THRY · Research
Thryv Sells the Yellow Pages but Its Small Business Software Has Stopped Growing
Print goes to Carolwood for $142M to cut debt, but software revenue is flat, retention is 90%, owner cash is near zero, and Paulson owns 19.0%.
Spreadsheet & IC Deliverables: Download the formula-linked THRY 5-tab Excel (.xlsx) workbook and printable 1-page tearsheet.
Download THRY .xlsx →- Owner cash after capex, capitalized software and all stock pay
- 1.2% of revenue
- Paulson & Co. stake
- 19%
- Software share of revenue
- 76%
1. Our call: Hold off until the print sale closes and software growth returns
Our call: Hold off, conviction 3 of 5. Thryv Holdings sells marketing and customer management software to small local businesses such as plumbers, salons and lawyers, and still runs the old Yellow Pages print directories in the U.S., Australia and New Zealand [1]. It is a Texas-based U.S. company and a small one, with total 2025 revenue of $785.0M [1].
The story changed on Sep 12, 2026, when Thryv agreed to sell the print directories business to an affiliate of Carolwood, a Los Angeles private equity firm, for $142M in cash, with the proceeds going to pay down debt [6][7]. That turns Thryv into a software company with a shrinking online marketing tail [7].
The problem is that the software line is no longer growing. Software revenue fell 0.5% in the second quarter to $114.5M, and the full year guide of $453.0M to $457.0M is below the $461.0M of 2025 [4][1].
We measure owner cash as operating cash flow, less capital spending including capitalized software, less all stock pay. For the six months to Jun 30, 2026 it was $3.7M, or 1.2% of revenue, our calculation [4].
There is no controlling holder, but Paulson & Co., the hedge fund run by John Paulson, owns 19.0%, and lead independent director John Slater was a Paulson partner for ten years
2. A small business software suite with a print tail
| Metric | Value |
|---|---|
| What it sells | Software for marketing, leads, payments and customer records; print and online directory listings [1] |
| Software clients | 95 thousand at Jun 30, 2026, with monthly revenue per client of $394, up 11.9% [4] |
| Seasoned net revenue retention | 90% in the second quarter [4] |
| Stock pay | $7.5M in six months, 2.4% of revenue, our calculation [4] |
| API quality | 2 of 5, our scoring: integrations with clients' existing customer systems and new partnerships with Ooma and Wix, but no open developer platform at scale [4][7] |
| Toll-gate score | 1 of 4 pass, 1 partial, 2 fail (section 2b) |
2b. Thryv does not own a gate
A gate is valuable when its owner holds a record others must check and prices what passes through. Thryv's clients are millions of tiny businesses that can switch to HubSpot, Wix or Google tools with little friction, so Thryv holds a useful record, the client's contacts and leads, but not one anyone else must check [1].
| Test | Result | Evidence |
|---|---|---|
| Owns the record | Partial | It holds each client's customer list and lead history, but those can be exported [1]. |
| Owns the connection | Fail | Seasoned net revenue retention of 90% means older clients spend less each year [4]. |
| Prices close to outcomes | Pass | The new Growth Platform is sold on leads won and return on marketing spend, and monthly revenue per client rose 11.9% [4]. |
| R&D on the gate vs features | Fail | Research and development was 5.0% of 2025 revenue, our calculation, small for a software company [1]. |
The verdict is one pass. Thryv sells tools that help, but its retention says clients drift away or trim spending, and its research budget is too small to build a moat [4][1].
The pricing shift toward outcomes is the one encouraging sign [4].
3. Revenue is shrinking as print winds down
Thryv bought Keap, a small business sales software maker, in late 2024 for $77.0M in cash, which is why 2025 software revenue jumped [1]. Without Keap, software growth has slowed, and the second quarter included what management called headwinds in older customer management products [4][1].
Inside software, the mix is improving. Quality customers, those paying more than $400 a month, made up 72% of software revenue excluding Keap, and the Market, Sell, Grow add-ons grew 21% excluding Keap [4].
The older customer management products are shrinking faster than the new ones grow [4].
Software gross margin was 63.5% in the second quarter, or 66.6% on the company's adjusted basis, low for software because of the service and support staff small clients need [4].
Marketing services revenue fell 62.0% to $36.2M in the second quarter, and print revenue fell 62.8% to $24.9M, partly on publication timing [2]
4. The quarter on Nov 9
Third quarter guidance is software revenue of $111.0M to $112.0M and software adjusted EBITDA of $8.5M to $9.5M, with marketing services revenue of $34.0M to $35.0M [4]. That software guide is below the $114.5M of the second quarter [4].
The release should also update the timing of the print sale close, expected in the fourth quarter, and how much of the $142M goes to debt [7]. The board replaced the 2026 bonus plan with a second-half plan with new targets after the restructuring, so management's incentives were reset mid-year [8].
5. GAAP results swung to a loss
Net income was $0.3M in 2025, after a $74.2M loss in 2024 [1]. For the first half of 2026 the net loss was $12.1M, and $16.7M in the second quarter alone [4].
Adjusted EBITDA fell 59.4% to $20.8M in the second quarter, a 13.8% margin [4]. Software adjusted EBITDA was $13.6M, an 11.8% margin, down from $23.4M a year earlier [4].
For the full year, management guides software adjusted EBITDA of $42.0M to $44.0M and marketing services adjusted EBITDA of $31.0M to $33.0M, our sum $73M to $77M, against $151.8M in 2025 [4][1]. Most of that drop is the print decline and the investment in the new platform [4].
The restructuring is meant to fix that. Management expects $20M to $25M of charges and $55M to $60M of gross annual savings once complete, starting in 2027 [4].
