Financebotresearch desk研究台

CSGP › analyze

CSGP · Analyze

ACCUMULATE Technology

Sleeve context: Re-Rating Plays sleeve-vacancy nomination. CSGP is down −66% from its 52-week high ($90.72 → $29.49; all-time high $101.05), the largest drawdown surfaced in the screen. The Re-Rating test requires a specific, fixable problem plus a gap-closing metric and date — not just "cheap." This report tests whether CSGP clears that bar.

Knowledge base check: python .mcp/kb.py find CSGP and find "data monopoly" returned no matches — nothing previously distilled on this name or field. Researched fresh. Pitfall check flagged one live, directly-applicable trap (below).

0. Data-quality flag before anything else

[[pitfall-yahoo-snapshot-fcf-field-diverges-from-quarter-sum]] fires on CSGP. Yahoo's snapshot freeCashflow field reads $412.7M (TTM). Reconciling from the same vendor's own quarterly cash-flow statements (mechanism-1 risk flagged in that pitfall: "whenever a large M&A quarter sits in the trailing four, recompute — never trust the snapshot field for an acquirer" — CSGP closed the ~$1.6B Matterport acquisition in Q3 2025, inside the trailing four quarters):

Quarter OCF Capex FCF (quarter's own field)
Q3 2025 $67.9M −$91.5M −$23.6M
Q4 2025 $162.1M −$66.5M $95.6M
Q1 2026 $152.0M −$54.0M $98.0M
Q2 2026 $115.0M −$58.0M $57.0M
TTM sum $497.0M −$270.0M $227.0M

Reconciled TTM FCF is $227.0M — 45% below the $412.7M vendor snapshot. Every FCF-based multiple below uses the reconciled $227.0M, not the snapshot field. This is a new instance of a known pitfall; the orchestrator may want to add CSGP to that note's ticker list.

Phase 1 — Fundamentals Analyst

Free Cash Flow — the backbone

Reported annual FCF looks broken on a straight CAGR read: $384.9M (2022) → $347.0M (2023) → −$245.0M (2024) → $41.0M (2025) → 3yr CAGR −52.6%. Taken at face value this screens as a collapsing business. It is not — it is a business that chose to spend its cash machine's output on two large, named programs:

  1. Homes.com residential build-out — net investment (company's own metric: revenue less directly attributable/allocated costs) ran ~$850M in FY2025 alone.
  2. Foundry Park, Richmond campus — a $460M+ owned real-estate construction program, breaking ground Nov 2022, expected substantially complete in H1 2026. This is a separate, finite capital program from Homes.com and explains part of the capex spike (capex jumped to $638M in FY2024, 23% of revenue, vs 4-6% in 2021-2023) that is independent of the residential thesis and should structurally roll off after H1 2026.

TTM (reconciled) FCF is $227.0M against TTM revenue $3,555.5M = 6.4% FCF margin — still well below the 2022-2023 norm (~15-18%), but a real, already-in-progress recovery from FY2025's 1.2% full-year margin. This is the honest read on §1: reported FCF is depressed by two overlapping, mostly-discretionary capital choices, one of which (Foundry Park) is already ending on its own schedule, and the other (Homes.com) has a management-committed, dated de-escalation plan (Phase 2/Sentiment, below).

Capital allocation (FY2025)

Use Amount Note
Operating cash flow $430M
Capex −$389M Homes.com + tail of Foundry Park
M&A (Matterport + other) −$2.35B funded from balance-sheet cash, not FCF
Buybacks −$500M
Debt ~flat minimal issuance/repayment
Net effect on cash $4.68B → $1.63B

Flag: FY2025 ran three simultaneous cash uses (Homes.com opex, a $2.35B acquisition, and a $500M buyback) against near-zero FCF, funded entirely by drawing down a cash pile built in the pre-Homes.com years. This was survivable only because the balance sheet started essentially unlevered (net cash). It is not a repeatable pattern — the cushion is now thinner (cash $1.27-1.63B vs debt $1.15-1.18B, roughly net-cash-neutral) than the ~$3.5B net cash position a year ago. Not a solvency risk (Debt/Assets 11.2%, current ratio 2.2x) but a reduced margin for a fourth simultaneous bet.

