Floor & Decor Holdings, Inc. (NYSE: FND) — The Sales Multiple Looks Cheap; the Cash Math Still Demands a Recovery

Independent equity research — Floor & Decor Holdings, Inc. Report date: 2026-09-03 · Price (2026-09-02 close): ~$47.95 · 52-week range: $43.49–$90.42 · Shares outstanding (2026-06-25): ~106.9M · Market cap: ~$5.1B · Enterprise value excluding operating leases: ~$5.0B · FYE: Thursday on or before December 31 (FY2026 has 53 weeks)

⚡ Claude’s Take

This block is the author’s own subjective opinion and general information only — not investment advice. The analysis that follows takes no position and carries no price target; this section is the single exception.

Verdict: HOLD / WATCH at ~$48—better value, but still not a clean buy. Conviction: medium. I would permit only a small starter position below ~$45 and reserve serious accumulation for ~$36–$42, or pay more after two positive-comp quarters and rising lease-adjusted returns confirm the turn. The directional zone reflects roughly 0.8–0.95× lease-excluded EV/sales and a discount to a conservative partial-recovery value; it is not a point target. The body that follows is analytical and carries no recommendation or price target.

Floor & Decor remains the best specialist operator in US hard-surface flooring. Its warehouse assortment, same-day job-lot inventory, direct sourcing and Pro service create a real but narrow cost-and-convenience moat. The proof is not rhetoric: adjusted gross margin held at 43.7% in Q2 2026 even as comps fell 2.1%, and installation materials continued to outgrow the more commoditized laminate/vinyl category. The counterproof is equally important: conservative FY2025 lease-inclusive ROIC was only about 5.4%, below a reasonable cost of capital, because negative comps and immature stores overwhelmed the sourcing advantage. This is not a wide-moat compounder immune to cycles; it is a good retailer whose fixed-cost model magnifies them.

Since the July memo, the price has done more work than the thesis. Q2 sales rose 3.0% to $1.25B, adjusted EBITDA rose 1.2% to $152M, and monthly comps improved from −5.1% in April to −0.3% in June. Yet Q3-to-date comps were back to −2.2%, management said it was too early to call a bottom, July existing-home sales remained only 4.06M SAAR, and Harvard’s remodeling indicator points to growth slowing to 0.5% by Q2 2027. The raised FY2026 adjusted EPS guide of $1.88–$2.13 includes an extra week worth about $0.08. None of that meets the earlier bull test of two positive-comp quarters with rising ROIC.

What has changed is the apparent margin of safety. FND is near the lowest sales and book valuations in its public history; it trades around 1.06× lease-excluded EV/trailing sales and 9.4× lease-excluded EV/after-rent adjusted EBITDA, while management repurchased $65.7M at an average $49.36 in Q2. Those are the economically consistent multiples: the July memo incorrectly added operating leases to enterprise value while dividing by after-rent EBITDA. But EBITDA is FND’s most flattering denominator—D&A is 46% of TTM adjusted EBITDA, and the five-year cash surplus after capex and SBC was negative. Clean trailing P/E is roughly 26×, 52-week FY2026 guidance implies about 25×, and no-growth EPV/reproduction value cluster around only $2.5B–$2.8B versus $5.1B market value. The wager remains explicit: pay for a recovery before it is visible.

The tape argues for patience. At $47.95 FND sits below its 21-, 50- and 200-day exponential moving averages, has lost 39% over 12 months, and carries high market, housing and rate sensitivity. The July rebound failed; this is again a falling knife. The bull trigger remains two consecutive positive-comp quarters with lease-adjusted returns or at least adjusted operating margin rising; the bear trigger is negative comps after the flooring market turns positive. Catchy version: “cheap on the aisle, expensive at the cash register.”

📈 Stock Price Action — Five-Year Event Map

FND round-tripped a growth-stock mania into a housing-cycle bust. It rose from roughly $92 at the start of 2021 to $143.31 in November 2021, fell to a five-year low of $43.49 on May 15, 2026, rebounded above $60 after Q2, and then returned to $47.95 by September 2. The stock is roughly 66.5% below the five-year high, only 10.3% above the low, and below its 21-day ($54.73), 50-day ($55.22), and 200-day ($59.45) EMAs. Price data are from AZI’s adjusted daily series, through 2026-09-02. Moves are facts; driver attribution is interpretation.

