Phreesia: $122.0 million of EBITDA, $2.3 million left for owners

Over the twelve months to July 31, 2026, Phreesia reported $122.0 million of adjusted EBITDA (earnings before interest, taxes, depreciation and amortization, after the company's own adjustments). Its free cash flow (FCF), the cash left after running the business and paying for investment, was $67.5 million on the company's definition. Then come two costs that definition leaves out: $59.4 million of stock-based compensation (SBC), which is paid in new shares instead of cash, and $5.8 million of principal on leased computer equipment. What remains for shareholders is $2.3 million. The cheap-looking stock on a screener and the expensive-looking stock on that last line are the same company.

At the October 9, 2026 close of $10.81, the shares trade within 1% of my triangulated Base value of $10.88. The Bear value is $7.07 and the Bull value is $14.57. The required margin of safety (MOS), the discount to fair value I need before buying, is 60%, so the maximum buy price is $4.35, about 60% below today's price. The question for this stock is not whether the restructuring is real. It is whether the price has already paid for it, and on my numbers it has.

Route: SF-07 software and technology. Subtype: hybrid healthcare IT platform, mature, low organic growth, recently free cash flow positive. Economic classification: Standard (3/12). Primary method: enterprise value (EV, the value of the equity plus net debt) as a multiple of next twelve months' adjusted EBITDA minus SBC. Secondary method: a ten-year discounted cash flow (DCF) model of unlevered owner free cash flow after SBC. Values are present fair value estimates, not 12 month price targets.

What Phreesia is

Phreesia sells the software a patient meets at check-in: intake forms, insurance and identity capture, consent and the payment screen. The company says it touched more than 180 million visits in calendar 2025, roughly one in six US visits. Three businesses run on that position. Subscription fees from providers were $52.7 million in the second quarter, 40.7% of revenue. Payment processing and, since November 12, 2025, patient financing through AccessOne brought $38.5 million, or 29.7%. Network solutions, which is marketing from pharmaceutical brands delivered to patients and clinicians at the point of care, brought $38.3 million, or 29.6%.

The fiscal year ends January 31, so fiscal 2027 (FY2027) runs from February 2026 to January 2027. The mix matters more than the total. Subscription revenue, the part a software investor would pay up for, fell 1.9% year over year in the second quarter because the company is deliberately moderating prices for providers under payer pressure. Network revenue grew 8.9%, but it depends on pharma brand budgets, and in March 2026 those budgets were the reason fiscal 2027 revenue guidance was cut from $545 to $559 million to $510 to $520 million, 6.7% lower at the midpoint.

AccessOne, bought for $153.2 million in cash, adds a financing layer that a pure software company does not have. On July 31 it carried $88.9 million of cardholder receivables at fair value against $147.3 million of unpaid principal, $89.5 million owed to providers, a $21.5 million deferred purchase price receivable and an off-balance-sheet securitization with PNC whose limit was raised to $300 million. Patient credit risk sits with providers, which have full recourse, and Phreesia is a joint and several co-guarantor of the servicer's obligations. AccessOne is expected to bring about $37 million of revenue in fiscal 2027, roughly 7% of the total.

The latest quarter

The second quarter looks like a turnaround on the headline lines. Revenue grew 10.4%, adjusted EBITDA rose by half and stock compensation fell by more than a quarter. Most of the revenue growth is AccessOne: on a pro forma basis, as if AccessOne had been owned all along, first half revenue grew about 1.5%.

Guidance for fiscal 2027 was maintained at $510 to $520 million of revenue and $125 to $135 million of adjusted EBITDA. The first half delivered $63.4 million of adjusted EBITDA, 47.0% to 50.7% of the range, so reaching the midpoint needs $66.6 million in the second half, 13.7% more than a year earlier. A restructuring of about 220 positions, roughly half of them contractors, started in May. Total charges are expected at about $10 million, of which $2.8 million was recognized in the second quarter.