H1 2026 correction already visible: buyback pace continued ($505M Q1 + $82M Q2) while diluted share count fell from 420.7M (FY2025 avg) → 414.0M (Q1'26) → 404.4M (Q2'26) — a real, fast reversal, consistent with the new $1.5B authorization being deployed opportunistically into the depressed price.

Revenue, margins, per-share

Metric 2022 2023 2024 2025 TTM (Q3'25-Q2'26)
Revenue $2.18B $2.46B $2.74B $3.25B $3.556B
Revenue YoY — +12.8% +11.4% +18.6% +18% (Q2'26)
Gross margin 81.2% 79.7% 79.6% 78.8% ~79%
Op margin (GAAP) 20.7% 11.5% 0.2% −2.2% improving (Q2'26 op margin 8.2%)
Net income $369.5M $375.0M $139.0M $7.0M —
Diluted EPS $0.93 $0.92 $0.34 $0.02 $0.18 (ttm)

Revenue CAGR is 14.2% over 3 years and accelerating (18.6% in FY2025, "61st consecutive quarter of double-digit revenue growth" per the Q2 2026 release — a 15+ year unbroken streak). Gross margin has drifted down ~2.4pp in three years (mix-shift toward lower-margin residential, not core-CRE compression — see Moat). Operating margin collapse tracks the Homes.com ramp almost exactly, and is recovering: Q2 2026 GAAP EBITDA was $161M vs $85M (Q1'26), $130M (Q4'25), $18M (Q3'25), $39M (Q2'25) — a clean sequential-improvement trend, not just an adjusted-metrics story.

Revenue/share (FY2025): $7.72. FCF/share (FY2025, as reported): $0.10 — depressed by the same choices above; on reconciled TTM FCF ($227.0M ÷ ~410M shares) ≈ $0.55/share.

Debt/Assets: 11.2%. Low leverage, no balance-sheet risk currently.

Shares outstanding trajectory

Rose from 397.75M (2022) → 420.7M (2025, SBC/Matterport-related dilution during the growth-investment phase), now reversing fast under the new buyback: 420.7M → 414.0M (Q1'26) → 404.4M (Q2'26). This is the single cleanest proof that management can and does flip capital allocation levers when it chooses to — directly relevant to the discretionary-vs-structural question below.

Phase 1 — Sentiment & Intelligence Analyst

The activist campaign is the actual origin of the "fixable problem"

This is not a self-initiated turnaround story — it was forced:

  • 2024: Homes.com launches (Q1), backed by a publicized ~$1B marketing campaign.
  • FY2025: Homes.com net investment reaches ~$850M for the year.
  • 2026-01-27: Activist hedge fund Third Point (Daniel Loeb) sends a public letter calling the residential push a "fiasco" enabled by a "feckless board" and a "quixotic quest" by founder-CEO Andy Florance.
  • Early 2026: D.E. Shaw joins the pressure, stating that by end of 2026 CoStar will have spent more than $3B on Homes.com, and explicitly calling for CoStar to divest or shut it down.
  • Response (~Jan 2026, 8-K): CoStar commits to cut Homes.com net investment by more than $300M in 2026 (from ~$850M in 2025), $100M+/year further cuts through 2030, targets Homes.com revenue exceeding expenses exiting 2029 and positive Homes.com adjusted EBITDA in 2030. Board also authorizes a new $1.5B buyback after completing the prior $500M program, plus governance and compensation-alignment changes.
  • 2026-04-13: Third Point drops its board fight and exits its entire stake — read as the activist judging the commitment credible enough to walk away rather than escalate further. D.E. Shaw's current position is not confirmed by this research and should be treated as an open monitoring item, not resolved.