Cycle narrative: The 2021 pandemic remodeling boom and growth multiple produced the peak. The 2022 rate shock broke both housing demand and valuation. A 2022–24 soft-landing rally briefly restored the multiple even while the earnings cycle weakened. The 2024–25 turnover freeze then made the operating problem visible: negative comps, fixed-cost deleverage, and falling ROIC. The early-2026 tariff shock and Q1 miss created the May low. Q2’s better monthly cadence and actual buyback execution powered another rebound, but the subsequent reversal—without a corresponding earnings warning—shows how decisively the stock remains a housing/rate instrument. The event map says the market has repeatedly anticipated a turn that the comp line has not yet delivered.

1. Executive Summary

Floor & Decor is a category-killer specialty retailer of hard-surface flooring—tile, laminate/vinyl, wood, natural stone, and the installation materials and decorative accessories needed to complete a job. At Q2 2026 it operated 281 warehouse stores, up from 270 at FY2025 year-end, against a stated long-term target of at least 500. The legacy box averages roughly 76,000 square feet and carries about 4,200 SKUs; the 2026 class is nearer 55,000 square feet. Direct procurement from more than 240 suppliers, five regional distribution centers, job-lot inventory and everyday-low pricing give FND a genuine category-scale advantage over independents. They do not create a defensible cost advantage over Home Depot or Lowe’s.

The central tension remains cycle versus structure. Revenue depends on project-based remodeling and is especially sensitive to home turnover; July existing-home sales of 4.06M SAAR were virtually unchanged from the depressed 2024–25 annual level, while the latest 30-year mortgage rate was 6.66%. Comps were −7.1%, −7.1%, and −1.8% in FY2023–25, then −3.7% in Q1 and −2.1% in Q2 2026. Operating income fell from $396.8M in FY2022 to $270.1M in FY2025 and conservative lease-inclusive ROIC from roughly 10.2% to 5.4%. The core merchandising engine held up; fixed occupancy, payroll, pre-opening expense and immature-store economics did not. But LVP oversupply now shows that gross margin is not invulnerable either.

Capital allocation has changed from promise to action. Management cut openings from about 31 annually to 20, shifted toward lower-cost and smaller boxes, authorized its first $400M repurchase, and spent $65.7M at $49.36 in Q2. H1 free cash flow improved to $141.8M from negative $5.6M as working capital reversed and capex fell. The funded balance sheet remains conservative—$320.6M cash against $200M gross term debt, no revolver borrowings, and $621.8M net ABL availability after letters of credit—although $1.82B of operating-lease obligations are real fixed commitments. The unresolved issue is not liquidity; it is whether the next 219 stores earn above the cost of capital after occupancy and cannibalization.

Valuation is still a split decision, now at a lower absolute price. Own-history P/S and P/B sit around the first percentile, while normalized trailing P/E remains about 26× and 52-week FY2026 guidance implies about 25×. The category is consolidating, but the earlier memo overstated LL Flooring’s exit: 211 stores closed, while F9 retained 219 and the Lumber Liquidators brand remains active. Meanwhile excess LVP supply has lowered price architecture. The investment question is therefore sharper: is FND a temporarily under-earning consolidator whose store cohorts re-accelerate with housing, or a maturing concept whose nominal whitespace produces sub-WACC incremental returns? Q2 improved the evidence but did not answer it. No recommendation or price target follows in the body.

2. Business Overview

What it is. Founded in 2000 and public since 2017 (an Ares Management / Freeman Spogli LBO, since fully exited), Floor & Decor is a high-growth, multi-channel specialty retailer of hard-surface flooring and related accessories, plus a commercial-surfaces distributor (Spartan Surfaces). At fiscal year-end (12/25/25) it operated 270 warehouse-format stores and five small-format design studios across 39 states, served by five port-adjacent distribution centers, a transload facility, FloorandDecor.com, and the Spartan commercial subsidiary (FY2025 10-K, filed 2026-02-19). [FACT]

The box is the strategy. Stores average roughly 76,000 square feet, carry about 4,200 SKUs, and hold roughly 1.0 million square feet of flooring and $2.7M of inventory at cost each (FY2025 10-K). High ceilings, warehouse racking and separate Pro pickup entrances produce low occupancy cost per square foot and enough on-floor inventory to fill an entire job the same day. Management’s “broadest in-stock assortment” claim is not independently quantified against HD/LOW, but the dedicated category footprint and job-lot depth plausibly differentiate FND from a generalist flooring aisle. [FACT / INTERPRETATION]

