TL;DR: Primary research generates variant perception through structured interviews, surveys, and channel checks. The $2.5 billion expert network industry sells access at $1,200 per call with 50–70% margins whilst pushing all research work onto analysts. Investment-grade primary research requires triangulation across multiple sources, MNPI boundary awareness, and measuring cost per useful insight rather than cost per call. Smaller funds outperform larger ones (5.91% vs 4.04% over 5 years) because research intensity matters more than scale.

Core Framework:

• Primary research produces first-hand intelligence from original sources (experts, customers, suppliers) to validate investment theses and generate non-consensus insights before they become widely known.

• Legal foundation rests on mosaic theory: combining non-material information fragments into material insight whilst avoiding material non-public information (MNPI).

• Research intensity calibrates to investment stage: broad screening during idea generation, deep conviction building for position sizing, ongoing monitoring for thesis tracking.

• Cost per useful insight matters more than cost per call: $1,200 calls with 40% useless outcomes cost $2,000 per useful insight versus $600 finished intelligence with 95% hit rates costing $630.

• Effective research requires triangulation (minimum three independent sources), source diversity over volume, and treating primary research as compounding organisational capability rather than transactional expense.

I've spent years on the buyside watching investment teams pay research prices for access.

$1,200 per expert call. 14 hours a month on scheduling and note-taking. Analysts doing all the real research work themselves. The middleman takes 50–70% margin. The analyst gets a pile of transcripts. The fund gets a memo that might move conviction, or might not.

This is backwards.

Primary research generates non-consensus insights that compress the timeline from hypothesis to conviction. It operates as systematic capability, not transactional expense. It produces finished intelligence that survives IC scrutiny, not raw data you still need to clean and interpret.

This piece maps the framework I use to think about primary research within hedge fund investment processes. If you're new to the field or reassessing how you use it, this is where we start.

What Primary Research Actually Is

Primary research gathers first-hand information directly from original sources to validate investment theses and generate variant perception.

It differs from secondary research (analysing existing reports, filings, publicly available data) and quantitative alternative data (processing large datasets through algorithms to identify patterns).

Primary research methods include structured interviews with industry experts, customer and supplier surveys, channel checks with distributors and retailers, and on-the-ground intelligence from field research.

The value lies in generating insights proprietary to you, at least initially. Secondary information is widely available to the market. Expert interviews, surveys, and site visits validate trends, assess risks, and uncover opportunities before they become consensus.

This matters because sustainable edge derives from superior research processes rather than privileged access. As markets become more efficient, the ability to systematically generate variant perception through primary research becomes compounding organisational capability.

Key Point: Primary research produces proprietary intelligence through direct source engagement, creating informational edge before consensus forms.

How Mosaic Theory Makes Primary Research Legal

Mosaic theory enables analysts to gather non-material information fragments from company insiders and combine those pieces to create material insight into company health.

This is the legal foundation that makes primary research possible.

SEC staff guidance recognises an issuer would not be conveying material information if it shared seemingly inconsequential data which, pieced together with public information by a skilled analyst with knowledge of the issuer and industry, forms a mosaic revealing material insight.

However, leading compliance advisers recommend clients not rely on mosaic theory except where non-materiality is clear-cut. The SEC has not formally endorsed the theory in the context of insider trading. It relies on determinations of materiality subject to after-the-fact second guessing.

This creates strategic constraint: effective primary research must be designed from inception to stay within legal boundaries.

You need to understand what constitutes material non-public information before you start asking questions, not after you've gathered the data.

Key Point: Mosaic theory permits combining non-material fragments into material insight, but requires designing research boundaries before execution, not after.

Where Research Stops and Risk Begins: The MNPI Boundary

Information is material if there is substantial likelihood a reasonable investor would find it important in making an investment decision by having significantly altered the total mix of information available.

Information is non-public if it has not been disseminated broadly to the marketplace via press release or analyst report and has not permeated proper channels. Information is not considered public until fully internalised by the market, until the price of the security has fully adjusted.

