Financebotresearch desk研究台

Discovery › screen

Screen quality smallmid value

Date: 2026-07-29 Criteria: Convert the ring-2 finding (US small-cap value at 14.9x vs S&P 26.2x) into names. Profitable, low-debt, non-speculative. Framework: analysis_notes.md §0 + full §1 health check on finalists

⚠️ This is a different asset class from the Microstocks sleeve. Output/Microstocks/ already holds 24 pre-revenue / turnaround names evaluated under §6, where the default prior is "most of these fail." Nothing here is that. Every name below is profitable, real-revenue, and screened for balance-sheet quality. Conflating the two would be the easiest way to misuse this screen.

Universe swept: 33 names across business services, niche software, specialty industrials, specialty insurance and healthcare services. Market caps $1.3B–$7.1B.


Executive Answer

One clear winner, one high-torque second, and one name I nearly recommended before the health check killed it.

The screen's most useful output may be the rejection: OTEX looked like the cheapest thing in the entire sweep (5.8x forward, 4.44% yield) and is a value trap. Revenue declining, FCF declining and below where it stood four years ago, 48% debt/assets, and returning more cash than it generates. §0 caught it; the price alone would not have.


🟢 Shortlist

# Company Ticker Sector Mkt Cap Price %off High fwd P/E ttm Yield Target Upside
1 Dolby Laboratories DLB Technology (IP licensing) ~$5.1B $52.54 −31.5% 11.5x 20.8x 2.75% $78.33 +49%
2 Calix CALX Infrastructure (broadband) ~$2.5B $35.55 −50.1% 15.4x 46.8x $62.33 +75%
3 Gibraltar Industries ROCK Industrials ~$1.3B $43.10 −42.6% 9.0x 20.6x $72.50 +68%
Progress Software PRGS Technology ~$1.9B $42.71 −14.4% 6.9x 20.7x $53.80 +26%
ICF International ICFI Technology (govt) ~$1.5B $83.03 −18.4% 10.8x 17.9x 0.70% $105.25 +27%

Health Check (§1) — 4 fiscal years, newest first

Metric DLB CALX ROCK PRGS ICFI 🔴 OTEX
Revenue trend 1.35 ← 1.27 ← 1.30 ← 1.25B ✅ steady 1.00 ← 0.83 ← 1.04 ← 0.87B ✅ recovered 1.14 ← 1.02 ← 1.05 ← 1.39B ⚠️ −6.5% CAGR 978 ← 753 ← 694 ← 602M ✅ +17.5% CAGR 1.87 ← 2.02 ← 1.96 ← 1.78B ⚠️ declining 5.17 ← 5.77 ← 4.48 ← 3.49B 🔴 declining
FCF trend 430 ← 297 ← 337 ← 259M ✅ 4yr HIGH, +66% 116 ← 50 ← 38 ← 13M ✅ +106% CAGR (8.8×) 121 ← 157 ← 206 ← 83M ⚠️ halved 229 ← 206 ← 168 ← 186M ✅ rising 120 ← 150 ← 130 ← 138M ⚠️ declining 687 ← 808 ← 655 ← 889M 🔴 below 4yrs ago
Gross margin 88.1% 56.8% (rising) 26.8% ✅ 80.8% 37.2% 72.1%
Debt / Assets 1.2% ($39M) 1.5% ($16M) 4.2% ($58M) 🔴 58.1% 27.9% 🔴 48.2%
Equity / Assets ✅ 81% ✅ 81% ✅ 68% 🔴 19.4% 50% 28.5%
Net margin ✅ 18.9% 1.8% (trough) net loss ⚠️ 7.5% (interest drag) 4.9% 8.4%
Shares outstanding −4.4% ~flat ($94M buyback) −6.9% ~flat ✅ −2.7% ~flat
Div covered by FCF? Yes — 29% payout n/a n/a 🔴 Dividend eliminated ($31M → $0.8M) ✅ Yes (8%) ⚠️ Total returns $815M > $687M FCF

Rationale

1. DLB (Dolby) — the pick 🥇

The best business in the entire 33-name sweep, and it's on sale.

