reshape how the entire investment industry does primary research.

Key Insights

• The origin story: GLG was founded in 1998 as a publishing house. It pivoted when it realised investors didn't want to read reports. They wanted to talk to the people who wrote them.
• The access era: For nearly twenty years, expert networks sold one product: a phone number and a time slot. The analyst did everything else.
• The transcript era: Starting around 2016, companies like Tegus and Third Bridge began recording, transcribing, and reselling expert calls, creating searchable libraries of thousands of conversations. AlphaSense's $930 million acquisition of Tegus in 2024 consolidated this model at scale.
• The limits of transcripts: A transcript is still raw material. Reading forty pages of someone else's expert call does not answer your specific research question. The format changed. The work didn't.
• The outcomes era: A new model is emerging where the provider owns the entire research process and delivers decision-grade insight, not access, not content, but answers. In days, not weeks.

Two Yale Graduates and a Publishing House

The expert network industry began, improbably, with a set of guidebooks that nobody wanted to read.

In 1998, Mark Gerson and Thomas Lehrman, both in their late twenties and fresh from Yale Law School, raised a million dollars to start a publishing company in New York. Lehrman had spent two years as an analyst at Tiger Global and understood what investors needed: reliable, sector-specific intelligence that went deeper than a broker note. The plan was to commission industry experts to write detailed guides for institutional investors.

The guides didn't sell. But something unexpected happened. The investors who received them kept calling to ask if they could speak directly to the authors. They didn't want the written analysis. They wanted the conversation.

Gerson later described the insight simply. "We thought it was kind of ridiculous that the hedge fund business got so much information by asking for favours. 'Could I please have 15 minutes of your time?' They would certainly pay for that information. And the people who have it would love to talk. We just thought there should be a way to get the two connected."

By mid-1999, the publishing business was gone. In its place was something entirely new: a subscription service that connected investors with a curated network of industry experts for paid one-hour phone consultations. They called it Gerson Lehrman Group. The expert network was born.

Alexander Saint-Amand, who joined the young company while still at college, is often credited with seeing the model's full potential. The story goes that he attended a medical conference to research his mother's Alzheimer's condition and found he learnt more from informal conversations on the margins of the event than from any of the formal presentations. It was the same insight, applied differently. The real value wasn't in the content. It was in the conversation.

The First Era: Selling Access

The business model that GLG built was elegant in its simplicity. Clients paid a subscription fee for access to a growing database of experts. When an analyst needed to understand a particular market, technology, or competitive dynamic, they submitted a request. GLG's team would search the network, identify a suitable expert, and arrange a call. The expert was paid an hourly rate, typically between $100 and $500. The client got the conversation.

The timing was fortunate. In 2000, the SEC introduced Regulation Fair Disclosure, prohibiting public companies from sharing material information selectively. Overnight, the supply of informal intelligence that investors had relied on for decades was cut off. Expert networks offered a compliant alternative: conversations with industry practitioners who could share their own knowledge and experience without crossing regulatory lines.

Growth followed the growth of hedge funds through the 2000s. By 2007, GLG had 200,000 experts in its database and reported revenues of $232 million. That year, Silver Lake Partners invested $200 million, valuing the company at $875 million. By 2010, the Wall Street Journal described GLG as dominating the US expert networking industry.

A second generation of competitors emerged from 2006 onwards, many founded by former consultants who had been clients of the first-generation networks. AlphaSights was started in 2008 by people from Bain and McKinsey. Third Bridge, originally called Cognolink, launched in 2007 with founders from Bain. Guidepoint had been operating since 2003. These newer entrants initially focused on private equity firms and strategy consultancies, segments underserved by GLG's hedge fund-centric model, and grew rapidly alongside the PE industry's decade-long expansion.

By 2024, the expert network market had grown to approximately $2.5 billion in annual revenue. GLG alone generates over $400 million. AlphaSights surpasses $300 million. Third Bridge exceeds $250 million. The industry employs thousands of people across dozens of countries and serves nearly every major hedge fund, private equity firm, and consulting house in the world.

But for all that growth, the core product had changed remarkably little from what Gerson and Lehrman stumbled upon in 1999.