How it makes money. Everyday-low-price retail on directly sourced product. FND procures the majority of merchandise directly from more than 240 manufacturers and quarries in roughly 20 countries, bypassing importers and distributors; no supplier exceeds 10% of sales (FY2025 10-K). The loop is direct sourcing → lower landed cost → EDLP → volume → scale benefits. Gross margin was 43.6% in FY2025, up from the 40.5% cyclical low in FY2022, even while demand weakened. FY2019 gross margin was 42.2%; the moat evidence is resilience through the recent downturn, not improvement from FY2019. [FACT / INTERPRETATION]

Product mix (FY2025 net sales by category; FY25 10-K):

The strategic tell in this table is installation materials & tools at 20% of sales — grout, mortar, backer board, trowels — the least-discretionary, most-recurring “supply-house” category, which grew in absolute dollars every year through the downturn. This is the Pro-anchored revenue that puts a partial floor under the box. [FACT]

Pro vs. DIY/BIY. FND defines Pros as professional installers and commercial businesses; DIY and “buy-it-yourself” are separate homeowner groups. Pro represented roughly 55% of Q2 sales and grew about 4%, continuing the FY2025 pattern in which Pro grew 9% while total sales grew 5.1%. Dedicated Pro desks, credit, free product storage, separate pickup access, delivery, loyalty and design support reduce search and downtime. That bundle creates repeat behavior but no contractual lock-in: Pros can multi-source. FND’s 2027 Pro app and continuing digital rebuild show that captivity is still being constructed. [FACT / INTERPRETATION]

Recurring vs. cyclical. This is not recurring revenue. It is big-ticket, project-based remodel/renovation demand — flooring is bought when a house changes hands or is renovated, with no subscription and no consumable reorder cycle. Management’s own stated demand drivers are “existing home sales, aging homes, rising home equity values, and the secular shift from carpet to hard surface flooring” (FY25 10-K). The Pro/installation-materials mix adds a semi-recurring floor, but the business fundamentally breathes with housing turnover — which is exactly why an excellent operator is currently earning trough returns. [FACT / INTERPRETATION]

Spartan (commercial). Acquired in FY2021, Spartan sells commercial surfaces (to architects, designers, GCs) outside the warehouse base and is excluded from comps. It is not a separately reportable segment — immaterial to consolidated results — and should be treated as an optionality lever, not a needle-mover. [FACT]

Verdict: A well-run, differentiated, single-category big-box retailer whose revenue is cyclical remodel demand, roughly 55% Pro-weighted. Direct sourcing funds EDLP, volume and job-lot availability. There is still no recurring revenue, network effect or contractual retention; the model breathes with housing turnover, and Pro workflow convenience only moderates that exposure.

3. Industry Dynamics

Market structure. US hard-surface flooring is split among Home Depot and Lowe’s, a handful of national/regional specialists, manufacturers and distributors, and a long tail of independent flooring and tile shops. Fragmentation gives a scaled specialist room to consolidate share, but low formal entry barriers and large generalists limit pricing power. The current market denominator and channel caveats are quantified below.

The secular tailwind (real, but slow). The carpet → hard-surface shift is a genuine multi-decade structural driver — hard surface rose from ~39% of the flooring mix (2002) to ~57% (2024) — and within hard surface the migration to waterproof rigid-core LVP favors precisely the categories where FND has the deepest in-stock assortment. This durable mix-shift expands FND’s addressable market irrespective of the cycle. But it is a low-single-digit annual grind, not an accelerant, and residential flooring volumes actually declined in 2023 and 2024 — even the secular winner bends to the cycle. [FACT / INTERPRETATION]

The cyclical driver. Hard-surface demand is levered to existing-home sales and repair-and-remodel activity. Existing-home sales ran 4.09M in 2023 and 4.06M in both 2024 and 2025, the lowest annual pace since 1995. July 2026 was still 4.06M SAAR—down 1.7% sequentially, up only 0.7% year over year—and July pending sales declined 2.3% sequentially and 2.2% year over year (NAR, August 11 and 18, 2026). The latest 30-year mortgage rate was 6.66% (Freddie Mac PMMS, August 27, 2026). The old housing stock and “improve, don’t move” behavior create a spending floor, but Harvard’s July LIRA expects improvement-and-repair spending growth to fade to just 0.5% by Q2 2027 (JCHS, July 23, 2026). Stabilization is not recovery.

