How to run private comps and precedent transactions
TL;DR
Private comps are built from precedent private transactions. Pick five to ten deals that share the target's model, growth and geography. Use EV/Revenue and EV/EBITDA at close. Show the range, then defend where your deal sits and why.
How to pick the deal set
Same industry is not enough. You want private deals with a similar model (recurring vs one-off), similar growth, similar profitability, similar deal type (venture, growth, buyout, minority, control) and similar geography.
Five to ten precedents. Fewer is thin. More adds noise from deals that are not really comparable.
Which multiples and why
EV/Revenue. Standard for venture and growth deals with limited or negative profitability.
EV/EBITDA. Standard for buyouts and mature private companies.
EV/ARR. Standard for private SaaS.
Always compute enterprise value at close, using announced deal value plus assumed net debt.
Where to source the transactions
Institutional data providers (PitchBook, Preqin, S&P Capital IQ, Mergermarket) if the fund pays for them. Regulator offer documents for take-private transactions (for example UK Takeover Panel filings). Trade press for disclosed venture and growth rounds. Fund LP letters when public. Never a single press release as the sole source.
Defending the range
Show the median and the range. Explain why your target trades at, above or below the median. Growth premium, margin discount, market position, deal structure. Be specific and cite the precedent you are anchoring to.
How this shows up on the job
Every private capital IC memo has a comps page. Getting the precedent set right is what separates a defensible valuation from a lazy multiples table an IC partner will reject.
How Reuben AI helps
Reuben AI builds private comps tables from primary transaction sources with the enterprise value calculation shown and every input linked to its source. Students can start on the Free workspace.
Sources
Frequently asked questions
How many precedents is enough?
Five to ten. Fewer is thin. More adds noise from deals that are not really comparable.
Announced or completed transactions?
Completed for the primary set. Announced can be a directional cross-check when the completion universe is thin.
How far back should precedents go?
Usually three years. Older transactions get noisy as market conditions change.
What if there are no clean private precedents?
Widen the geography or the model, and disclose the widening. A defensible thin set beats a padded one.
How does AI help with private comps?
Pulling and normalising the inputs from primary transaction sources. The precedent selection and defence still need judgement.
Learn on the platform you will use on the job.
Start with the Free workspace to practise IC memos, LBOs, DCFs and private-capital workflows. Faculty, program leads and student investment club leaders can book a demo to discuss teaching seats and program access.