Why EBITDA after stock compensation is the metric

The company's adjusted EBITDA adds back stock compensation, acquisition and restructuring costs and everything in other expense. Each add-back has a stated reason. Together they explain how $12.5 million of trailing GAAP operating income becomes $122.0 million of adjusted EBITDA. Depreciation and amortization account for part of that gap. Stock pay accounts for $59.4 million of it, and stock pay is not a one-off here: it was between $58.8 million and $71.6 million in every fiscal year from 2023 to 2026.

The table is the reason the company definition of FCF is not the value anchor. On a $671.0 million market value, the headline FCF yield is 10.1%. On the last line it is 0.3%. So I value Phreesia on EBITDA minus the full SBC charge, and on owner cash flow after SBC, leases, capital spending and cash taxes.

One item sits below operating income and inside the add-backs: the $2.8 million fair value loss on AccessOne instruments in the second quarter, against about zero in the first. The filing does not split it into interest rate and credit effects. Economically it is the cost of the receivable book, carried at $88.9 million against $147.3 million of principal. One quarter is not a run rate, so it is not deducted in the Base, but it is tested below and it is on the dashboard as red.

Trailing twelve months also understates the forward run rate, because it carries about $12.9 million of acquisition and restructuring cash costs, about $6 million of interest on the AccessOne bridge loan and SBC well above the current pace. The bridge from trailing to normalized owner cash flow:

Every step after the first is my estimate. Revenue growth contributes almost nothing. The forward cash comes from lower stock pay, cost cuts and one-offs fading, which is why the verdict hinges on those three things.

Shares, dilution and the balance sheet

The diluted count I use is 66.456 million: 62.071 million shares outstanding on August 27, 2026, 2.787 million restricted stock units, 1.423 million performance stock units at target and 0.175 million shares from in-the-money options under the treasury method. Weighted basic shares rose from 49.89 million in fiscal 2022 to 59.74 million in fiscal 2026, 4.6% a year, and cover-page shares rose 3.6% in the last twelve months. A buyback of up to 2.5 million shares was authorized in March 2025. None has been executed. An automatic shelf registration was filed on September 3, 2026, with no announced use.

Stock compensation is falling. It was $67.5 million in fiscal 2026, $25.4 million in the first half of fiscal 2027 against $33.5 million a year earlier, and $11.9 million in the second quarter. The second quarter was probably helped by forfeitures in the restructuring (377,299 restricted stock units forfeited in the first half against 1,067,797 granted), so the first half average of $12.7 million a quarter is the better signal. Unrecognized compensation of $46.9 million on restricted stock units and $19.7 million on performance units supports a run rate of $45 to $50 million, not a return to $59 million.

Because the full SBC charge is deducted and the diluted count already includes unvested units, the multiple carries no further dilution penalty. Using basic shares instead would give $11.65, which shows the size of that overlap.

The balance sheet is not the problem. On July 31 the company held $72.9 million of cash against $60.7 million drawn on a $275 million Capital One revolver at 6.4% and $4.6 million of finance leases. With $7.2 million of remaining restructuring charges added, adjusted net debt is -$0.40 million, a small net cash position. The $4.95 million ConnectOnCall data breach settlement is matched by a recorded insurance receivable of the same amount, so it nets to zero. Debt is 0.54 times trailing adjusted EBITDA. The federal net operating loss carryforward (NOL), past losses that can offset future taxable profit, was $587.2 million at January 31, 2026.

Ownership has one notable feature. Pale Fire Capital, a Czech investor, disclosed a 14.7% stake (8,924,329 shares, at an aggregate cost of about $103.5 million, roughly $11.60 a share) in its latest amendment, for an event dated March 31, 2026. Its filing states that it considers the shares undervalued and may engage with the board and management on capitalization, board composition and operations. No later amendment was found, so the current stake may differ. AQR Capital Management reported 5.1%.

Primary valuation: EV to EBITDA after stock compensation

The metric is next twelve months' (August 2026 to July 2027) adjusted EBITDA of $131.9 million minus a $50.0 million SBC run rate, which gives $81.9 million, a 15.75% margin on $520 million of revenue. The adjusted EBITDA figure is the second half guidance midpoint plus the first half of fiscal 2028 at 3% above this year's first half.

A peer multiple table was not part of this analysis, so the multiple is built explicitly and tested from 7x to 11x.