Market size, with denominator discipline. A defensible 2025 broad flooring market is about $23.5B at first point of distribution, derived from resilient flooring’s $8.59B and 36.5% dollar share. Hard surface excluding rubber was approximately $14.4B, the closest public product-market serviceable market (Floor Covering News, July 1, 2026). FND’s roughly $2.75B of FY2025 surface-product retail revenue corresponds to a deliberately broad 11%–19% directional share bracket: 19% compares retail revenue directly; about 11% converts the numerator using consolidated gross margin. Neither is a precise share estimate because the industry denominator is at distribution, category margins differ, and FND also sells tools, services and accessories. The useful conclusion is that FND is meaningful but nowhere near saturated nationally; the exact share trend remains undisclosed.

Competitive set.

• Home Depot / Lowe’s — the real threat. They carry flooring, can match FND on price where they choose (flooring is a traffic category for them), and own the Pro relationship in adjacent trades. What they lack is FND’s breadth and same-day job-lot depth in hard surface specifically — flooring is one aisle for them, the entire store for FND.
• Independents & distributors — the fragmented tail FND is built to kill. They cannot match direct-sourcing scale, assortment, or price.
• Lumber Liquidators—the capacity event, not a full exit. LL Flooring filed Chapter 11 in August 2024. F9 retained 219 stores while 211 were slated to close, and Lumber Liquidators continues to advertise more than 200 locations. Bed Bath & Beyond has since agreed to acquire F9 brand assets. The event removed material capacity, but the surviving chain remains a competitive asset and may gain procurement and digital support. The July memo’s “dead rival” framing was too strong.

Marathon capital-cycle lens—mixed, not simply constructive. Capacity is rationalizing at the specialist-retailer layer: 211 LL doors closed, independents remain pressured, and Tile Shop’s FY2025 operating loss illustrates weak tail economics. But FND is itself adding about 20 stores annually, HD and LOW continue to invest in Pro and home services, and a major product layer is oversupplied. Management describes laminate/vinyl as “devalued”: excess supply lets higher-quality LVP sell at opening price points and may persist through H1 2027. Industry consolidation can shift share toward FND without restoring price or return on capital. This is late-bust/early-recovery in retail capacity and still excess capacity in product manufacturing. [FACT / INTERPRETATION]

Current peer signal is inconclusive. FND’s Q2 comp of −2.1% lagged Home Depot’s +1.7% global/+1.3% US comp and Lowe’s +0.2% comp. Yet HD and LOW have broader small-project, appliance, services and trade exposure, so total-company comparisons do not establish flooring share. More useful is the common pattern: Pros held up better than DIY, smaller projects outperformed big remodels, and online/service growth offset softer discretionary demand. A 2025 laminate-channel proxy is favorable—specialty-retail share rose while HD/LOW channel shares fell—but it does not isolate FND. The honest answer is that management may be gaining category share, but no consistent public five-to-eight-year retail sell-through series allows Greenwald’s share-stability test to be scored.

Tariffs. Broad 2025 tariffs hit many sourcing countries; ceramic tile from India separately drew final countervailing-duty rates around 3.06%–3.45% in April 2025. FND negotiates with vendors, re-sources and passes residual cost through price, using the same global procurement network that supports its moat. The 2026 IEEPA refund reverses much of the immediate headline exposure and temporarily lowers inventory cost, but it does not eliminate product-specific duties or future policy risk. Whether smaller rivals are hurt more is plausible, not established. [FACT / INTERPRETATION]

Verdict: A fragmented industry in a mixed capital cycle. The secular shift to hard surface and specialist-retailer contraction favor a scaled survivor, but LVP manufacturing capacity remains excessive and HD/LOW keep investing around the Pro. Existing-home turnover is stable at a historically depressed level and remodeling growth is decelerating. FND can gain share without gaining pricing power or industry profit. Good category structure, weak demand, mixed supply discipline.

4. Competitive Position / Moat

Moat type (Greenwald taxonomy): a narrow cost/scale advantage — dominant local/category economies of scale in a fragmented niche — with an emerging, not-yet-proven customer-captivity layer. Not a wide moat.

The cost advantage is real, and it shows up in the financials. FND’s differentiation is direct global sourcing at scale plus the low-occupancy warehouse format. Gross margin rose from roughly 40.5% in FY2022 to 43.6% in FY2025 while cumulative comps declined sharply. That validates a durable sourcing edge over fragmented independents, which cannot readily match 240 supplier relationships, five regional distribution centers and national purchasing volume. Q2 adjusted gross margin of 43.7%, down 20bp, also sets the boundary: FND has procurement skill, not unrestricted pricing power. Imports create tariff, customs, forced-labor, freight and currency exposure; the IEEPA refund does not erase product-specific duties such as ceramic-tile trade remedies.