The 2022 SEC Risk Alert highlighted common deficiencies across the industry:

• Poor documentation of MNPI policies

• Lack of controls around alternative data and expert networks

• Untrained access persons who do not recognise MNPI when they encounter it

Private markets fund managers need comprehensive analysis of the ways MNPI flows into firms (via discussions with public company executives or expert networks) and must prevent such information from flowing in, ring-fence it upon receipt, and monitor activities of individuals who have accessed it to mitigate misuse risk.

This is not theoretical. The legal boundaries around primary research create natural barriers to entry. They also create real risk if you get them wrong.

Key Point: MNPI violations carry enforcement risk. Design research to avoid material non-public information from inception, not through post-collection filtering.

What You're Actually Buying from Expert Networks

The $2.5 billion global expert network industry has grown 16% compound annually over the last decade. Private equity firms and strategy consultants drove adoption. Hedge funds and corporate strategy teams contributed to growth.

Expert networks now rank among the highest-paid research firms, with major players in the top-20 research providers. Hedge funds were once the primary clients. Traditional long-only asset managers now drive sector growth.

Expert networks' average payment per client grew 38% after two years of substantial growth. 80% of the research universe is long-only asset managers. The remainder is hedge funds.

Here's what the growth obscures: you're paying for access, not answers.

Expert networks connect investors with industry experts for one-on-one calls or Q&A, often within days, handling compliance screening and scheduling. This cuts time to find the right expert and ensures conversations stay within legal boundaries.

But $1,200 per call gets you an introduction and compliance screen. Not a verified answer. Not triangulated insight. Not finished intelligence you drop into an IC memo.

You still vet the expert, design questions, conduct the interview, take notes, fact-check claims, and synthesise output. The middleman takes 50–70% margin. You do all the work.

Key Point: Expert networks sell access and compliance screening, not verified intelligence. Research work and synthesis remain entirely on your team.

How Primary Research Maps to the Investment Lifecycle

Primary research operates differently depending on where you are in the research process.

Idea Generation and Initial Screening

At this stage, you're identifying promising opportunities and eliminating obvious misdirections quickly.

Primary research here involves broad industry expert interviews to map competitive dynamics, quick channel checks to validate or refute initial thesis assumptions, and targeted surveys to gauge market sentiment or emerging trends.

The goal is rapid triangulation: find three independent sources that support or contradict your initial hypothesis.

Deep Dive and Conviction Building

This is where primary research intensity increases significantly.

You conduct detailed expert interviews with former executives, customers, suppliers, and competitors. You run comprehensive surveys to quantify market size, growth rates, or customer satisfaction. You perform site visits and channel checks to verify operational claims.

The objective shifts from screening to conviction building: develop proprietary insight that justifies position sizing.

Ongoing Monitoring and Position Management

Primary research continues after you establish the position.

Regular check-ins with key industry contacts track whether your thesis remains intact. Periodic surveys monitor shifts in customer behaviour or competitive positioning. Channel checks provide early warning signals of deterioration or acceleration.

Repeating similar primary research over time creates compounding advantage through pattern recognition and tacit knowledge.

Exit Timing and Post-Investment Review

Primary research informs when to reduce or exit positions by detecting inflection points before they appear in financial statements.

Post-investment reviews use primary research to assess what you got right and wrong, building institutional memory that improves future research design.

Key Point: Research intensity calibrates to investment stage: breadth for screening, depth for conviction, consistency for monitoring, reflection for institutional learning.

The Landscape of Primary Research Providers

The array of primary research options is overwhelming. Here's how I organise the landscape:

Expert Networks and Platforms

These connect you with industry specialists for one-on-one consultations, charging per hour or per call.

The value proposition is speed and compliance screening. The limitation is you're still doing all research work: vetting, interviewing, note-taking, and analysis.

Survey and Panel Providers

These offer access to consumer or B2B panels for quantitative research, charging per complete or per project.

The challenge is quality control. Panel fraud, respondent fatigue, and middleman margin stacking degrade data quality significantly.

Specialist Research Firms

These conduct primary research on your behalf, delivering finished reports rather than raw access.

The distinction is critical: you're buying outcomes, not introductions. The economics shift from paying for access to paying for verified intelligence.

