Good businesses at a low price

What I screened. All 1853 companies in Oracle, through the BEST PRICE view: only businesses the app already rates strong (quality 80 or better), then ranked by what you pay for them. 128 clear that bar. Then I applied two filters of my own, because a "quality and cheap" screen has two reliable ways of lying, and both of them were firing.

Data as of 2026-09-09 23:13. Every figure below is measured from the companies' own filings by the app; none of it is my recollection.

Your instinct is right, and the screen can prove it

You said you were hunting underdogs. The screen agrees — cheapness is concentrated at the small end, and it is not subtle:

The largest companies in the screen cost roughly twice the earnings multiple of the smallest. That is the whole reason your Alphabet question was a good one.

What I threw out first

Trap 1: the insurers, where cheap and good are the same fact

25 of the 128 are insurers, lenders, miners, tankers or brokers, and they dominate the top of the raw list:

Three of those — HCI, Heritage and Slide — are Florida homeowners insurers. For a property insurer, a low price-to-earnings ratio and a high return on equity after a quiet hurricane season are not two pieces of good news. They are one piece of good news counted twice: the same benign loss year inflates the earnings that make the multiple look low and the return that makes the quality look high. One bad season inverts both at once.

Hamilton and RenaissanceRe are Bermuda reinsurers with the same shape. Qfin is a Chinese consumer lender at 2.5× earnings, which is a governance and regulatory discount, not an oversight.

None of these are underdogs. They are bets on weather, credit and commodity prices, and they belong in the CYCLICAL view, not this one.

Trap 2: four share counts that do not reconcile

Market capitalisation is price times share count, and if the count is wrong the company appears cheap for no reason at all. I cross-checked every candidate's cover-page share count against the diluted count in the same filings; these four disagree, so I excluded them:

Booking is a false alarm — it split during the year, so a weighted-average diluted count legitimately disagrees with a point-in-time one. The other three need looking at before anyone trusts their multiples.

This check earned its keep. Yesterday HEICO showed a 12.7× P/E and Duolingo 6.2×; both were reading a live price against a stale share count — HEICO's was from 2015. Corrected, they are 65.1× and 16.9×. Two invented bargains, gone.

The seven that survive

"Price" is the app's valuation score, which is deliberately broader than P/E: it reads price against book value, against net cash, and as an earnings yield. A company can have a low P/E and still score modestly here if it owns little.

One at a time

YELP — the strongest case on this page

YELP INC · Personal Services · $1.1B · 7.7× earnings · 39% off its 12-month high · figures from the year to 2025-12-31

• return on equity 20.5% · return on assets 15.2% · return on capital employed 25.3%
• net margin 9.9% · margin expanding 6.1pp over the window · profitable in 5 of 5 years · operating income positive
• free cash flow $0.32B · FCF margin 22.1% · OCF/net income 2.56x · FCF positive in 5 of 5 years
• 3Y revenue CAGR 9.2% · grew in 4 of 4 years · revenue $1.46B
• debt/equity 0.02 · net cash $0.30B — fortress balance sheet
• P/B 1.59 · net cash 27% of market cap · earnings yield 12.9%

Twenty-seven percent of the market capitalisation is net cash, debt is effectively nil, and it has converted profit to cash at more than twice reported earnings. Strip the cash and the operating business is on roughly 5.6 times earnings while growing revenue every year for four years and expanding margin by six points. The market is pricing a slow decline that has not begun.

LRN (Stride) — down by half, still compounding

Stride, Inc. · Educational Services · $3.5B · 10.2× earnings · 50% off its 12-month high · figures from the year to 2026-06-30

• return on equity 20.7% · return on assets 13.9% · return on capital employed 22.0%
• net margin 13.4% · margin expanding 7.1pp over the window · profitable in 5 of 5 years · operating income positive
• free cash flow $0.43B · FCF margin 17.2% · OCF/net income 1.28x · FCF positive in 5 of 5 years
• 3Y revenue CAGR 10.5% · grew in 4 of 4 years · revenue $2.52B
• debt/equity 0.26 · net cash $0.54B — fortress balance sheet
• P/B 2.12 · net cash 16% of market cap · earnings yield 9.8%

Online K–12 schooling. A 50% drawdown against a business that grew revenue in four of four years and expanded margin by seven points. That gap between price and trajectory is what an underdog is.

NICE — software at ten times earnings with no debt

NICE Ltd. · Business Services (incl. Software) · $6.1B · 10.0× earnings · 34% off its 12-month high · figures from the year to 2025-12-31

• return on equity 15.8% · return on assets 12.0% · return on capital employed 16.7%
• net margin 20.8% · margin expanding 10.4pp over the window · profitable in 5 of 5 years · operating income positive
• free cash flow $0.70B · FCF margin 23.7% · OCF/net income 1.17x · FCF positive in 5 of 5 years
• 3Y revenue CAGR 11.3% · grew in 4 of 4 years · revenue $2.95B
• debt/equity 0.00 · no interest burden · net cash $0.42B — fortress balance sheet
• P/B 1.57 · net cash 7% of market cap · earnings yield 10.0%

Debt-to-equity of zero and no interest burden at all, with margin expanding more than ten points. Enterprise software rarely trades at ten times earnings unless the market believes the growth is finished.