Q2 2026 print (2026-07-28) — real progress, real market skepticism

  • Revenue $925M (+18% YoY); net income +817% YoY; GAAP EBITDA +441% YoY.
  • Commercial segment: $481M revenue (+8% YoY), $172M adj. EBITDA (+7% YoY), 36% adj. EBITDA margin — stable, healthy, unremarkable-in-a-good-way.
  • Residential segment: $444M revenue (+33% YoY), $12M adj. EBITDA — its first profitable quarter since Homes.com launched in Q1 2024.
  • Full-year adj. EBITDA guidance raised to $780-820M (up ~$30M at the midpoint from the February outlook); adjusted EPS guidance raised to $1.32-1.39.
  • Tension: despite the "profitability inflection" framing, secondary coverage describes full-year revenue guidance being trimmed alongside the EBITDA raise, attributed to a 21% cut to the Homes.com salesforce, Ten-X restructuring, and Apartments.com "price integrity" (walking back promotional pricing). The two figures were not fully reconcilable from public secondary sources in this pass — flagging rather than resolving. What is unambiguous: a wave of sell-side downgrades landed the next day (2026-07-29): Citigroup Buy→Neutral (PT $70→$33), KBW Outperform→Market Perform (PT $41→$29), Baird Outperform→Neutral (PT $48→$34), William Blair Outperform→Market Perform. The Street's own revealed view, one day after the "inflection" quarter, was more skeptical, not less — a genuine tension this report does not wave away: the market is worried the margin gain is coming from trading away growth, not from pure efficiency.

Insider activity — a golden flag

Date Insider Role Action Price
2026-08-04 Andrew Florance CEO Buy 83,300 sh $29.89
2026-05-01 Andrew Florance CEO Buy 71,430 sh $35.17-35.82
2026-02-27 Andrew Florance CEO Buy 55,720 sh $44.52
2026-03-02 Frederick Saint Officer Buy 20,000 sh $45.33
2026-08-05/07 Hill, Cann Director/Officer Sell (small, ~$100-150K each) $30-31

Founder-CEO Florance bought stock four separate times in 2026 as the price fell from the mid-$40s to under $30 — roughly $7.5M of open-market purchases at successively lower prices. This is a meaningfully stronger and more sustained insider-conviction signal than a single opportunistic buy, and it directly overlaps the period of activist attack and guidance uncertainty — he was buying through the controversy, not after it resolved. The only insider selling in the same window was two small officer/director sales an order of magnitude smaller in dollar terms.

Analyst coverage

Average rating 2.14/5 ("Buy"), 20 analysts, mean target $37.30 (+26.5% from $29.49), range $25-$53 — a wide dispersion that itself reflects genuine, unresolved uncertainty rather than consensus complacency.

Phase 2 — Moat Analyst

Quantitative base

  • Gross margin ~79%, essentially stable (down only ~2.4pp in 3 years, attributable to residential mix, not core erosion).
  • Commercial segment adj. EBITDA margin 36%, growing in line with segment revenue (+7% EBITDA vs +8% revenue YoY) — the core is not decaying.
  • ROIC (return_on_inv_capital, roic.ai): −18.4% (FY2025) vs +3.9% (FY2024) — the headline number looks like moat collapse, but this is mechanically the Matterport goodwill add ($4.94B, up from $2.53B) plus depressed consolidated operating income landing in the denominator/numerator of a blended, company-wide ROIC. It is not evidence the core CRE data business's returns on capital have deteriorated — it is evidence that a large, recent acquisition and a deliberate residential investment are diluting a blended metric. Treat blended ROIC as uninformative here; the segment-level EBITDA margins are the cleaner read.
  • "61 consecutive quarters of double-digit revenue growth" (company disclosure, Q2 2026) is the single best piece of quantitative moat evidence in this report — a 15+ year unbroken record survives multiple recessions, rate cycles, and the entire post-2022 CRE downturn.