  • 88.1% gross margin — IP licensing, not manufacturing. Dolby Atmos and Vision are embedded standards across the content-production and device chain, which is a switching-cost moat by construction (§2: intangibles + switching costs).
  • Essentially debt-free: $39M against $3.23B assets (1.2%). Equity is 81% of assets. Invested capital equals equity — there is no leverage in the structure at all.
  • FCF at a four-year high of $430M, up 66% over three years while revenue grew only 2.5% — that's margin and working-capital discipline, not a revenue cycle.
  • Dividend covered at a 29% payout ratio, and buybacks have shrunk the share count 4.4% over three years. Total shareholder returns are 67% of FCF — returning cash without straining.
  • −31.5% off its high at 11.5x forward with a 2.75% yield and a +49% consensus target.

Honest knocks: revenue growth is slow (2.5% 3yr CAGR) — this is a royalty annuity, not a grower, so the return comes from the multiple, the dividend and the buyback rather than from expansion. The reported ROIC of ~10.7% understates the underlying economics because a large cash pile sits in invested capital. And licensing models carry concentration risk to a handful of device and studio partners.

Portfolio fit: genuinely new territory — capital-light IP licensing, in no existing correlation cluster (not AI capex, not cyber, not AI-services). Income-producing. Earnings July 30.

2. CALX (Calix) — the high-torque second

The sharpest inflection in the sweep. FCF went $13M → $38M → $50M → $116M — an 8.8× increase over three years, with a +106% CAGR — while revenue recovered to a fresh high of $1.00B off an $831M trough. Debt-free ($16M on $1.06B assets), 81% equity, gross margin 56.8% and rising, and it bought back $94M of stock last year.

Why it's cheap: operating income is only $21M on $1.0B of revenue — margins are at a trough, which is exactly why the trailing P/E reads 46.8x. If operating margin normalises toward the gross-margin structure, the earnings leverage is enormous. That's the bet, and it's why consensus carries a +75% target at −50.1% off the high.

Honest knocks: it's a hardware vendor to rural telcos and ISPs — lumpy order patterns, and the federal broadband-funding backdrop (BEAD) is politically variable. Thin trailing profitability means a stalled recovery leaves you holding a 46x stock. Higher risk than DLB; size accordingly.

3. ROCK (Gibraltar) — fortress balance sheet, cyclical trough

Debt/assets of 4.2% with 68% equity and a −6.9% share count over three years — an unusually clean balance sheet for a building-products cyclical. At 9.0x forward, −42.6% off its high, with a strong_buy consensus and +68% target.

But the fundamentals are genuinely soft: revenue is −6.5% CAGR and below where it stood four years ago, FCF has halved ($206M → $121M), gross margin is a thin 26.8%, and the most recent year shows a net loss (operating income stayed positive at $123M, so this is a below-the-line impairment). This is a §6-style turnaround wearing a strong balance sheet — the safety is real, the earnings recovery is a forecast.

Not recommended

PRGS — real growth (+17.5% revenue CAGR) and 80.8% gross margin, but 58.1% debt/assets with equity at only 19.4%, net margin compressed to 7.5% by interest expense despite $174M of operating income, and the dividend was eliminated ($31M → $786K) to service the debt. The 6.9x forward multiple is on adjusted numbers that step around the leverage. A leveraged acquisition roll-up, not a quality compounder.

ICFI — cheap at 10.8x with a decent balance sheet, but revenue is already declining and it's government consulting, carrying the same federal-budget/DOGE risk already flagged on BAH and correlated with owned ACN. The portfolio does not need a fourth expression of that bet.


🔴 Traps, Named

Ticker The lure The reality
OTEX 5.8x forward, 4.44% yield — cheapest in the sweep 🔴 Revenue declining (5.77 → 5.17B), FCF declining and below its level 4 years ago, 48.2% debt/assets from the Micro Focus deal, and returning $815M against $687M of FCF. "Hold" consensus, +12% target. This is the CAG/GIS/KHC pattern in software form: a fat yield as compensation for a shrinking business carrying acquisition debt.
FORR (Forrester) 11.3x forward, −11.4% off high 🔴 Target $6.00 is 42% BELOW the $10.26 price. No trailing earnings. Research/advisory — the same AI-disruption bucket that collapsed Gartner's contract value. Avoid.
NSP (Insperity) 4.56% yield 🔴 Target $43.75 vs $53.18 spot — 18% below. No trailing earnings.
KFY (Korn Ferry) 13.4x, 2.71% yield 🟡 Target $81 below the $84.69 spot; only −1.2% off high. Fine business, no discount.
PINC (Premier) 🔴 37.7x forward, 294.6x trailing, −0.4% off high.

