Why does a success fee create a race to close? Because it pays everyone around the deal for one outcome, completion, and nothing for the other, walking away. Once every incentive in the room points the same way, so does the advice. For an acquirer the question is not whether the people are honest. It is whether a structure that rewards only closing can be expected to surface the reason not to.
Manufacturing and software engineering met this exact problem and solved it a generation ago, with two moves that have nothing to do with finance. Find the defect as early as possible, because the cost of fixing it climbs with every stage it survives. And give the person who finds it an incentive that is indifferent to whether the line stops. An acquisition process that fails both tests is not badly run. It is built the way quality engineers learned not to build. The alternative is the one they did build: a sequence of gates, each placed where a defect is still cheap to find, each with the authority to stop the line.
The Mechanism
A success fee is a call option on the deal. The broker holds it, the sell-side banker holds it, and often the buy-side advisor holds it too. The option pays only if the transaction completes. It expires worthless if the buyer declines. Every week of work raises the sunk cost; nothing raises the payoff for saying no.
Follow the incentive through the people in the room:
- The sell-side advisor is paid a percentage of enterprise value. Curating the CIM is not dishonesty; it is the mandate.
- The buy-side advisor on a success fee is paid on the same event as the seller. The buyer has hired someone whose interest in closing matches the counterparty’s, and whose interest in the buyer walking is nil.
- The deal team is measured on deployment. Fees have been running since the NDA. Walking at week eight means explaining the spend; closing means a portfolio company.
- The target has told its staff a sale is coming. Every week of exclusivity makes a collapse costlier for management, and management says so.
None of these parties needs to act in bad faith. The structure does the work. A red flag found in week eight is weighed not against the price of the deal but against the price of not doing it, which by then includes everyone’s fees, time and reputation. So it becomes “manageable”, “priced in”, “a post-close workstream”. Due diligence stops asking should we? and starts asking how do we get comfortable?
What the Evidence Says
This is not a diligence consultant’s intuition. The academic record on advisor fee structure is thin but consistent, and it has pointed the same way for thirty years.
Contingent fees buy completion, not performance. Rau (2000), in the Journal of Financial Economics, found that an investment bank’s M&A market share rises with the contingent fees it charges and with its past completion rate, but is unrelated to how the acquirers it advised went on to perform. In tender offers, acquirer performance was negatively related to the bank’s contingent fee. Rau’s reading: the fee structure ensures that banks focus on completing the deal.
The contract itself carries the conflict. McLaughlin (1996) modelled the incentives in tender-offer fee contracts and found substantial banker-client conflicts, with simulated losses of up to 16.7% of target value, held down only by the banker’s reputation capital. His 1992 study had already concluded that fee contracting is only a partial solution to the agency problem.
People persist past bad news, and experience does not help. Haunschild, Davis-Blake and Fichman (1994) ran a simulated acquisition in which participants received negative information about the target mid-process. Personal responsibility for the decision, competition for the target and having gone public with it each increased commitment despite the bad news. Acquisition experience made no difference. Their remedy is the one this piece is about: assign target selection and due diligence to different people.
The counter-evidence, and what it does not test. Calomiris and Hitscherich (2005), on US cash tender offers, found that fee structure did not predict the premium paid and read fees as ordinary compensation for risk. It is the best case for the benign view. It is also a study of sell-side fees in deals that closed, which is precisely the selection this argument is about: the deals the structure should have stopped are not in a sample of completed deals.
The literature does not contain, as far as we can find, a direct measurement of buy-side success fees against buyer walk-away rates. The chain runs from contingent fees to completion focus, from fee contracts to conflict, and from responsibility and publicity to persistence past bad news. It does not need the missing link to point in one direction.
What Quality Engineering Learned First
Strip the finance out and the deal process is a production line with an inspection step at the end. The target is the product. The data room is final inspection. And the inspectors are paid only if the product passes. Three bodies of work say what happens next.