Adversarial stress-test

Core commercial data business (CoStar Suite, LoopNet, Ten-X, CoStar Sales/Leasing/Tenant): "You are a well-funded rival — how easily do you displace this?" Hard. CoStar's moat here rests on (1) a large, labor-intensive proprietary research/verification operation that has taken decades to build — expensive and slow for a rival to replicate even with capital, because the data is not simply scrapeable; (2) deep switching costs — CoStar data is embedded in the daily workflows of brokers, appraisers, lenders, and government agencies; (3) network effects (more verified listings attract more subscribers, which funds more research); (4) a demonstrated history of outlasting a well-funded direct rival — Xceligent, a venture-backed competitor that tried to build a rival CRE database, collapsed into bankruptcy in 2017 amid data-scraping litigation with CoStar, evidence the market structurally supports only one dominant player at scale (an efficient-scale moat). Live disruption vector worth naming: AI/LLM-driven data aggregation could lower the cost of building "good enough" alternative CRE data over a 5-10yr horizon — CoStar's rising R&D spend (+84% since 2022, to $406M in 2025) is the correct defensive response, not yet a proven one.

Residential (Apartments.com + Homes.com) vs. Zillow: This is the opposite kind of market. Zillow is the entrenched, brand-genericized incumbent with 247.7M monthly visits vs. Homes.com's 34.8M (≈7x), and Realtor.com beats Homes.com on engagement (visits/user) as well. CoStar's own earlier claim of Homes.com traffic parity/leadership (104M monthly unique visitors, cited on the Q1 2025 call) is hard to reconcile with the more recent, independently-sourced July 2026 traffic data and should be treated skeptically rather than taken at face value — a possible instance of a company-reported KPI diverging from third-party measurement (a pattern this playbook has flagged before in other names — see pitfall-third-party-panel-contradicts-issuer-disclosure). This segment does not have a data-monopoly moat. It is a late (2024) entrant into a winner-take-most consumer marketplace against a 15+ year incumbent with genuine network effects of its own. Apartments.com is the more defensible piece within Residential (an established, profitable rental-listings subscription business, not itself in an open war with Zillow) — the segment's Q2 2026 swing to +$12M adj. EBITDA is most plausibly Apartments.com profit outweighing a shrinking Homes.com loss, not evidence Homes.com itself has found its footing (company guidance explicitly puts Homes.com's own EBITDA breakeven at 2030, not now).

Evergreen assessment

Core commercial data segment: yes, a durable, close-to-monopoly cash machine — 36% segment margins, 79% gross margins, a 61-quarter growth streak, and a moat that has already outlasted one well-funded direct challenger. Residential (specifically Homes.com): unresolved. It may become a permanently profitable #3 portal by 2030 as guided, or it may prove that no amount of disciplined spending closes a 7x traffic gap against an incumbent with its own network effects — in which case the "fix" is not a margin recovery but an eventual write-down/exit, exactly what D.E. Shaw argued for. Both outcomes are consistent with everything observed so far; this is not yet decided.

Phase 2 — Valuation Analyst

Which models apply

No dividend (yield 0%, payout ratio 0%) — DYT and DDM are N/A. Trailing GAAP EPS ($0.18) is depressed by deliberate opex choices and acquisition-related amortization, not a proxy for owner earnings — Graham's Intrinsic Value (√(22.5 × 0.18 × 19.59) = $8.91) is not meaningful here and is reported for completeness only, not weighted. This is a case for FCF/EBITDA-based and segment (SOTP) approaches, per the framework's guidance for high-growth/distorted-earnings names.

Sum-of-the-parts, explicitly stated assumptions

CoStar does not disclose a clean Apartments.com-vs-Homes.com split, so this SOTP is necessarily approximate — stated as a range, not a point estimate.