The client still submitted a request. The network still found an expert. A call still got scheduled. And the analyst still sat on that call, alone, trying to extract something useful from a conversation they had prepped for themselves, with an expert they had never spoken to before, on a timeline dictated by the network's matching speed rather than their own research needs.

The access model created an industry. But it also created a problem that nobody was incentivised to fix. The expert network got paid when the call happened. Whether the analyst walked away with anything useful was, quite literally, someone else's business.

The Industry's Reckoning

The access model's limitations were exposed long before anyone built an alternative.

In the years following the 2008 financial crisis, the FBI began investigating whether expert networks were being used to pass inside information to hedge fund managers. The resulting prosecutions sent shockwaves through the industry. Several experts were found to have crossed the line from sharing their own professional knowledge to disclosing material non-public information about their employers. The investigations led to convictions and destroyed careers. GLG itself was never charged, but the reputational damage to the industry was significant.

The compliance response was dramatic. Expert networks invested heavily in screening, monitoring, and training programmes. Call recordings became standard. Compliance officers reviewed transcripts. New rules were introduced to prevent experts from discussing their current employers' material non-public information.

But the more important lesson was subtler. The insider trading scandals revealed something structural about the access model. When the product is a conversation between two people with no oversight, no moderation, and no quality control, the quality of the output depends entirely on the skill of the two people on the call. Most analysts are not trained interviewers. Most experts are not trained communicators. And nobody in the middle has any incentive to make the conversation better, because the network gets paid regardless.

The access model was a marketplace. It connected supply and demand. But it took no responsibility for what happened after the connection was made.

The Second Era: Selling Transcripts

The first serious challenge to the access model came not from improving the call itself, but from recording it.

In 2016, two brothers in Chicago launched Tegus. The insight was borrowed from an unlikely source: Costco's membership model. Rather than charging per call, Tegus offered clients a subscription to a growing library of expert call transcripts. The economics were clever. Clients who ran calls through Tegus were incentivised (and increasingly required) to allow those calls to be recorded, transcribed, and published to the library after a brief compliance review. Each call cost the client $400 to $500. But Tegus owned the transcript and could resell it to every other subscriber.

The flywheel was powerful. Subscribers read transcripts, identified follow-up questions, ran additional calls, and generated new content for the library. Meanwhile, Tegus employed its own analysts to conduct interviews on popular companies, ensuring there was always fresh material. By April 2024, the library contained over 100,000 transcripts covering thousands of public and private companies.

Third Bridge had arrived at a similar destination from a different starting point. Its Forum product offered curated expert interviews on specific companies and sectors, authored by Third Bridge's own research team. Where Tegus built a user-generated content library, Third Bridge built an editorial one. GLG launched its own Library product. Guidepoint followed. The transcript model had become the industry's second act.

In June 2024, AlphaSense acquired Tegus for $930 million, consolidating the two largest players in the transcript library market. By October 2025, AlphaSense reported $500 million in ARR, 6,500 customers, and over 200,000 transcripts. The company has been discussed as an IPO candidate.

The transcript era solved a real problem. It gave analysts a way to access expert insight without scheduling a call. It was faster than the access model. It was often cheaper. And it was scalable in a way that one-to-one calls could never be. An analyst covering a new sector could read twenty transcripts in a morning and develop a baseline understanding that would have taken weeks of calls to assemble.

But the transcript model has its own structural limitation. And it is the same limitation that defined the access model, just in a different shape.

A transcript is not an answer. It is raw material.

Reading forty pages of someone else's expert call does not answer the specific question an analyst is trying to resolve. The call was conducted by a different analyst, with different thesis drivers, at a different point in time. The most relevant insight might be buried on page thirty-seven between two unrelated tangents. Or it might not be there at all, because the original interviewer didn't ask the right question.

The transcript model changed the format. It didn't change the fundamental problem. The analyst is still doing all the work. They are still responsible for finding the signal, extracting the insight, and connecting it to their own investment thesis. The expert network, or the transcript library, has still taken no responsibility for whether the analyst actually gets an answer.

The industry moved from selling a phone number to selling a document. The work moved from the phone to the screen. But it didn't move off the analyst's desk.

The Gap Nobody Talks About

There is something the expert network industry has never been willing to say out loud: most expert calls do not produce a useful outcome.