At 9.0x, EV is $737.1 million. Adding the $0.40 million of net cash and dividing by 66.456 million shares gives $11.10 per share.

Sensitivity: value per share

Read down the 9.0x column: $75.0 million of EBITDA after SBC gives $10.16 and $89.0 million gives $12.06. Every dollar of stock pay comes straight off that metric, which is why SBC matters as much as revenue here.

Secondary valuation: ten-year owner cash flow model

The second method values the cash conversion, the margin path and the tax shield explicitly. Revenue starts at $520 million and grows 2.5% in year 2, 3% in years 3 to 5 and 2.5% in years 6 to 10. The EBITDA margin after SBC rises from 15.75% to 19.0% by year 5 as restructuring savings land and SBC drifts toward 8% of revenue. From that I deduct capital spending and capitalized software of 4.7% of revenue, equipment leases of 1.0%, recurring other items of 0.5% and cash taxes of 1.5% while the NOL shelters federal tax. The terminal year is taxed in full at 24%. The discount rate is 10% and terminal growth 2.0%.

The ten years are worth $382.3 million today. The terminal value, on $64.6 million of fully taxed year 11 cash flow, is $807.2 million, or $326.4 million in present value, 46.1% of EV. Total EV is $708.7 million, which with net cash and the diluted count gives $10.67 per share, an implied 8.65x EBITDA after SBC. Removing the NOL entirely, with a 24% tax on operating profit every year, lowers this leg to $9.86, so the shield is worth about $0.81 a share here and $0.41 in the blended Base.

Sensitivity: value per share

A one point move in the discount rate shifts this leg by 11% to 14%. At 8% it reads $14.09, which is the most generous single input in the analysis and the clearest route to a higher value.

History and growth diagnostics

A history of the chosen metric barely exists: adjusted EBITDA was negative until fiscal 2024 and company FCF turned positive only in fiscal 2025. The 52 week range of $7.77 to $24.25, and a 53.5% fall over the year, describe how far sentiment moved, not a fair multiple. The historical layer gets no weight.

On reported earnings the stock is 64.9 times trailing GAAP earnings of $0.17 a share, which makes a price to earnings to growth (PEG) ratio meaningless. The non-GAAP PEG of 0.33 adds back SBC, which is 38% of next year's adjusted EBITDA, so it compares numbers on different bases and is not used. Growth durability is normalizing and acquisition affected: it comes from cost cuts, lower SBC and AccessOne, not from organic demand.

Free cash flow moved from -$116.3 million in fiscal 2023 to $54.4 million in fiscal 2026, so a growth rate on it is not meaningful. Per share, company FCF was $1.081 over the trailing twelve months, while the hardest measure after SBC and leases was $0.037. My normalized forward figure is $0.557 a share, a 5.5% yield on the market value, against 13.7% on the company definition. The quality score for that FCF growth is 30/100 and its reliability is low, because the base is tiny and the bridge is mostly estimates. None of these diagnostics sets the Base, the MOS or the verdict.

Valuation triangulation

The two methods land 3.9% apart. That looks like strong agreement, but both start from the same $81.9 million of EBITDA after SBC, so the convergence shows the model is internally consistent, not that the starting point is right.

What is the market already pricing in?

At $10.81, the market EV is $718.0 million. Running both methods backwards on the same net cash and share count shows what that price assumes:

The market is pricing almost exactly my Base: SBC down to about $50 million, guidance met at the midpoint, the margin after SBC up about 3.5 points over five years and no recurring credit cost from AccessOne. The 5.4x adjusted EBITDA multiple that makes the stock look cheap only holds if you treat $50 million a year of stock pay as free. To find a discount, you need to believe in more than the restructuring: a year 5 margin of 23.0%, NTM adjusted EBITDA of $141.6 million or more, or a control buyer.

That last route is the strongest bull argument. A private equity buyer would replace equity pay with cheaper cash pay and capitalize adjusted EBITDA before SBC: at 8x that is $15.89 a share, at 10x $19.85. A control price is not intrinsic value to a minority shareholder without an offer, so it is not in the Base.