But pressure-test it against the giants. FND’s cost advantage is clear versus independents and smaller specialists, but not clearly versus Home Depot and Lowe’s, which have far more aggregate purchasing scale and can use flooring as a traffic category. FND’s edge over the majors is assortment depth and same-day job-lot availability in hard surface specifically, not absolute landed cost. In Greenwald’s terms this is a local/category economies-of-scale advantage, not a fortress. [INTERPRETATION]

Switching costs—aspirational, under construction, not yet proven. Credit, free inventory storage, Pro desks, delivery and loyalty create workflow friction, but management describes switching costs as expected to increase and the new Pro app is not due until 2027. Pros are structurally multi-source and price-sensitive; a rewards program is table stakes, not lock-in. FY2025 Pro growth of 9% is consistent with share gains, but it is not proof of captivity and may also reflect customer/category mix. [FACT / INTERPRETATION]

The financial-outcome test. A moat must tie to a financial outcome that deteriorates without it. Gross margin says the sourcing advantage is real. Return on capital says the overall moat is not currently earning its keep: conservative lease-inclusive ROIC fell from roughly 12.7% in FY2021 to 5.4% in FY2025, below a reasonable WACC. The honest reading is a narrow product-cost advantage overwhelmed by SG&A deleverage, negative comps and new-store drag. That does not prove the moat is fake, but it proves it is not wide enough to defend excess returns through a downturn. [INTERPRETATION]

Is it a moat, or just a well-run retailer? Mostly a very well-run category retailer with a genuine but penetrable cost edge. The sourcing system and box are replicable in principle but costly to reproduce at scale; FND’s 25-year head start, vendor base and distribution network matter. There is no network effect, protected intellectual property or regulatory barrier. Customer captivity varies: low for infrequent DIY, moderate for design-assisted BIY, and emerging for recurring Pros. Management’s own “street fight” description fits.

Verdict: A real but narrow moat—category/local scale, sourcing efficiency, assortment and job-lot convenience, with a modest emerging Pro-captivity layer. It is durable against small independents, but not a fortress against Home Depot, Lowe’s or a recapitalized specialist. Lease-adjusted ROIC around 5%–6% fails the strongest economic-outcome test today. Call it a well-run retailer with a defensible edge, not a wide-moat compounder.

5. Growth History and Forward Opportunities

Two engines, one now stalled. FND’s growth algorithm is new-unit growth × comparable-store sales. Revenue compounded from $2.05B (FY19) to $4.68B (FY25) — a ~2.3× run. Historically both engines fired; today only units do, and units are decelerating.

Comp trajectory — three years of pain, tentatively bottoming (FY25 10-Ks; transcripts):

FY2025’s −1.8% comp—with average ticket +1.8%—was a genuine sequential improvement, and Q2 2026 improved from Q1. The quality of H1 growth remained narrow: total sales rose 1.2% because non-comparable-store sales added $92.8M against a $65.1M comp-sales decline. By product, tile grew 5.5%, wood 4.7%, installation materials 8.1%, and decorative accessories 2.2%, while laminate/vinyl fell 7.4%. H1 ticket rose 1.4% and transactions fell 4.2%; Q2 ticket rose only 0.8% as transactions declined 2.9%. This is unit, Pro and attachment growth—not broad organic demand recovery. Q3-to-date comps of −2.2% and the unchanged FY2026 −4%-to-flat guide preserve that conclusion. [FACT]

New-unit economics and deliberate deceleration. New warehouses follow a four-to-five-year maturation curve, initially diluting sales productivity, margin and consolidated returns. Management cut openings from about 31 annually to 20 in FY2025 and 2026, uses more second-generation real estate, and reduced the 2026 average format to about 55,000 square feet from the 76,000-square-foot legacy average. Half the FY2025 base was less than five years old, and the 10-K says post-2022 cohorts produced lower first-year sales and initial returns than earlier classes amid higher construction and occupancy cost. Management says the smaller format preserves productivity, but the evidence disclosed is insufficient to prove it: 24 stores opened since the prior-year quarter contributed $60.5M of Q2 non-comp sales while non-comp SG&A rose $26.7M, with no cohort gross profit, rent or invested-capital bridge.