Case Commercial value Residential value (net of remaining Homes.com burn) + Net cash Implied equity value Per share (~406M sh)
Bear/floor $700M EBITDA × 15x = $10.5B ~$0 (market currently implies zero-to-negative) +$0.3B ~$10.8B ~$27
Base $700M EBITDA × 20x = $14.0B +$1.5B (Apartments.com profit, less PV of remaining committed Homes.com losses through 2029) +$0.3B ~$15.8B ~$36-38
Bull $700M EBITDA × 25x = $17.5B (data-monopoly comp multiple: MSCI/Verisk/FactSet trade 18-35x EBITDA) +$2.0B (Homes.com proves durable) +$0.3B ~$19.8B ~$48-49

The base case (~$36-38) lands close to the independent analyst mean target ($37.30) — a useful cross-check that this SOTP is not an outlier. Current price ($29.49) sits at/just above the bear-case floor and meaningfully below the base case — consistent with a market that is currently pricing something close to "Residential is worth nothing," while giving no credit for Apartments.com's standalone profitability or for the possibility the committed spend-cuts hold.

Forward multiple cross-check

Trailing P/E (163.8x) and trailing EV/EBITDA (30.1x) are artifacts of the deliberately depressed base-year earnings and are not meaningful entry signals either way. Forward P/E is 17.0x (price $29.49 ÷ consensus FY2027 EPS $1.73) — cheap for a business with an 79%-gross-margin core still compounding at double digits, if the earnings recovery materializes on the guided path. EV ÷ FY2026E adj. EBITDA (using the $780-820M guided range) ≈ 14.4-15.2x — reasonable, not demanding, for the quality of the core asset.

Bogle's Expected Return (qualitative, no dividend)

Entirely a function of earnings-growth recovery ± multiple change, since yield is zero. If the guided EPS trajectory (adjusted EPS $1.32-1.39 FY26 → forward consensus $1.73 FY27, ~25-30% growth) is delivered and the multiple holds or re-rates modestly, expected return is strongly positive. If the multiple continues compressing on distrust (as the July 2026 downgrade wave suggests is still a live risk), earnings growth alone will not be enough. Wide dispersion of outcomes — this is a genuine "close call," not false conviction in either direction.

Phase 3 — Debate round: is the Homes.com spend discretionary or structural?

Fundamentals + Sentiment argue discretionary (a real switch, now proven): Management has already demonstrated it can throttle this spend on command — the 21% Homes.com salesforce cut delivered a sequential swing to the segment's first-ever profitable quarter within two quarters of the commitment; the buyback reversed the share count within two quarters too; the Foundry Park capex program is separately, mechanically ending in H1 2026 regardless of Homes.com's fate; and the balance sheet (net cash, 11.2% Debt/Assets) means none of this is forced by financial distress. This is optionality CoStar is choosing to fund from a real cash machine, not a company burning down its balance sheet to survive.

Moat + the sell-side downgrade wave argue structural risk remains live: The spend rate is discretionary, but the strategic outcome is not something spending discipline alone can guarantee. Zillow's 7x traffic advantage and its own network effects are a competitive-position problem, not a cost-control problem — cutting the Homes.com salesforce 21% may simply be an admission that the $1B marketing bet already failed to move the traffic needle enough to justify continued spend at that level, which is a different and more troubling story than "we found efficiencies." D.E. Shaw's explicit call to divest/shut down — a position not shown to have been withdrawn — is a standing bet that the honest resolution of "discretionary vs. structural" is that the money is mostly already spent and mostly already lost, and the 2026-2030 plan is optimized to make that loss shrink gracefully rather than to produce a business worth what was invested.

Synthesis: Both are true at once, at different levels. The rate of spend is discretionary and management has just proven it can flip that switch. Whether the resulting, smaller Homes.com is ever worth the ~$3B+ cumulative investment is structural and will not be known until the 2029-2030 checkpoints the company itself set. Calling this cleanly "discretionary" would be false comfort; calling it cleanly "structural" ignores the real, already-delivered proof of spending discipline and the core business's unrelated strength. The correct framing for sizing: treat the core Commercial business as the reason to own this name, and Residential as a multi-year call option with a real chance of expiring worthless — which is exactly the shape a Re-Rating sleeve position should take.

