Private equity deal sourcing: a take-private turned into a screen

Permira and Warburg Pincus agreed on December 20, 2025 to take Clearwater Analytics private at $24.55 a share, about $8.4 billion of enterprise value and a 47% premium to the unaffected price. The DCF model and valuation pages test that price. This page asks the question a deal team asks the Monday after a deal like it is announced: which listed companies look the way Clearwater looked at signing?

Deal sourcing is how a fund finds companies to buy, and it runs on three routes. Bankers’ auctions put the same book in front of every bidder. Relationships bring a call from a founder or a board. A screen built on a thesis finds the companies a fund can approach before anyone runs a process. A fund’s investment criteria are the size, growth, margin, valuation and ownership that screen filters on. This page runs the third route end to end, on public data, using the Clearwater deal as the thesis.

Every company named below is the output of a screen and a filing review on public data as of October 19, 2026. Passing the screen means a company resembles Clearwater at signing on those measures. It says nothing about whether any company will be sold, and nothing here is an investment recommendation.

Investment criteria from one take-private, run as a company screen

Compound started by writing Clearwater at signing down as filters.

Clearwater is delisted, so no screen can read its numbers. Compound read them from the filings instead. At the unaffected date the stock was 51.0% of its 52-week high, organic ARR grew 17.0%, the 2025 adjusted EBITDA margin guide was 34%, and net leverage was 3.2x. Its pre-IPO sponsors, Welsh Carson, Warburg Pincus and Permira, reported 70.1% as a group in Welsh Carson’s first 13G in February 2022. By August 2025 Welsh Carson’s own stake was 1.6%. And the company had just closed the Enfusion acquisition.

The screen turns those facts into two kinds of filter. Four are filters every company must pass, counting the universe definition, and the price filter is the one that keeps the deal’s main feature: a good business the market had marked down. The other five are traits: revenue growth of 8% or more, an adjusted EBITDA margin of 15% or more, net debt of at most 3.5x EBITDA, a sponsor holding 5% or more since 2021, and an acquisition closed since October 2024. A company needs three of them, because no listed company matches a deal on everything. Clearwater’s growth is its organic ARR, because its reported revenue includes Enfusion. The other companies are measured on LTM revenue growth.

Fourteen companies pass. Every row in the workbook states why it passed, and every figure opens the filing or price series it came from.

The most useful check is the company that just missed. CCC Intelligent Solutions met all five traits and failed only the price filter, at 82% of its high. On October 15, 2026 it announced a take-private by GTCR and Elliott at $7.75 a share, a 44% premium to its close on July 9, before reports of a sale. The miss is a timing artifact: the screen runs on October 19 prices, which already include the deal. At its unaffected price, about $5.38 by our arithmetic, CCC sat near 58% of its high and would have passed the price filter. A screen that nearly catches a live deal is at least looking in the right place.

A screen on fields cannot see who controls a company, who is pressing it, or what its debt does on a sale. Some of the fourteen cannot be bought at any price.

What the filings show that a screen cannot

The next step reads what each passer has filed.

Each answer becomes a cited column in the Finalists tab, and the answers take the 14 down to six. These are steps 6 to 8 of the funnel above.

Three companies leave on voting control, which no screen field holds. The Bentley family controls 62.8% of Bentley Systems’ voting power through 29-vote Class A shares, per its 2026 proxy. Doximity’s Class B shares carry 10 votes each, and Vertex’s founding family controls it through Class B shares.

CoStar leaves next. Third Point and D.E. Shaw campaigned there in 2025 and 2026 over its spending on Homes.com and the board’s deference to its founder, and asked for Homes.com to be sold, not the company. Shareholders re-elected the whole board in June 2026, and Third Point has since exited. No 13D records either campaign, so the filing review did not find it. Compound rated CoStar medium feasibility for the same reason as the next four: nobody on its register is pressing for a sale.

Those four are ServiceTitan, Procore, Guidewire and SPS Commerce. ServiceTitan and Procore have pre-IPO growth investors on the register, ICONIQ at 6.7% and 8.5% in its 13Gs, but no holder pressing for a sale. Procore is also absorbing DroneDeploy, an $845 million purchase that closed in September 2026. Guidewire and SPS Commerce show no sponsor, activist or strategic review in their filings. The six left each have a sponsor, activist or past bidder on the register: BlackLine, Alkami, nCino, Waystar, Blackbaud and EverCommerce.

Two of the six show what the filing read adds. At BlackLine, no holder controls the company. Fivespan Partners filed a 5.1% Schedule 13D in March 2026 calling the shares undervalued, and holders of the 2029 convertible notes can require repurchase at par on a fundamental change, so a buyer has to refinance them. At Alkami, General Atlantic holds 18.1%. JANA Partners has held a 6.3% 13D stake since June 2026, when it said it would press the board to explore a sale. Its October 16 amendment adds talks on profitability, capital allocation and AI partnerships, and Reuters reported in August 2026 that Alkami had begun a sale process. Alkami’s 2030 convertible notes carry the same put.