Cannibalization makes the same point. Management said eight of 16 districts and both East and West regions were positive excluding cannibalization, while reported Q2 comp was −2.1%. That demonstrates local demand health in parts of the fleet but does not establish system economics, because the cannibalization drag was not quantified. If a new box transfers sales from a mature nearby store, consolidated revenue can grow while mature-store productivity and incremental ROIC fall. The required disclosure is sales, four-wall EBITDA, invested capital, lease-adjusted ROIC, cannibalization and payback by the 2022–26 cohorts and by 50k/55k/75k-square-foot format.

The annual rows show gross openings; year-end counts are net of closures. Q1 and Q2 are sequential additions within FY2026.

Forward runway. The white-space story is intact on paper: 281 stores at Q2 versus at least 500, leaving 219 locations or roughly 78% unit upside. Layered on:

• Pro penetration—roughly 55% of sales, the most resilient customer, with installation materials, loyalty, a future app, credit and fulfillment as levers. This is the highest-quality growth vector.
• Commercial (Spartan) — small; optionality only.
• Connected-customer / design services—online penetration reached 20.3%, up 170bp year over year, during an 18–24 month digital transformation. These are conversion and retention tools, not standalone growth engines.
• Commercial field coverage—the regional-account-manager team has reached about 80 people, while Spartan Q2 sales grew only about 2%. Commercial is credible optionality but not separately disclosed at a scale that warrants a distinct valuation.

The decisive question: is unit growth into a weak market creating or destroying value? Right now it is diluting measured returns. The roughly 12.7% to 5.4% lease-inclusive ROIC decline partly reflects adding immature, sub-fleet-average boxes onto a shrinking comp base while carrying full SG&A and D&A. The nominal whitespace is real, but the claim that every trough-era store is ultimately value-accretive is not yet evidenced. Management’s slower cadence, cheaper boxes and first buyback acknowledge the tradeoff. Value creation is deferred and conditional on positive comps, maturation and controlled cannibalization. [INTERPRETATION]

Verdict: Low-quality growth today, with unproved recovery optionality. Comps have been negative for three years, unit growth has outrun store productivity, and returns are sub-WACC. The 500-store target is genuine physical whitespace but not yet demonstrated economic whitespace. Pro, smaller boxes and digital tools can improve the outcome; housing normalization and controlled cannibalization are still required. Growth quality is a leveraged bet on both demand and cohort execution, not a self-help result already in hand.

6. Financial Quality

Revenue growth & composition. Revenue grew every year 2019–2025 ($2.05B → $4.68B), but the quality of that growth degraded sharply: from +24.2% (FY22) to +0.9% (FY24) to +5.1% (FY25), and the composition flipped from comp-led to entirely unit-led (three years of negative comps). This is the single most important fact in the financials — the top line kept rising only because the store count kept rising. [FACT]

Margins — the whole story is SG&A, not gross margin. This bears repeating because it is counter-intuitive and it governs the thesis:

Gross margin rose roughly 310bp from FY2022 to FY2025 while operating margin fell roughly 350bp. The compression is SG&A deleverage—occupancy and payroll on a store base that grew 41%, plus pre-opening and immature-store cost. Q2/H1 2026 SG&A reached 38.3%/38.9% of sales versus 37.1%/37.6% a year earlier. The FY2025 10-K says newer stores have higher construction, occupancy and operating cost and may have lower profitability. The pattern is consistent with a cyclical fixed-cost squeeze, but calling it entirely cyclical would assume what needs to be proved: declining average-unit-volume proxies and persistent cannibalization may be partly structural. Lease-inclusive ROIC uses after-tax EBIT divided by average debt plus equity plus operating-lease liabilities less cash; it does not add imputed lease interest to NOPAT, so it is deliberately conservative and definition-sensitive. [FACT / INTERPRETATION]

Store productivity is the missing bridge. A simple disclosed-segment AUV proxy—Retail revenue divided by average beginning/end warehouse count—fell from approximately $20.5M in FY2023 to $18.0M in FY2024 and $17.1M in FY2025. The comparable H1 annualized proxy fell about 6.5% year over year to $16.6M as Retail revenue rose 1.4% while end-store count rose 9.3%. Staggered openings and seasonality make this an analytical proxy, not a company KPI. Still, it demonstrates why positive total sales cannot substitute for cohort economics: unit growth is currently outrunning productivity.