This is not because the experts are bad. It is not because the analysts are lazy. It is because the model is designed wrong.

An expert call works when four things align. The right expert is matched to the right question. The interviewer understands the thesis well enough to probe for incremental insight. The call is moderated in a way that steers toward signal rather than generic commentary. And the output is synthesised in a format the analyst can actually use.

In the traditional access model, the network handles only the first of these. In the transcript model, a third party handled the first two, but for a different client with different questions. In neither model does anyone take responsibility for all four.

The result is an industry where the client pays for the process and hopes for the outcome. Where fill rates and match speeds are tracked religiously, but nobody measures whether the analyst actually learned something useful. Where billions of dollars change hands every year for a product whose success is defined by whether a call happened, not whether it mattered.

Every analyst who has used an expert network knows this. They have sat through calls that went nowhere. They have read transcripts that were tangentially relevant at best. They have spent hours preparing questions for an expert who turned out to be the wrong person. And they have done it all again the next day, because there was no alternative.

Until now.

The Third Era: Selling Outcomes

The third era of primary research starts with a different question. Not "can we connect you with an expert?" Not "would you like to read a transcript?" But: "what do you need to know?"

The shift sounds simple. It changes everything.

In an outcomes model, the provider owns the entire research process. The analyst gives a brief. Fifteen minutes, not fifteen hours. The provider takes that brief and handles the rest: identifying the right experts, conducting the interviews, moderating the conversations to probe for what matters, and delivering the output in a format that is ready to use. Not a transcript. Not a recording. Clear, decision-grade insights that connect directly to the analyst's thesis.

The economics change too. When the product is an outcome rather than an hour of access, the provider's incentive aligns with the client's. A call that produces nothing useful is a failure for both parties, not a billable event for one and a wasted hour for the other. The provider is motivated to make every interaction count, because their value is measured by the quality of the answer, not the volume of the activity.

Speed changes as well. The access model operates on the expert's schedule. The transcript model operates on the library's publication cycle. The outcomes model operates on the client's timeline. When an analyst needs an answer ahead of earnings, or a PE team needs to validate assumptions before an investment committee meeting, or a sell-side advisor needs diligence for a live transaction, the research arrives in days, not weeks. Because the provider controls the process end to end, they can compress timelines without asking the client to do more work.

And cost changes. This is perhaps the least intuitive part. An outcomes model can deliver more insight at lower cost than the access model, because it eliminates the hidden cost that nobody accounts for: the analyst's time. The true cost of a traditional expert call is not the $500 or $1,000 that appears on the invoice. It is that figure plus the three to five hours the analyst spent sourcing, scheduling, prepping, conducting, and extracting value from the call. When you multiply the analyst's hourly cost by the hours consumed, the real cost of a single call often exceeds several thousand pounds. The outcomes model replaces that entire process with a 15-minute brief and a set of delivered insights.

What This Means for the Industry

The expert network industry is not going to disappear. It is a $2.5 billion market for a reason. The demand for primary research among hedge funds, private equity firms, and capital markets professionals is growing, not shrinking. Diligence windows are compressing. Investment decisions are getting more complex. The need for ground-truth insight has never been greater.

But the model is going to change. The access era built the industry. The transcript era scaled it. The outcomes era will reshape it.

The firms that thrive will be the ones that stop measuring success by calls completed, transcripts produced, or experts matched. They will measure it by the only thing that matters: did the client get an answer they could use?

That is a harder business to build. It requires trained moderators who understand thesis drivers and know how to extract signal from a conversation. It requires operational systems that can compress timelines without sacrificing quality. It requires a willingness to take ownership of the output rather than just the input. And it requires aligning your economic model with your client's, so you only win when they win.

Not everyone will make that transition. The access model is profitable precisely because it avoids these responsibilities. Selling a phone number is simpler than selling an answer. But the analysts, portfolio managers, and deal teams who use primary research every day are already making the shift. They are tired of doing the work themselves. They are tired of paying for process and hoping for outcomes. And they are moving toward providers who will own the result.

The first era asked: can we connect you? The second era asked: can we show you what others have learned? The third era asks: what do you need to know?

That is where primary research is going. And that is what we are building at Woozle.