Rejected on price / no discount (all within ~10% of highs or >20x forward): EXPO 23.9x · AIT 28.3x · WTS 25.5x · SXI 27.6x · KAI 22.2x · FELE 20.5x · MEDP 30.3x · AGYS 35.5x · DSGX 24.0x · HSTM 34.1x · USPH 23.4x · CRAI 19.7x · MWA 16.0x · MLI 14.9x · TNC 13.2x · IDCC 23.2x

Specialty insurers priced out: SKWD (11.4x but −2.2% off high, +4% target) and PLMR (12.8x, −3.2% off high) are good businesses that never sold off. Worth a revisit on a real drawdown — the sector has no exposure in either account.

No data returned: CSGS, CSWI, CRVL (incomplete forward estimates).


Portfolio Fit

Check Result
AI-capex cluster? ✅ None of the three
Cyber sleeve (already 4-deep)? ✅ None
AI-services cluster (ACN/CTSH/EPAM/IBM)? ✅ None — and this is why ICFI was cut
Overlaps Microstocks sleeve? ✅ No — all profitable, real revenue
New sub-sectors added IP licensing (DLB) · rural broadband infra (CALX) · building products (ROCK)
Income added DLB 2.75%, covered at a 29% payout

DLB in particular sits in no existing correlation cluster, which is rare at this point — most additions to this portfolio deepen a bet that's already on.


Recommendations

Add to Watchlist:

Ticker Sleeve Suggested entry Trim Conviction
DLB 🏛 Evergreen Compounders Attractive at $52.54; add $46-50; strong buy <$44 $75+ [—] pending /analyze
CALX 🔧 Re-Rating Plays Stage $32-36; strong <$29 $60+ [—] pending /analyze
ROCK 🔧 Re-Rating Plays Watch only — wait for revenue to turn [—]

Do not add: OTEX, FORR, NSP, PRGS, ICFI, PINC.

Next steps

  1. /analyze DLBmy recommendation. Best business found, in a zone, in a new sleeve, with earnings July 30 making it timely. The §2 moat work (how durable is an embedded audio/video standard against open codecs like AV1?) is the question that decides it.
  2. /analyze CALX — resolve whether the margin recovery is structural or a funding-cycle artifact before committing.
  3. /compare DLB CALX — if you'd rather pick one; they're not really substitutes though (annuity vs inflection), so two separate analyses probably serve better here.

Bottom Line

  1. DLB is the find — 88% gross margin, 1.2% debt/assets, FCF at a four-year high, share count −4.4%, a covered 2.75% yield, −31.5% off its high at 11.5x forward, and it lands in no correlation cluster the portfolio already owns.
  2. CALX is the higher-torque option — FCF up 8.8× in three years, debt-free, −50% off its high, +75% target — but trough margins mean a stalled recovery leaves you in a 46x trailing stock.
  3. The most valuable output was a rejection. OTEX screened as the cheapest name in the sweep (5.8x, 4.44% yield) and the §1 check exposed a shrinking business with 48% debt/assets returning more cash than it earns. Price alone would have bought it.
  4. The small-cap discount is real but thin on quality. Of 33 names swept, only three cleared both §0 tests — most quality small/mids never sold off (SKWD, PLMR, MLI, TNC, KFY all sit within 10% of highs), and most of the deeply discounted names are discounted for cause.
  5. Consistent with the two prior spirals: the cheap-looking tail is mostly traps, and the finds are one step off the obvious. DLB is not what "small-cap value" conjures — it's a high-margin royalty business that happens to be small.

Screen run 2026-07-29. 33 names swept, 6 taken to full §1 health check. Prices live; health data via .mcp/fin.py (4 fiscal years, newest-first). Follows the ring-2 finding in screen-global-vs-us-value-2026-07-29.md.