FCF — capex-driven volatility, recovering off a low base. Operating cash flow is healthy, but growth capex has swung FCF wildly:

FCF is depressed by choice, but that choice is economically real. Capex peaked at $548M in FY2023 and fell to $318M in FY2025. Across FY2021–25, cumulative OCF of $2.202B barely exceeded $2.177B of capex, leaving only $25.9M; after charging $133.2M of SBC, cumulative economic surplus was negative $107.3M. Working capital made individual years volatile—FY2022 built inventory and FY2023 reversed it—but the five-year total avoids cherry-picking. The capital-cycle inflection is now constructive: capex/D&A fell from 3.45× in FY2021 to 1.32× in FY2025, and FY2026 guidance of $240M–$275M against roughly $250M of D&A is 0.96×–1.10×. FCF can rise even without an earnings recovery. The risk is that lower spending merely harvests the estate while newer cohorts still earn below hurdle rates.

Quality-of-earnings checks. The underlying accounting is generally clean: SBC remains low, the share count has grown modestly over six years, inventory discipline improved through FY2025, and cash generation reconciles sensibly across the cycle. Q2 2026 is the important exception. An IEEPA tariff refund produced a $56.2M cost-of-sales benefit; the non-GAAP reconciliation records a $45.2M pretax tariff adjustment, a $1.3M debt-extinguishment add-back, and a combined $11.0M tax line, yielding a $32.9M combined after-tax difference between GAAP and adjusted net income. A further $28M reduced inventory cost, of which $6M sold through Q2 and most of the remaining roughly $22M should support H2 margin. Reported Q2 gross margin of 48.2%, operating-income growth of 51.4%, and EPS of $0.89 are therefore not run rate. Adjusted gross margin was 43.7% and adjusted EPS $0.58. Any valuation that capitalizes the refund as recurring profit overstates earnings power. [FACT / INTERPRETATION]

H1 cash flow improved, but normalize working capital. H1 operating cash flow was $278.4M versus $155.3M a year earlier; capex was $136.7M versus $160.8M, producing $141.8M of free cash flow versus negative $5.6M. The improvement is directionally real because new-store capex is easing, but cash also benefited from working-capital timing. The tariff receivable was roughly $80M at quarter-end and collected afterward, so it was not yet in the quarter-end cash balance. FY2026 capex guidance of $240M–$275M comprises roughly $140M–$165M for new/relocated/future stores, $60M–$65M for existing stores and distribution centers, and $40M–$45M for IT/e-commerce. Maintenance and growth are intertwined; treating all capex as maintenance understates owner earnings, while ignoring ongoing remodel/DC/technology needs overstates it.

Balance sheet—conservative funded debt, meaningful fixed leases. Q2 cash was $320.6M against $200M gross term debt, with the term loan refinanced to 2033 at 5.64%. The $800M revolver maturing 2031 was undrawn. Operating-lease liabilities were $163.2M current plus $1.654B long term, or $1.817B. The lease term averages roughly 12 years and the weighted discount rate is 6.1%. Excluding leases, FND is net cash and has ample liquidity. Including them, it has a large fixed occupancy commitment that matters if stores underperform. Both statements are true; solvency is not the thesis risk, but lease-adjusted returns are.

Verdict: Solvent, gross-margin-resilient, but medium-low earnings quality today. The merchandising engine, funded balance sheet and capex moderation are positives. Negative traffic, falling store-productivity proxies, SG&A deleverage, a large one-time Q2 benefit, working-capital-aided H1 FCF and roughly 5%–6% lease-inclusive ROIC are the counterweight. A meaningful portion is cyclical; the undisclosed cohort economics prevent concluding that all of it is. Return recovery, not reported sales growth, is the proof point.

7. Capital Allocation

The core story: FND is a reinvestment engine that grew its asset base into a falling-return trough — and management has, for the first time, blinked.