DECK — the best business in the group

DECKERS OUTDOOR CORP · Rubber & Plastics · $11.3B · 11.0× earnings · 32% off its 12-month high · figures from the year to 2026-03-31

• return on equity 41.0% · return on assets 27.8% · return on capital employed 50.5%
• net margin 18.7% · margin expanding 4.4pp over the window · profitable in 5 of 5 years · operating income positive
• free cash flow $1.10B · FCF margin 20.1% · OCF/net income 1.15x · FCF positive in 5 of 5 years
• 3Y revenue CAGR 14.8% · grew in 4 of 4 years · revenue $5.47B
• P/B 4.51 · net cash 17% of market cap · earnings yield 9.1%

The highest quality reading here: 50.5% return on capital employed and margins still expanding. It is the least distressed of the seven, which is also why it is the least cheap on book value at 4.5×.

LULU — cheaper than Deckers, and here is the catch

lululemon athletica inc. · Apparel · $10.8B · 6.8× earnings · 55% off its 12-month high · figures from the year to 2026-02-01

• return on equity 31.8% · return on assets 18.7% · return on capital employed 44.6%
• net margin 14.2% · margin compressing 1.4pp over the window · profitable in 5 of 5 years · operating income positive
• free cash flow $0.92B · FCF margin 8.3% · OCF/net income 1.01x · FCF positive in 5 of 5 years
• 3Y revenue CAGR 15.4% · grew in 4 of 4 years · revenue $11.10B
• P/B 2.18 · net cash 8% of market cap · earnings yield 14.6%

Down 55%, on an earnings yield of nearly 15% and a 44.6% return on capital. On the surface a better bargain than Deckers. But read the margin line: LULU's is compressing, Deckers' is expanding. For a brand, compressing margin is the first sign of lost pricing power, and it is the single most important difference between these two.

If you look at one pair, look at this one. Nearly the same price, opposite margin direction.

TTD (Trade Desk) — the highest risk and the highest growth

Trade Desk, Inc. · Business Services (incl. Software) · $6.9B · 15.6× earnings · 75% off its 12-month high · figures from the year to 2025-12-31

• return on equity 17.8% · return on assets 7.2% · return on capital employed 23.7%
• net margin 15.3% · margin expanding 3.8pp over the window · profitable in 5 of 5 years · operating income positive
• free cash flow $0.80B · FCF margin 27.5% · OCF/net income 2.24x · FCF positive in 5 of 5 years
• 3Y revenue CAGR 24.7% · grew in 4 of 4 years · revenue $2.90B
• P/B 2.78 · net cash 19% of market cap · earnings yield 6.4%

Down 75% from its high, and yet the fastest grower in the group at nearly 25% a year with a 27% free-cash-flow margin. A fall that size is the market saying the growth is over, not that it was mispriced. This one is a judgement about competition in advertising technology, and the screen cannot make it for you.

CHKP (Check Point) — quality, cheap, and slow

CHECK POINT SOFTWARE TECHNOLOGIES LTD · Business Services (incl. Software) · $14.0B · 13.3× earnings · — off its 12-month high · figures from the year to 2025-12-31

• return on equity 36.7% · return on assets 13.5% · return on capital employed 28.8%
• net margin 38.8% · margin expanding 1.1pp over the window · profitable in 5 of 5 years · operating income positive
• free cash flow $1.17B · FCF margin 43.0% · OCF/net income 1.13x · FCF positive in 5 of 5 years
• 3Y revenue CAGR 5.9% · grew in 4 of 4 years · revenue $2.73B
• P/B 4.86 · net cash 31% of market cap · earnings yield 7.5%

A 38.8% net margin, a 43% free-cash-flow margin and net cash worth 31% of the company. Superb economics. It is also the only name here growing under 6% a year, which is very likely why it is cheap. This is the honest one: cheap for a reason you may find perfectly acceptable.

What this cannot tell you

• There are no forecasts in here. Oracle has no analyst estimates and no consensus feed. Everything above is history, and history is where cheapness hides but also where value traps hide.
• A 52-week high needs live price data. Check Point has no drawdown figure because no range came back for it.
• I excluded Exzeo (XZO) despite a quality score of 96 — its return figures have needed capping before, and I do not trust the reading.
• Cheap on assets is not the same as cheap on earnings. The app's price score leans on book value and cash, so an asset-light business will look expensive there even at a modest P/E.
• This is a screen, not a recommendation. It says where to point a microscope, nothing more.

Where the numbers come from

Every figure is computed by Oracle from SEC XBRL filings — the period each company's figures were taken from is printed beside its name, and the app stamps each one with the filing it came from. Prices and 12-month ranges come from Interactive Brokers delayed data. Share counts are cross-checked as described above.