6. Owner cash has nearly disappeared
Owner cash is our calculation; Thryv discloses capital spending and capitalized software as one line [1][4]. It fell from 10.1% of revenue in 2023 to 3.9% in 2024, 0.7% in 2025 and 1.2% in the first half of 2026 [1][4].
The print business was the cash engine. Selling it cuts debt and interest but also removes cash flow, so the remaining software business has to earn its own way after the 2027 savings [7][4].
Stock pay fell to $7.5M in the first half from $13.7M a year earlier [4].
7. Capital: debt with a 6.75% spread, and a sale to cut it
The term loan had $236.3M outstanding at Mar 31, 2026, at SOFR plus 6.75%, maturing May 1, 2029 [3]. At Jun 30, 2026 the balance sheet carried $229.7M of term loan, our sum of current and long-term parts, plus $14.1M drawn on an asset-based revolver [4].
Cash was only $9.1M at Jun 30, 2026 [4]. Cash interest was $31.6M in 2025, down from $57.0M in 2023 [1].
The $142M print sale should cut that sharply if most proceeds repay the loan [7].
Other obligations matter at this size. Pension obligations were $41.4M at Jun 30, 2026, and the current portion of unrecognized tax benefits fell to $1.8M from $28.3M at year end [4].
Interest paid to the related party lender was $2.5M in the second quarter [4].
The balance sheet also carries $253.8M of goodwill, much of it from past deals, and $137.6M of deferred tax assets that depend on future profits [4]. A weaker software outlook could put both under review
8. The main risks are churn, a thin cash cushion and execution
Churn. Seasoned net revenue retention of 90% means existing clients shrink each year, and small businesses fail often [4][1].
Cash cushion. With $9.1M of cash and a revolver, Thryv has little room if the sale is delayed [4].
Execution. The Growth Platform launched in August 2026, the restructuring runs into 2027, and a services agreement keeps Thryv supporting the print business for Carolwood after closing [4][7].
Restructuring timing. About 10% of the $20M to $25M of charges had been taken by August, 40% falls in the second half of 2026 and 50% in 2027, so reported results stay noisy for over a year [4].
Competition. Wix, HubSpot and Google all sell tools to the same small businesses, and Thryv is now partnering with Wix rather than competing head on [4][1].
Retention. In January 2026 the board approved cash retention bonuses equal to half of salary for the president, the operations chief and the legal chief, paid in two parts through August 2027, a sign it feared departures
9. Management has run the company for twelve years
Joe Walsh has been chief executive since October 2014 and chairman since December 2021; he earlier ran Yellowbook and chaired Cambium Learning Group [3]. He owns 7.2% including options [3].
Chief financial officer Paul Rouse says the restructuring should lift margins and free cash flow [4]. President Grant Freeman, operations chief John Wholey and legal chief Lesley Bolger round out the named officers [3].
The cost base is shrinking. Thryv moved its headquarters to Grapevine, Texas, closed the former Keap headquarters and will not renew that lease when it ends on Dec 31, 2026 [1].
It also employs a workforce in the Dominican Republic [1].
10. Superintelligence tools are the new pitch
Thryv now describes its product as a growth platform with superintelligence running underneath, built to turn each lead into revenue and show return on marketing spend [4]. It launched the platform in August 2026 [7].
For small businesses, superintelligence agents could answer calls, book jobs and run ads, which is what Thryv sells, but the same agents will be offered free or cheap by Google, Wix and others [1]. Thryv's edge would have to be its 95 thousand paying relationships and local sales force, not the technology itself [4].
11. No controller, but Paulson is the largest holder and close to the board
Vote control. Each share has one vote and there is no controlling holder [3].
Paulson & Co. owned 8,443,835 shares, 19.0%, on its 13G [3]. Massachusetts Financial Services held 6.4% and BlackRock 6.2% on the proxy count, and BlackRock disclosed 7.0% in July 2026 [3][10].
Board ties. Lead independent director John Slater was a Paulson partner from 2009 to 2019 and has sat on the board since 2016, after an earlier stint from 2013 to 2015 [3].
He was most recently chief investment officer of Hum Capital [3].
Related parties. BlackRock, a holder of more than 5%, owns 40.0% of the term loan, so it is both a lender and a shareholder [3].
In 2025 Thryv paid $35.0M of principal and $28.3M of interest on the loan [3].
Buyout material. The print sale to Carolwood is the main deal; the buyer is not a related party
12. The questions we would ask management, and the calls we are tracking
Question one: how much of the $142M will repay the term loan, and what will interest cost after? [7]
Question two: when does software revenue grow again without acquisitions? [4]
Question three: why is seasoned retention only 90%, and what is the target? [4]
Question four: does Paulson want a sale of the whole company once print is gone? [3]
Question five: can the software business earn positive owner cash on its own in 2027? That answer sets our conviction.
Prediction one: third quarter 2026 software revenue is at or above $111.0M (70%). [4]
Prediction two: the print sale closes by Mar 31, 2027 (80%). [7]
Prediction three: term loan principal at Dec 31, 2026 is below $150M (55%). [3]
Bottom line: a cleaner company after the sale, but the software has to grow
Selling print for $142M fixes the balance sheet and leaves Thryv as a small business software company. But software revenue is flat to down, older clients spend less each year, and owner cash is near zero.
We would turn positive on two quarters of software growth and positive owner cash once the sale has closed.
Stock Profiles & Comparable Companies
THRY Stock Profile →More Published Research
View All →Mobility Global Owns the CARFAX Record and the Debt That Paid for the Spin
CARFAX grew 8% and no customer tops 10% of revenue, but the spin left $2.0B of notes and S&P Global's finance chief on the board.
Cerence Looks Fixed but a Samsung Patent Check Did Most of the Work
An in-car voice software supplier whose nine month growth falls to about 4% without one settlement, with no activist, a slipping pay vote and an ex-Intel board tie.