Bear, base and bull scenarios

Three of the six halves in the next twelve months are bounded by guidance, so the scenarios move mostly through the first half of fiscal 2028, SBC and the multiple.

The Bear lowers both EBITDA and the multiple because they share a cause: payer pressure on subscription pricing and weaker pharma budgets cut earnings and the durability of the business at the same time. It does not also raise the discount rate. The Bull needs first half fiscal 2028 revenue growth of 4.6% against about 1.5% organic today, which relies on the network rebound management describes. I assign no probabilities to the three cases.

The scenarios are not the whole downside. These tests change one assumption at a time, with the verdict band for HOLD running from a Base of $9.19 to $12.43:

The bear case on trailing numbers is the one to take seriously. Using trailing adjusted EBITDA of $122.0 million, SBC of $59.4 million and no margin improvement gives $6.64. The reason it is not the Base: first half SBC of $25.4 million and the unrecognized compensation schedule point to $45 to $50 million, adjusted EBITDA guidance has been held three times through a revenue cut, the 220 position restructuring is done, and net cash with $214 million of undrawn credit rules out financial stress. Each of those can still fail, which is why the verdict is hold and not add.

How the buy price is set

A Confidence score of 13/25 is the lowest grade C, one point above D, which puts the base MOS at 50%, the top of the C range. I add 10 points for accounting and valuation opacity: AccessOne cash flows that sit outside FCF, fair value losses that sit outside adjusted EBITDA, the off-balance-sheet securitization guarantee and two securities class actions with no loss estimate. Nothing is added for leverage, which is low. The required MOS is 60%, the framework's practical maximum, so the maximum buy price is $10.88 times 0.40, or $4.35.

Today's quote is 2.48 times that price, and I treat it as indicative rather than a target to wait for. At this level of uncertainty, the honest position for new money is the watchlist. A cheaper multiple on adjusted EBITDA does not shrink the uncertainty, because the uncertainty is in the adjustments.

Dividend and capital return

Phreesia has never paid a dividend, and its revolver restricts dividends. The 2.5 million share buyback authorized in March 2025 has not been used while the share count grew 3.6% to 4.6% a year. The only capital return in the filings is debt reduction: $31.5 million of revolver repayment in the first half of fiscal 2027.

Reasons to own PHR

• A hard position to replicate. Phreesia sits in the check-in workflow of roughly one in six US visits. Average healthcare services clients grew 6% to 4,744 and revenue per client 4%, and that reach is what pharma marketers pay for.
• Operating leverage has shown up. Adjusted EBITDA went from $36.8 million in fiscal 2025 to $122.0 million over the trailing twelve months, and quarterly SBC fell from $16.2 million to $11.9 million.
• A clean balance sheet. Debt fell from $90.6 million to $60.7 million in six months, the AccessOne bridge was refinanced into a $275 million revolver at 6.4%, and the ConnectOnCall settlement is covered by insurance.
• A large tax shield. $587.2 million of federal NOLs shelter federal tax for years, worth about $0.41 a share in the Base.
• A 14.7% holder with a stated view. Pale Fire Capital's filing calls the shares undervalued and reserves the right to engage on capitalization and the board. A control transaction at 8x adjusted EBITDA before SBC would be worth $15.89 a share, and it is not in the Base.

What could go wrong?

• Thesis breaker: stock pay does not fall. At the trailing $59.4 million the Base is $9.39, inside HOLD by $0.20. At $60.7 million or more it becomes Slightly Overvalued.
• AccessOne credit cost. The $2.8 million second quarter fair value loss is excluded from adjusted EBITDA. Recurring at the second quarter pace of $11.2 million a year, the Base is $9.11. Combined with SBC at the trailing level, even $5.6 million a year gives $8.50.
• Pharma budgets and payer pricing. Brand-specific and regulatory dynamics in pharma already forced one guidance cut. Payers squeeze provider budgets, and subscription revenue is down 1.9%.
• The off-balance-sheet layer. Phreesia co-guarantees the servicer obligations in a $300 million PNC securitization and must keep $50 million of minimum liquidity. A servicing failure or provider defaults would hit cash.
• Legal and dilution. Two securities class actions filed in May and July 2026 have no loss estimate. Dilution runs at about 4% a year with an unused buyback and a new automatic shelf in place.