Internal Research Capabilities

Some funds build dedicated primary research teams handling sourcing, interviewing, and analysis internally.

This provides maximum control and proprietary methodology. It requires significant fixed cost and ongoing talent retention.

Key Point: Primary research vendors split between access providers (networks, panels) and intelligence providers (specialist firms, internal teams). Economics and outcomes differ fundamentally.

Why Most Primary Research Fails to Generate Alpha

Primary investment research has been labour-intensive: extensive data collection, interviews, communications with experts. Investors rely on financial statements, earnings calls, market reports, and expert networks. This approach, whilst thorough, requires significant time and resources to gather and analyse information.

The problem is execution, not concept.

One primary challenge in alpha generation is lack of thorough research. To identify investment opportunities generating alpha, you need comprehensive research and analysis: financial statements, industry dynamics, macroeconomic factors.

But thorough research and comprehensive research are not the same as effective research.

I've watched teams conduct 30 expert calls on a single name and miss the critical insight. They optimised for volume, not triangulation. They collected data points, not built a mosaic.

The hedge fund industry has grown from $265 billion in 2000 to over $5 trillion. The industry comprises around 15,000 hedge funds. In 2024, further concentration of flows is anticipated, with a small percentage of managers attracting 90% of net assets.

Smaller managers have delivered significantly better performance than larger managers. Through November 2023, smaller funds outperformed larger funds over 1 year (4.52% vs 3.30%) and 5 years (5.91% vs 4.04%).

This performance gap suggests research intensity and proprietary insight matter more than scale. Smaller funds move faster, dig deeper, and generate variant perception larger competitors cannot replicate.

Key Point: Research volume does not equal research quality. Alpha comes from triangulated insight building, not call count maximisation.

The Shift from Access to Finished Intelligence

The future of primary research is not more expert calls or bigger panels.

It's the shift from paying for access to paying for verified, decision-ready intelligence.

Alpha potential is greater in fixed income than equities due to the pronounced presence of constrained investors in bond markets. Active managers gain edge by exploiting scale, tilts, behavioural biases, fundamental research, and institutional analytics.

The same logic applies to primary research: edge comes from superior research processes, not superior access.

When you hand over a 10-minute brief and receive verified answers from correctly profiled experts (structured, fact-checked, ready to drop into a memo), you've eliminated the middleman tax. You've recovered 14 hours a month your analysts were spending on logistics. You've cut your cost per useful insight in half.

More importantly, you've shifted from transactional research spend to systematic capability building.

Key Point: Finished intelligence eliminates middleman tax and analyst logistics burden whilst converting research from transactional expense to compounding capability.

Core Principles for Investment-Grade Primary Research

These principles recur across every primary research project I've seen succeed:

Design for triangulation from inception. One expert is an anecdote. Three independent sources with aligned insights is a pattern. Five sources with quantified agreement is conviction.

Calibrate intensity to decision stage. Idea generation requires breadth. Conviction building requires depth. Ongoing monitoring requires consistency. Match research design to where you are in the process.

Embed MNPI boundaries in research design. Do not ask questions eliciting material non-public information and then manage risk after the fact. Design questions staying within legal boundaries whilst generating insight.

Measure cost per useful insight, not cost per call. A $1,200 call where 40% are useless has real cost per useful insight closer to $2,000. A $600 finished intelligence project with 95% hit rate has real cost per useful insight of $630.

Treat primary research as compounding capability. Repeating similar research over time builds pattern recognition and tacit knowledge. The tenth channel check in a sector is exponentially more valuable than the first.

Prioritise source diversity over source volume. Ten former executives from the same company give you one perspective repeated ten times. Two executives, three customers, two suppliers, and three competitors give you seven different lenses.

Key Point: Investment-grade research requires triangulation design, stage-appropriate intensity, proactive MNPI avoidance, true cost measurement, compounding repetition, and source diversity.

Where This Leaves You

If you're using expert networks and survey platforms the way most funds do (paying $1,200 per call, absorbing 40% useless outcomes, doing all the research work yourself), you're leaking returns.