A paper LBO on the deal’s own debt terms

Each finalist then gets a quick buyout test on Clearwater’s own financing.

A paper LBO is the short version of a buyout model: it tells you whether a price, a debt package and a growth rate can return the fund’s money, and it is too coarse to sign on. The terms matter more than the arithmetic, and here Compound caught a trap in its own data room. The folder labeled credit agreements held Clearwater’s April 2025 Enfusion facility at SOFR + 2.25%, the debt the buyout repaid. The take-private terms are in the debt commitment letter: a $2,700 million term loan at SOFR + 4.50% with a 0.50% floor, no amortization, and at least 40% equity. The loan is 8.0x management’s 2026 adjusted EBITDA, and 10.9x the $247 million of 2025 adjusted EBITDA Clearwater guided to. The valuation page’s LBO puts it at 10.5x 2025 consolidated EBITDA, which adds $10 million of public-company savings to management’s $248 million. The model applies the 10.9x to each finalist’s last twelve months of EBITDA, because that is the one figure every finalist reports, and none has published a plan to set against management’s.

The paper LBO does not cut the list. It tests whether Clearwater’s financing fits each finalist, and it feeds the ranking, where the two return measures carry 25% of the weight. Every finalist can carry the debt at a 30% to 35% premium and still cover its year 1 interest, so none drops out. The ranking also comes out in the same order with or without the returns: BlackLine, Alkami, EverCommerce, nCino, Blackbaud and Waystar. What the model adds is where the debt is strained. This page leaves out the per-share offer prices and returns the workbook computes for each name, because some of these companies may be in a sale process.

Read across, the six columns put the weak point in the debt. Adjusted EBITDA here adds back stock-based compensation, and the model replaces a quarter of it in cash after close, as PJT’s illustrative LBO did for Clearwater. Stock pay is 56% of BlackLine’s adjusted EBITDA, and at Alkami it is larger than the adjusted EBITDA itself. A lender pricing 9.5x to 10.9x of that EBITDA is lending against stock pay. Interest cover here is year 1 adjusted EBITDA, after a year of guided growth and before stock pay, over cash interest at 8.03%. That rate holds SOFR flat at the 3.53% implied by Clearwater’s own cost of debt in early 2026. On that basis five of the six finalists cover year 1 interest only 1.3x to 1.5x, which leaves little room. Blackbaud covers it 2.0x because the equity floor holds its debt to 6.7x. Counting the stock pay replaced in cash would lower it further.

The numbers work for a buyer on some names, on generous exits. Whether anyone is ready to sell is a question for the calls.

Earnings call signals: what Clearwater’s own calls showed

The last step reads the earnings calls, starting with Clearwater’s.

The usual idea is that a company about to sell shows strain on its calls: guidance cut, targets withdrawn, analysts asking about options. Compound read Clearwater’s last four calls first, as the benchmark, and found none of the classic distress signals: no guidance cut, no targets withdrawn, no analyst asking about a sale. Clearwater raised its 2025 guidance on the Q2 and Q3 2025 calls, and the stock fell from $20.42 to $18.79 the day after the Q2 raise. The signs were subtler: disclosure narrowing, and a $100 million buyback in September 2025 to support the price, traced on the earnings call analysis page. Compound’s score counts two signals in four calls.

That changes how much weight the calls deserve. A company that sells at a premium after raising guidance does not announce it on its calls, so call signals are the weakest evidence here. Compound’s ranking gives them 10%, tied for the lowest weight with the IRR at today’s multiple. The screen fit, the price against its high, the filing review and the IRR at the entry multiple get 15% each, and a sponsor, activist or bidder on the register gets 20%. The calls are better at catching a change than at predicting one: a new 13D, a CFO leaving or a buyback raised into a falling price is worth reading when it happens.

How to keep a target list current

The list is right on October 19, 2026. A new 13D, a pending acquisition or a refinancing on any of these names changes it.

Compound follows each company for the event types and SEC forms you pick and keeps your note of what matters in them. The screen is saved as a view the team opens in the app, and the company table holds a rating and a deal stage per name that every later task reads back. When JANA amends its 13D or a convertible note is refinanced, the next task starts from the table, not from a blank screen.

What one sourcing run produced

The run took a single deal to a cited shortlist of six.

One deal written down as nine filters found 14 of 77 listed software companies. Their filings cut the list to six with a holder who could start a sale. A paper LBO on the deal’s own commitment-letter terms showed where the returns come from (the exit multiple) and where the risk sits (debt sized on EBITDA that adds back stock pay). Clearwater’s own calls showed that a company can raise guidance on the way to a sale, so the calls rank last as evidence. The method’s value is in the order of the work: the filings and debt terms of every name are read before the list is cut, and the closest miss turned out to be a live deal.

FAQ