How should I discuss losses and failed investments with LPs?

What an allocator reads in the loss data itself

Before any narrative, the numbers are read on their own. Each line below is a signal rather than a verdict, and none has a right answer this page will state, because loss patterns depend on stage, vintage, geography, and construction and no verified source publishes a universal figure.

• Loss ratio. What share of positions returned less than cost. Read alongside stage: the same figure means different things at pre-seed and at growth.
• Write-offs. Positions taken to zero, with the date. A record with none, in a portfolio old enough to have some, prompts a question about the valuation process rather than about the picking.
• Failure rate against thesis. Whether the losses cluster where your thesis said risk lived, which is coherence, or somewhere else, which is a question.
• Concentration. Whether the losses are in the largest positions, since a loss in a top position costs the fund far more than the count suggests.
• Follow-on decisions. Explain whether you invested more, what evidence you considered and why you chose that course.
• Reserves consumed. What the losses cost in follow-on capital as well as in initial cheques.
• Markdowns before write-offs. Whether the valuation process caught deterioration in advance or the position went from carried at cost to zero in one step.
• Timing. Explain how the assessment changed as new information arrived, including decisions to hold, sell or write down a position. Separate what was knowable then from what became clear later.

The six questions, and what a credible answer sounds like

Prepare these six questions for material losses. They help you explain the decision and its outcome without relying on hindsight.

• Why was the investment made? The reasoning as it stood at the time, ideally from a memo written then rather than reconstructed now. An answer that sounds wiser than the decision was is the tell allocators listen for.
• What changed? Market, team, product, financing environment, or execution. Be specific about which, because the category determines whether the loss says anything about your selection.
• What did you learn? One thing, stated plainly. A list of five lessons from one company reads as narrative rather than as analysis.
• What changed in your process? The hardest question and the most valuable answer. A lesson with no process change behind it is a sentiment. If nothing changed, saying that and explaining why is more credible than inventing a change.
• Was the loss thesis-consistent? Losses inside your stated risk are evidence the thesis is honest. Losses from a risk your thesis said you avoided are the ones that need a real answer.
• Was it avoidable? Sometimes yes. Managers who occasionally answer yes are believed when they answer no.

Candour without minimising, which is a narrow path

Two presentation problems are worth checking for: leaving losses out and explaining away their significance.

The first is omission: a track record showing only companies that worked, a narrative that moves quickly past the write-offs, or a loss column that appears nowhere in the materials. This is usually discovered rather than disclosed, and being discovered converts a normal outcome into a credibility problem.

Over-explanation can make responsibility hard to follow. Describe the result, the decisions within your control and the factors outside it. State any process change that followed; if none did, explain why without inventing a lesson.

The path between them is descriptive rather than defensive. State what was invested, what was returned, what happened, and what you would do differently, then stop. The restraint is the signal.

Missed opportunities, which are a different conversation

Allocators ask about deals you passed on that succeeded, and the question is not about the miss.

It tests two things: whether you record your reasoning at the time, and whether your process changed when it should have. A manager who can produce the pass note, explain what they saw and did not see, and say whether the process has changed since is demonstrating exactly the discipline the question is looking for.

If you did not see the company, explain the limits of your sourcing. If you would still have passed, describe the criteria behind that decision. The question is how the opportunity relates to the strategy you actually run.

What to put where

The public and confidential versions differ in detail and not in candour.

• In the deck: the loss count and the write-offs inside the track record table, so the sample is complete. No company-confidential explanation.
• In the data room: the position-level record including cost, proceeds, status, dates, and the memo or note behind the original decision where one exists.
• In conversation: the fuller story, including anything a founder would consider confidential, which should be discussed rather than written down without consent.
• Handle accounts involving founders fairly and within confidentiality obligations. Distinguish documented facts from your interpretation, and be prepared for a reference to offer a different perspective.

What limited partners are testing

The discussion helps an LP assess how you describe decisions, uncertainty and responsibility when the outcome is disappointing.

• Does the manager raise the losses, or wait to be asked?
• Does the account match what the founder and co-investors say?
• Is there a written record of the original reasoning, or only a current recollection of it?
• Did anything actually change in the process, and can the manager name it?
• Does the manager distinguish an unlucky outcome from a bad decision, in both directions?

What discussing losses well does not do

It does not offset the losses. Candour makes a record credible; it does not make it stronger, and a manager whose portfolio underperforms is not rescued by describing it accurately.

This page states no loss ratio, no failure rate, and no view on what proportion of a venture portfolio should be written off, because those depend on stage, vintage, geography, and construction and no verified source publishes a universal figure.

This page is educational and general. It is not legal, tax, accounting, valuation, or investment advice, and any presentation of investment outcomes should be reviewed by counsel where it will be used in marketing.