Phase 4 — Synthesis & Explicit Answers

1. §1 health bar, given reported FCF is depressed by choice. Health is better than the reported trailing figures suggest, and already improving on a clean read. Reconciled TTM FCF is $227.0M (6.4% margin), not the vendor-snapshot $412.7M nor the misleadingly-low FY2025 full-year $41.0M — both wrong in different directions if quoted uncritically. Core segment economics (79% gross margin, 36% Commercial EBITDA margin, 61 consecutive quarters of double-digit revenue growth) are intact and not part of the depression. Balance sheet is low-risk (11.2% Debt/Assets, 2.2x current ratio) though the net-cash cushion has thinned materially in the last 12 months from funding Homes.com + Matterport + buybacks simultaneously.

2. Buy zone or watch? Buy zone (accumulate), sized for the sleeve's risk level. Current price ($29.49) sits at/near the bear-case SOTP floor (~$27) and meaningfully below the base case (~$36-38) and analyst mean target ($37.30). Entry zone $27-33 (≈14-19x forward EPS).

3. Discretionary or structural, and the gap-closing metric/date? Both, at different levels — see Phase 3. The gap-closing checkpoints, in the company's own committed terms: FY2026 Homes.com net investment cut to below ~$550M (>$300M reduction from FY2025's $850M), tracked quarterly; Residential segment adj. EBITDA staying positive (first cleared in Q2 2026 at $12M — watch for backsliding in Q3/Q4); Homes.com specifically reaching revenue-in-excess-of-expenses exiting 2029 and positive adj. EBITDA in 2030 — the real structural test, still four years out.

4. Does it beat the weakest Re-Rating members, INGN (4.5) / LULU-DBX-FIS (5.0)? Yes, by roughly a point to a point and a half. The edge is twofold: (a) a quantitatively stronger, longer-tenured core moat (a 61-quarter growth streak and a monopoly-adjacent data business that has already outlasted a funded direct rival) than a typical Re-Rating nominee brings, and (b) a fixable problem that is unusually well-specified — dated, dollar-quantified, board-committed, activist-validated (Third Point's exit), and already showing a first data point of real progress (the Q2 2026 residential swing to profitability) rather than being purely a forward promise.

5. Proposed conviction: 6.0/10. Above the weakest Re-Rating cohort on moat quality and specificity of the fix; held below high conviction (7+) by the genuine, unresolved structural risk in Residential, the four-year runway to the real resolution date (2030), the sell-side's own skeptical reaction the day after the "inflection" quarter, and the thinner-than-before balance sheet cushion. "This is a close call because the core business's quality is not in question, but the residential bet's ultimate payoff genuinely is not knowable yet" — the framework's preferred honest framing over false conviction in either direction.

Key risks (named, not buried)

  • Zillow's traffic/network-effect moat may simply be too entrenched for Homes.com to ever clear breakeven, regardless of spending discipline — the 2030 target could slip or fail outright.
  • Revenue deceleration from the strategic pullback (salesforce cuts, Ten-X restructuring, Apartments.com price integrity) could persist and compress the multiple further even as margins improve — exactly what the July 2026 downgrade wave is pricing.
  • Thinner net-cash cushion than a year ago reduces the margin for a fourth simultaneous capital call (further M&A, a fresh residential re-investment, or a downturn in the core commercial data business tied to CRE transaction volumes).
  • Governance/activist overhang only partially resolved — Third Point exited satisfied, but D.E. Shaw's current stance is unconfirmed by this research.
  • Company-reported Homes.com KPIs (traffic) may not withstand third-party scrutiny — treat management's own residential progress claims with the same skepticism applied to the financials.

Sources