How Buy-Side Firms Structure Expert-Network Usage Caps by Deal Stage: 7 Models

Buy-side firms rarely spend on expert-network calls at a uniform rate across the life of a deal. A name at idea-screening does not warrant the same research budget as a name two weeks from an investment committee vote on a USD 500M commitment, and treating them the same is the fastest way to either overspend on dead ideas or underspend on live ones. The structural question is not how much to spend on expert calls in aggregate, but how to gate access at each stage of the funnel.

Below are seven models in active use across private equity, hedge funds, and long-only shops. They are not mutually exclusive , most firms run some combination , but the underlying logic differs enough that the choice of primary model shapes analyst behavior, compliance overhead, and vendor selection.

1. Flat per-stage call quotas

The simplest structure assigns a fixed number of calls to each stage of the funnel: three at screening, ten at initial diligence, an expanded allocation at deep diligence, and effectively unlimited at IC prep. This model is common at mid-market private equity shops that carry retainer relationships with Guidepoint or Third Bridge and want predictable monthly usage rather than variable spend.

The appeal is administrative. A deal team lead can approve a call without escalating to a partner, the operations team can forecast usage against the retainer, and analysts learn quickly which questions are worth the quota. The weakness is that call counts are a poor proxy for research value. A one-hour call with the right former operator can substitute for five calls with adjacent experts, and a rigid quota punishes the analyst who picks well.

2. Dollar-cap escalation by stage

A dollar cap replaces the call count with a spend ceiling that widens as the deal progresses. Screening might sit at roughly USD 5,000, initial diligence at roughly USD 25,000, and IC prep unlocked subject to partner sign-off. This structure is a better fit for firms that use premium networks, where per-hour rates at AlphaSights and similar providers sit in the USD 1,000 to 1,500 range and a single deep-expert call can consume a screening budget on its own.

The dollar model gives analysts room to trade off call length and expert quality against volume. It also makes the LP-facing conversation easier: management fee offset discussions increasingly focus on research spend as a controllable line item, and a per-stage dollar cap is a defensible answer to the question of how the fund controls it.

3. Deal-probability-weighted budgets

Some hedge funds, particularly those with more systematic overlays on fundamental research, tie the cap to the deal team's own conviction score. A name flagged at 20% probability of making the book gets a modest budget; a name at 70% conviction unlocks a wider allocation.

The logic is that research spend should track expected value, and the deal team's own scoring is the cheapest available signal of that. The risk is well known to anyone who has run this model: analysts learn to game the conviction score to unlock the budget they want, which is why the firms that use it typically pair the score with a post-deal review that grades calibration.

4. Sector-adjusted caps

Not every vertical is equally expert-dependent. A healthcare services deal, an industrial roll-up, or a specialty chemicals thesis typically requires more primary research than a software deal where the customer base is small, public, and reachable directly. Firms that acknowledge this run a sector multiplier on top of whatever stage-gated structure they use, so the healthcare team's diligence cap might sit 50% to 100% higher than the software team's.

The sector adjustment is where the choice of vendor also flexes. Third Bridge and Guidepoint have deeper benches in some verticals than others, and firms that run sector-adjusted caps tend to also run sector-specific vendor preferences rather than a single firm-wide contract.

5. Pre-committed commercial due diligence packages

For deep-diligence work, the fixed-fee commercial due diligence bundle has become an accepted alternative to metered expert access. Dialectica and Third Bridge market packaged CDDs , a defined scope of expert calls, survey work, and synthesis, delivered as a single deliverable at a fixed price , and Inex One's data puts the global volume at roughly 36,000 CDDs per year.

The packaged model shifts the stage-gating decision from per-call approval to a single purchase decision at the point where a deal moves from initial to deep diligence. It works well for private equity, where the deep-diligence stage has a clear beginning and end and the deliverable feeds directly into the IC memo. It works less well for public-market investors, where diligence is continuous and the packaged deliverable is stale within a quarter.

[]

6. Rolling 12-month firm-wide caps with stage sub-allocations

The multi-strategy pod model , Citadel, Balyasny, and similar platforms , imposes a rolling 12-month cap at the pod or firm level and then sub-allocates it across deal stages. A pod might get a firm-wide research budget of a defined size, with internal rules that no more than a set percentage can be consumed at the screening stage, forcing discipline about which ideas graduate.

This structure reflects two realities of the multi-strategy world. First, PM turnover means the firm needs a budget it can hand cleanly from one PM to the next without renegotiating vendor contracts. Second, the compliance overhead of running dozens of pods against a single set of expert-network contracts is easier when the caps are firm-wide and the sub-allocations are internal.

7. Unlimited-at-IC override

The most economically honest model treats the diligence-stage cap as a hard limit and the IC-stage as functionally uncapped subject to partner approval. The reasoning is direct: the marginal expert call before a USD 500M commitment costs USD 1,500 and shifts the decision by a non-trivial amount in expectation. Capping it is false economy.

What this model actually caps is the number of names that reach IC prep, not the research spend on the names that do. The gating happens earlier in the funnel, and the IC-stage override exists to make sure the last-mile research on a live deal is never the constraint. Firms that run this model typically pair it with a strict diligence-stage cap that forces analysts to kill names rather than pad the case for advancing them.

What the choice of model actually shapes

The seven models overlap in practice, but the primary structure a firm picks tends to correlate with vendor selection. Flat call quotas and firm-wide rolling caps favor networks with subscription-plus-usage hybrids , the model GLG and Guidepoint have long operated , because the retainer smooths the accounting. Dollar-cap escalation and IC-override models favor premium per-hour networks where the marginal call is expensive but the quality justifies the price. Sector-adjusted caps push firms toward multi-vendor rosters. Packaged-CDD models pull budget toward Dialectica, Third Bridge, and the boutiques that specialize in the deliverable. Transcript-first tiers such as Third Bridge Forum fit the screening stage cleanly, where the marginal question is often better answered by reading a prior call than by scheduling a new one.

The more useful question for a research operations lead is not which model is correct in the abstract, but which model matches the firm's actual deal flow. A PE shop that runs twenty diligence processes a year and closes three has a different problem from a pod running two hundred names a year and holding twenty. The cap structure that produces good research on the first firm's twenty processes will starve the second firm's twenty holdings.