The middleman model extracts margin whilst pushing risk and labour back onto your team. Not a research partnership. A tax.

Primary research generates non-consensus insights that compress your timeline from hypothesis to conviction. It operates as systematic capability that compounds over time. It delivers finished intelligence that survives IC scrutiny, not raw access you still need to process.

The question is whether your current vendor stack is built around your decision process, or around their volume metrics.

What would change if you only paid for research that genuinely moved conviction, sizing, or timing on a position?

Frequently Asked Questions

What is primary research in hedge funds?

Primary research gathers first-hand information directly from original sources (industry experts, customers, suppliers, competitors) to validate investment theses and generate variant perception before it becomes consensus. This differs from secondary research (analysing existing reports and filings) and quantitative alternative data (processing datasets algorithmically).

How much do expert networks charge for primary research?

Expert networks charge approximately $1,200 per expert call. With 40% of calls being useless, the real cost per useful insight approaches $2,000. Expert networks take 50–70% margin whilst analysts still vet experts, design questions, conduct interviews, take notes, fact-check claims, and synthesise outputs.

What is mosaic theory and why does it matter for primary research?

Mosaic theory enables analysts to gather non-material information fragments from company insiders and combine them into material insight. This is the legal foundation making primary research possible. However, compliance advisers recommend not relying on mosaic theory except where non-materiality is clear-cut, as the SEC has not formally endorsed it in insider trading contexts.

What is material non-public information (MNPI) in primary research?

Information is material if a reasonable investor would find it important in making investment decisions by significantly altering the total mix of available information. Information is non-public until disseminated broadly via press release or analyst report and fully internalised by the market (price fully adjusted). The 2022 SEC Risk Alert highlighted MNPI policy documentation failures, lack of controls around expert networks, and untrained personnel.

How should primary research intensity change across the investment lifecycle?

Idea generation requires breadth (broad expert interviews, quick channel checks, targeted surveys for rapid triangulation). Conviction building requires depth (detailed expert interviews, comprehensive surveys, site visits to justify position sizing). Ongoing monitoring requires consistency (regular check-ins, periodic surveys, channel checks for early warning signals). Exit timing requires reflection (detecting inflection points, post-investment reviews for institutional learning).

Why do smaller hedge funds outperform larger funds?

Through November 2023, smaller funds outperformed larger funds over 1 year (4.52% vs 3.30%) and 5 years (5.91% vs 4.04%). This performance gap suggests research intensity and proprietary insight matter more than scale. Smaller funds move faster, dig deeper, and generate variant perception larger competitors cannot replicate.

What is the difference between access and finished intelligence?

Access means paying for expert introductions and compliance screening whilst analysts still vet experts, design questions, conduct interviews, take notes, fact-check, and synthesise. Finished intelligence means paying for verified, structured, fact-checked answers ready to drop into IC memos. Access requires 14 hours per month of analyst logistics. Finished intelligence eliminates this burden.

How do you measure the real cost of primary research?

Measure cost per useful insight, not cost per call. A $1,200 call with 40% useless rate costs $2,000 per useful insight. A $600 finished intelligence project with 95% hit rate costs $630 per useful insight. Include analyst time spent on vetting, scheduling, interviewing, note-taking, and synthesis when calculating true research costs.

Key Takeaways

• Primary research generates variant perception through first-hand intelligence from original sources, creating informational edge before consensus forms.

• Expert networks sell access ($1,200 per call, 50–70% margins) whilst pushing all research work onto analysts, creating real cost per useful insight of $2,000.

• Mosaic theory permits combining non-material fragments into material insight, but MNPI boundaries must be designed into research from inception, not managed post-collection.

• Research intensity calibrates to investment stage: breadth for screening, depth for conviction, consistency for monitoring, reflection for institutional learning.

• Effective research requires triangulation (minimum three independent sources), source diversity over volume, and measuring cost per useful insight rather than cost per call.

• Smaller funds outperform larger funds (5.91% vs 4.04% over 5 years) because research intensity and proprietary insight matter more than scale.

• Finished intelligence eliminates middleman tax and 14 hours monthly analyst logistics burden, converting research from transactional expense to compounding capability.