Management execution

The record has both directions in it. Fiscal 2026 adjusted EBITDA guidance was raised from $87 to $92 million to $99 to $101 million, and the company delivered $101.5 million. The fiscal 2027 adjusted EBITDA range of $125 to $135 million has been held through three updates. Against that, fiscal 2027 revenue guidance was introduced in December 2025 at $545 to $559 million and cut to $510 to $520 million four months later. The adjusted EBITDA definition widened to add back acquisition costs from the third quarter of fiscal 2026 without restating earlier periods, the buyback has not been used while the share count rose, and two securities class actions allege misleading statements about the business and outlook. The allegations are unproven. Assessment: 8/20, mixed. The next test is the third quarter of fiscal 2027, reported in early December 2026, with the second half EBITDA run rate, SBC and any further fair value loss.

Stock Analyza scorecard

Revenue grew 22.5% a year from fiscal 2022 to fiscal 2026, which would mechanically score 2, but that growth is acquisition affected and organic growth is now about 1%, so it scores 1. Return on invested capital is about 3.4% before tax on $371.5 million of capital. Subscription revenue is 40.7% of the total, below the 50% needed for the recurring revenue points.

Business quality is 35/100: moat and pricing power 8/20 (wide reach, but subscription pricing under pressure and subscription revenue falling), return versus cost of capital 4/20, balance sheet 9/15 (low leverage, offset by the securitization guarantee and liquidity covenant), cash flow quality 5/15, growth and runway 4/15, management and capital allocation 4/10, dilution, governance and alignment 1/5. The valuation score is 47/100: discount to Base 20/40 (the price is at Base), protection against the Bear case 6/20 (the Bear value is 34.6% below the price), agreement between methods 9/15 (close, but on a shared anchor), historical valuation 5/10 (neutral, no usable history), reverse valuation 5/10 (the market and the Base assume the same margin), data and model quality 2/5.

Confidence score

The filings are complete and reconcile, but the input that decides the value, forward EBITDA after SBC, is an estimate, and the company gives no FCF guidance. Predictability is low after a 6.7% revenue guidance cut and with network revenue tied to pharma budgets. Accounting transparency scores 2 because the adjusted EBITDA definition widened, fair value losses sit outside it and part of the AccessOne cash flows sit outside FCF.

What I would watch from here

Today two lines are red. The share count is up 3.6% year over year against a 3% red line, and the $2.8 million AccessOne fair value loss in the second quarter is above the $2.7 million quarterly line, which is the pace at which it would flip the verdict on its own. SBC is yellow: the second quarter's $11.9 million was green, but the first half averaged $12.7 million and was likely helped by forfeitures. Owner FCF after SBC and leases is yellow at $2.3 million, subscription revenue is yellow at -1.9%, adjusted EBITDA against the guidance midpoint is yellow until the second half delivers, and the class actions are yellow. Network revenue (+8.9%), revenue per client (+4%), net cash at July 31, the insured ConnectOnCall settlement and the securitization are green. The ConnectOnCall final approval hearing is on November 17, 2026.

The three most important thesis breakers

1. NTM adjusted EBITDA at or below $121.2 million, an SBC run rate at or above $60.7 million, or recurring AccessOne fair value losses at or above $10.7 million a year. Each alone takes the Base below $9.19. The milder versions compound: SBC at the trailing $59.4 million plus fair value losses of $5.6 million a year gives $8.50.
2. A securitization or legal event creating a cash outflow above about $113 million, which takes adjusted net debt above $112.3 million and the Base below $9.19. Fair value losses on the receivable book would usually show up first.
3. On the upside, fiscal 2028 guidance implying NTM adjusted EBITDA of $141.6 million or more with SBC at or below $50 million, or a control offer, would move the stock toward Undervalued and require a new valuation.

