InvestorsFree
Investment Post Mortem and Attribution
Send the original memo and the realization, and get the multiple attributed to each assumption the memo committed to in writing.
An investment post mortem template is almost always a set of discussion prompts. What went well, what went badly, what we would do differently, who owned it. Those produce a meeting and a narrative, and the narrative is written by whoever is most articulate about the deal. None of the results ranking for this query asks the memo what it actually promised. So none of them can tell you whether the outcome turned on a call the committee made, or on a variable nobody underwrote at all.
Thornbury Analytics is an invented buyout, entered at 22.0m for 40.0 percent of the equity. The memo underwrote 18.0m of EBITDA, a 9.0 times exit multiple, 6.0m of net debt and 34.0 percent ownership after expected dilution, which together are a 2.4109 times multiple in year five. It realized 2.0704 times in year six, on proceeds of 45.549m. That is a shortfall of 0.3405 turns, and every post mortem template on page one stops right there.
Attributed to the memo's own variables it decomposes exactly. EBITDA coming in at 14.4m instead of 18.0m cost 0.5007 turns. The exit multiple landing at 11.5 times rather than 9.0 added 0.5564. Net debt at 21.0m instead of 6.0m cost 0.2318, and ownership at 31.5 percent rather than 34.0 cost 0.1643. Those four terms sum to the 0.3405 with no residual and no balancing line, which is the property that makes the lesson transferable to the next underwriting.
The memo's risk section pointed at the variable that helped
Read the attribution against what the memo flagged and the order inverts. Thornbury's risk section named multiple compression as the principal downside, and the multiple was the largest positive term in the bridge at plus 0.5564 turns. The two terms that actually cost money were the EBITDA build, which the memo supported with a page of detail, and net debt, which the memo did not mention at all because the acquisitions that created it were not underwritten. One of those is a forecasting error. The other is a process gap.
Time is the term that hides inside the headline. The multiple fell 14.12 percent, from 2.4109 to 2.0704, and the annualised return fell 6.35 points, from 19.24 percent to 12.90 percent. Split that: holding the underwritten multiple and stretching the hold from five years to six costs 3.45 points on its own, and the multiple shortfall costs the remaining 2.90. So more than half the return miss was the extra year, and the extra year appears nowhere in a template that asks what went well.
Two rules apply the moment a single-deal figure leaves the building. A firm showing extracted performance must provide, or offer to provide promptly, the performance results of the total portfolio it was extracted from-1). And the adviser has to keep the working papers necessary to form the basis for or demonstrate the calculation of any performance or rate of return it presents. An attribution built line by line at realization is that record, written while the inputs are still at hand.
How it works
Send the memo
The underwriting as it was written, with the assumptions and the risk section as they stood.
Send the outcome
Realization proceeds and the exit inputs, plus the interim reporting where you still have it.
Bridge the multiple
One variable at a time, in a stated order, until the terms sum to the shortfall exactly.
Score the assumptions
Each stated assumption marked against the outcome, and each unstated driver named on its own.
What you get
- The realized multiple attributed to each variable the memo underwrote, summing with no residual
- Every stated assumption scored against the outcome, with the page it was written on
- The variables that moved the result and were never in the memo, listed separately
- The return miss split between the multiple shortfall and the longer hold period
- What the memo's own risk section named, ranked against what the bridge actually shows
- One transferable underwriting change per finding, stated as a rule rather than a reflection
Common questions
Why attribute to the memo rather than to what happened?
Because a lesson has to be about a decision to be worth carrying. What happened is history. What the memo said would happen is a commitment the committee made with the information it had, and the gap between the two is the only part a future underwriting can act on. That is why the memo is the scoring key rather than the background.
What if the bridge does not sum to the shortfall?
Then it is not finished, and the run says so rather than plugging the gap. A residual means a variable is missing, an order was not held constant, or a figure changed definition between the memo and the realization. Each of those is findable. A bridge with a balancing line in it is a narrative wearing arithmetic.
Does the order of the bridge steps change the answer?
Yes, and that is why the order is stated. A sequential attribution assigns interaction effects to whichever term moves later, so a different order gives different individual terms and the same total. The output prints the order it used. The valuation methodology pack applies the same discipline to a quarterly mark bridge.
What about the variables the memo never mentioned?
They get their own section, because they are usually the finding. Thornbury's net debt went from 6.0m to 21.0m on acquisitions that were never underwritten, costing 0.2318 turns. Nothing in the memo was wrong about that. There was simply no row for it, which is a different kind of problem and a different fix.
Can this go into a track record we show investors?
The arithmetic can, with care. A single-deal figure is extracted performance, so the rules require offering the total portfolio performance alongside it. Restating a fund's own figures onto a comparable basis is what fund performance and benchmark analysis is for, and this attribution is the working paper underneath one line of it.
Does it work on a deal that went well?
It is arguably more useful there. A 3.4 times outcome built on one variable nobody underwrote is luck the firm should know it had, and the bridge shows that as plainly as it shows a shortfall. The same terms, the same sum, and the same question about which of them the process can actually claim.
Where does the exit multiple itself get examined?
Not here. This takes the realized multiple as a fact and attributes around it. Where the question is whether a multiple was defensible, comparable company and multiple analysis builds the set behind it, and secondary sale analysis handles a realization that is a transfer rather than a sale.
Investment Post Mortem and Attribution
Fill in the form and your workspace opens with the work already underway.