The cost of a defect rises with every stage it survives. Barry Boehm and Victor Basili’s 2001 summary in IEEE Computer put a number on it: finding and fixing a problem after delivery is often around 100 times more expensive than at the design stage, closer to 5 to 1 for small systems. The multiplier has been argued over for twenty years. The gradient has not. The deal equivalent is exact: a pricing pattern spotted before the LoI costs one conversation and perhaps a price adjustment. Found in week eight of exclusivity, it costs the QoE, legal and management time spent on the assumption it was not there, plus the partner’s credibility. Nothing about the finding changed. Its position on the line did.
Inspection at the end does not create quality, and incentives tied to throughput destroy it. W. Edwards Deming’s third point is to cease dependence on inspection and build quality into the process; his eleventh is to eliminate numerical quotas, because a workforce paid on units shipped will ship units. Confirmatory diligence after exclusivity, run by people compensated on deals closed, violates both at once.
Stopping the line must cost the person who stops it nothing. The Toyota Production System gave every worker the authority to pull the andon cord and halt production on a defect. The insight was not the cord but the incentive around it: pulling it is the correct act, the team lead’s job is to come and help, and nobody’s pay falls because the line stopped. A worker penalised for stopping lets the defect pass and lets the customer find it. A success fee is the opposite design: a personal cost for pulling the cord, for everyone who could pull it, scaled to the size of the deal.
The standard acquisition process is therefore a textbook case: inspection at the most expensive point on the line, run by people on throughput incentives, in a structure that penalises stopping. Quality engineering’s answer was never to hire more honest inspectors. It was to move inspection upstream and make stopping free.
How the Race Shows Up in Diligence
The fee structure does not announce itself. It shows up as timing, framing and omission.
Work migrates to the end. An advisor paid at close concentrates effort where it supports closing: the fairness opinion, the SPA, the final QoE reconciliation. The early question, whether this target deserves an LoI at all, has no fee attached and gets answered by whoever has time.
Findings get reclassified. A permanent-sale pricing pattern that would have ended interest in week one becomes a “post-close pricing workstream” in week eight. An EPR registration gap becomes “a known cost, estimated at”. A review base that does not reconcile to order volume becomes “a marketing-attribution question”. The finding is identical. What changed is how much everyone has already spent getting to it.
The CIM is read as evidence rather than advocacy. Everything in it is true. The risk lives in what is absent, what has been averaged until the pattern disappears, and what is framed as a strategic choice rather than a liability. A buyer whose own advisors share the seller’s payoff has nobody in the room with a reason to read it that way.
There is one window where the buyer can still see clearly and act freely: before the LoI, when walking costs nothing and the price is still open. That window is exactly where a success-fee advisor has least reason to be.
Why a Fixed Fee Changes the Answer
A fixed fee for a screen is indifferent to the outcome. The fee is the same whether the report says “clean, proceed”, “proceed at a different price” or “walk”. That is the only arrangement under which a negative finding is worth as much to the advisor as a positive one, and so the only arrangement under which it can be expected to reach the buyer undiluted. It is the andon cord with the penalty removed.
It also changes when the work happens. Paid at delivery rather than at close, the work can sit where information is cheapest to act on: before the LoI, while a no still costs nothing. That is the cost-of-change gradient working for the buyer instead of against it.
Three consequences:
- The cost is visible and declinable. A fixed fee is an expense you can choose not to incur. A success fee is a price you have already agreed to pay the moment you decide to proceed, which is one more reason to proceed.
- The deliverable is a decision, not a deal. A risk map, a sharper set of questions and a view of what the known issues should cost. “Don’t” is a successful outcome. So is “yes, at a different price”.
- A no is a result, and so is a priced yes. A screen that never says no is not screening. But the common outcome is a qualified pass: the target proceeds with its issues named, costed and reflected in the offer. The fee for three screens is cheaper than the fee buried inside the one deal a buyer regrets.
The screen does not replace due diligence. It aims it. The QoE, legal and commercial work after the LoI is still needed; the screen makes sure the decision to commission it was taken while the buyer could still decline to.
Three Gates Before the LoI
The incentive problem does not start at the LoI. It starts at sourcing. A target that arrives through a broker, an auction or a banker’s “we thought of you” call arrives pre-curated by someone paid on completion.