Adversarial review

Inputs come from Phreesia's Form 10-Q for the quarter ended July 31, 2026, the Form 10-K for fiscal 2026, the quarterly 8-K earnings releases, the restructuring 8-K of May 11, 2026, SEC XBRL data for the multi-year series, and the Schedule 13D filings of Pale Fire Capital and the 13G of AQR. The insurance receivable against the ConnectOnCall settlement and the $2,812 thousand fair value loss were read directly from note 11 and the cash flow statement of the 10-Q. The closing price was confirmed by two independent sources. Every table was recomputed independently, and the key figures reproduce to the cent.

The values marked as estimates are: NTM adjusted EBITDA of $131.9 million and revenue of $520 million beyond the guided half; the $50 million SBC run rate; the 10.0x reference multiple and the resulting 9.0x; the 10% discount rate and 2.0% terminal growth; the margin path to 19.0% after SBC; cash taxes of 1.5% of revenue during the NOL period; the normalization of working capital and securitization timing; and the timing of the remaining $7.2 million of restructuring cash. Not available in the filings or sources used: a peer multiple table, a synchronized history of EV to EBITDA, recurring revenue metrics such as annual recurring revenue and net revenue retention, AccessOne's standalone results for fiscal 2027, the split of the fair value loss between rates and credit, any estimate of class action exposure and any Pale Fire filing after early April 2026. The federal NOL figure is taken from the 10-K tax note.

The weakest point of the model is that both methods start from the same $81.9 million of EBITDA after SBC, and that figure rests on two things that have not yet happened: SBC settling at $50 million and second half EBITDA meeting the guidance midpoint. The 10.0x reference multiple has no external anchor. The verdict is also asymmetric in where it breaks: SBC staying at the trailing level leaves HOLD by only $0.20, while a discount rate of 8%, inside the tested range, would read Undervalued at $12.59.

Arithmetic: 81.902 times 9.0 is 737.12, plus 0.40 divided by 66.4555 million shares is $11.098; the DCF gives an EV of 708.72 and $10.671; 0.5 times 11.098 plus 0.5 times 10.671 is $10.884; $10.884 times 0.40 is $4.354. The publication gate status is PASS WITH WARNING: the balance sheet, share count and cash flows reconcile to the filings, and the value rests on an estimated SBC run rate, an estimated forward EBITDA and an estimated multiple.

Final verdict: hold at $10.81

Phreesia has done the hard part of a turnaround: adjusted EBITDA went from $36.8 million in fiscal 2025 to $122.0 million over the last twelve months, on a definition that has since widened, debt is down, and stock pay is finally falling. What it has not done yet is turn that into cash for shareholders, $2.3 million over the last twelve months. At $10.81 the market already pays for SBC of about $50 million and a 3.5 point margin gain, which is my Base.

Verdict: HOLD at $10.81. Holders: hold, do not add. New money: watchlist. Triangulated Base value $10.88, Bear $7.07, Bull $14.57, Confidence C (13/25). Required margin of safety 60%, maximum buy price $4.35.

The cost cuts are real and the price already knows about them. What nobody knows yet is whether $50 million of stock pay is a run rate or one good quarter of forfeitures.

Sources

Primary sources include Phreesia's Form 10-Q for the quarter ended July 31, 2026 (balance sheet, share count, stock compensation, AccessOne balances, the fair value loss, the ConnectOnCall settlement and insurance receivable, litigation and the securitization), the second quarter fiscal 2027 earnings release of September 2, 2026 (guidance, adjusted EBITDA, client metrics and restructuring charges), the Form 10-K for fiscal 2026 (multi-year cash flows, the NOL and the tax note), the first quarter fiscal 2027 release, the fourth quarter fiscal 2026 release and the earlier quarterly releases on SEC EDGAR. Ownership comes from Pale Fire Capital's Schedule 13D and its amendments, and AQR's Schedule 13G of August 12, 2026. The closing price comes from StockAnalysis and Yahoo Finance.

The reference market price is the October 9, 2026 regular session close of $10.81. Balance sheet figures are as of July 31, 2026, and shares outstanding are as of August 27, 2026. Third quarter fiscal 2027 results were not yet available. The forward EBITDA, the SBC run rate, the multiple, the discount rate, the margin path and the tax rates are analytical estimates, not company guidance.