Haunschild and colleagues’ remedy, separating the people who pick the target from the people who test it, is the organisational form of the same quality principle: the inspector should not be the person whose output is inspected. Quality engineering adds a second principle. Inspection is not one event at the end of the line but a sequence of gates, each placed where a defect is still cheap to find, each with a pass criterion and the authority to stop. The population shrinks at every gate, and the expensive work is spent only on what survived the cheap checks.
Tronvik’s process is built as that sequence: three gates before the LoI, all paid on delivery, none on completion.
Gate 1: exhaustive deal search. The buyer’s thesis is run against the whole addressable universe of e-commerce operators, built from public evidence rather than any intermediary’s inventory. This is incoming inspection, and it fixes a sampling problem before it exists: a broker’s longlist is a sample of owners who have decided to sell and advisors who have decided the buyer is a plausible closer. A longlist built from the full population is the buyer’s own.
Gate 2: pre-outreach screen. Before any contact with the target, each candidate is screened on outside-in evidence alone: pricing behaviour, review base, delivery promise, marketplace dependency, regulatory exposure. This is the in-process check and the cheapest place on the line to find a defect: no NDA, no management meeting, nobody at the target aware the buyer exists. The Flash Screen does this work, and its output is a clean pass, a qualified pass with the issues named, or a reason not to make the call.
Gate 3: deep-dive pre-LoI screen. For candidates under NDA, the Full Screen reconstructs unit economics and compliance position from evidence the seller does not control: the observations ledger, the GP1 to GP3 waterfall and EU regulatory exposure, delivered as a management-meeting questionnaire, a risk-scoped DD brief and priority checks. This is final test, run before the LoI rather than after it, so exclusivity fees, counsel and QoE time are committed on findings rather than on the CIM.
Then your due diligence, aimed by the three gates and run by your counsel and QoE team on terms you set, on a target that has passed three independent inspections with the questions already written and the price still open.
A gate is a qualification, not a veto. Factory inspection sorts parts into pass, rework and reject, and rework is the common outcome. A target with a known EPR registration gap or a pricing practice that must stop is not disqualified by the finding. It qualifies with the finding attached, and the cost of fixing it goes into the price and the SPA rather than into a post-close surprise. The point of every gate is that the buyer knows, early enough to price what it knows. The targets that reach the LoI have been qualified three times, and arrive with their known issues already in the price.
The Pre-LoI Question Every PE Fund Should Ask
Not “who is advising us”. The question is: of the people who have looked at this target on our behalf, how many are paid more if we buy it than if we don’t?
If the answer is everyone, the buyer has not yet heard from anyone with a reason to pull the cord. The evidence has been curated once by the seller and once by the structure. The window in which that can still be corrected closes at the LoI.
This analysis is part of Tronvik’s pre-LoI screening methodology. Nothing in this article constitutes legal, financial or investment advice. To start a fixed-fee screen on a specific acquisition target, contact info@tronvik.com.
Sources
- Rau, P. R. (2000). Investment bank market share, contingent fee payments, and the performance of acquiring firms. Journal of Financial Economics 56(2). RePEc
- McLaughlin, R. M. (1996). Adverse contract incentives and investment banker reputation. Journal of Financial Research 19(1). RePEc
- McLaughlin, R. M. (1992). Does the form of compensation matter? Journal of Financial Economics 32(2). DOI
- Haunschild, P. R., Davis-Blake, A., and Fichman, M. (1994). Managerial overcommitment in corporate acquisition processes. Organization Science 5(4). DOI
- Calomiris, C. W., and Hitscherich, D. M. (2005). Banker fees and acquisition premia for targets in cash tender offers. NBER Working Paper 11333. NBER
- Boehm, B., and Basili, V. R. (2001). Software defect reduction top 10 list. IEEE Computer 34(1). DOI
- Deming, W. E. (1986). Out of the Crisis. MIT Press. The 14 points, ASQ
- Toyota Motor Corporation. Toyota Production System