Growth capex is the whole game. Capex ran $548M in FY2023 and $318M in FY2025 as openings fell toward 20 annually. GAAP EBITDA less capex was negative roughly $24M in FY2023 and only about $42M in FY2024; as openings slowed it recovered to roughly $193M in FY2025. More rigorously, cumulative five-year OCF barely covered capex and did not cover capex plus SBC. Meanwhile assets rose roughly 90% from FY2020 to FY2025 and conservative lease-inclusive ROIC fell from about 12.7% to 5.4%. That is a Marathon asset-growth warning: the firm expanded the capital base faster than economic profit. [FACT / INTERPRETATION]

Marathon capital-returns lens—the asset-growth anomaly. Heavy capital deployment drove assets up roughly 90% from FY2020 to FY2025 and doubled store count while consolidated returns fell below the cost of capital. Expansion during compression destroys value unless returns are cyclically depressed and will mean-revert, or new-store four-wall economics remain above WACC despite an immature-estate drag. Management asserts its capital hurdle exceeds WACC, but does not disclose cohort returns. The slower cadence is a constructive response; it is not proof that the remaining openings create value. [INTERPRETATION]

M&A: modest checks, deteriorating outcome. Spartan Surfaces cost $77.7M in 2021, three small commercial sellers cost $4.6M in 2022, and Salesmaster cost $20.1M in 2023—about $102.4M disclosed consideration in total. This is not empire-building, but it is more than one acquisition. Commercial-segment revenue rose from $195.6M in FY2023 to $243.5M in FY2025 while operating margin fell from 9.8% to 5.1%; H1 2026 margin was only 1.0%. The segment’s growth has not yet demonstrated synergies or moat. Underwrite no acquisition value until margins stabilize. [FACT / INTERPRETATION]

Buybacks: a philosophical shift, now executed. The April 2026 $400M open-ended authorization was FND’s first. In Q2 it repurchased 1,330,975 shares for $65.7M at an average $49.36, leaving $334.3M. The current $47.95 price is below management’s demonstrated execution level. This is stronger information than an unused authorization, but not proof of intrinsic value: the full authorization is large relative to trough FCF and continued deployment would consume cash or incremental borrowing. A disciplined policy should compare each repurchased share with the risk-adjusted return on a new store, not simply defend the quote. At today’s valuation, the first tranche looks economically sensible; the board should preserve liquidity for an elongated housing trough. [FACT / INTERPRETATION]

Insider activity since Q2. Executive Chair Thomas Taylor exercised 218,189 options at $21 and sold all resulting shares at a $62.30 weighted average on August 7, retaining 229,820 directly owned shares. The options were set to expire on April 27, 2027, and the filing affirmatively marks the trade as not under a Rule 10b5-1 plan (SEC Form 4, August 10, 2026). Expiration within 8.5 months supplies a liquidity motive, but the full, unplanned exercise-and-sale is a moderate negative and partially offsets the CEO/CFO’s May open-market purchases. The other post-quarter filing was routine tax withholding by the CIO.

Incentives—less protective than the prior memo stated. FY2025 annual bonus was weighted 20% to sales and 80% to EBIT. Annual LTI was split equally between RSUs and PSUs; the PSU half was 80% adjusted EBIT and 20% three-year average ROIC, with the metrics vesting independently. Thus ROIC determines only 10% of total annual LTI value. Earlier PSU designs included a minimum ROIC gate that zeroed FY2022/FY2023 awards, but FY2024/FY2025 grants removed that gate: missing ROIC no longer blocks the EBIT portion. That is a material governance downgrade for a retailer whose central question is incremental capital efficiency. SBC is modest, the independent-board majority is positive, and there is no dividend. [FACT / INTERPRETATION]

Ownership. The PE sponsors have fully exited; the 5% register comprises non-controlling institutional holders—BlackRock 9.2%, Vanguard 8.9%, FMR 7.4%, Capital World 6.4%, and Principal 5.7%. Some are active managers, so “entirely passive” would be inaccurate. No holder controls the company and no dual-class overhang remains. [FACT]

Verdict: Mixed and on watch. The positives are a net-cash funded balance sheet, extended maturities, lower capex, modest SBC and an initial buyback below $50. The negatives are essentially nil five-year cash surplus after capex/SBC, weakening store productivity, deterioration in the acquired commercial segment, removal of the ROIC gate, and large unused debt capacity. Solvency discipline is proven; incremental-return discipline is not.

8. Changes and Headwinds — Last Two Years

CEO transition (orderly, external, complete). Thomas V. Taylor, CEO since 2012 and the architect of FND’s growth, moved to Executive Chair effective Dec 26, 2025. Bradley S. Paulsen — ex-CEO North America at Rentokil Initial, prior CEO of Rexel USA, COO at HD Supply, nine years at Home Depot — joined as President (April 28, 2025) and became CEO (Dec 26, 2025). President Trevor Lang retired March 1, 2025 (forfeiting unvested equity, no severance); CAO Luke Olson resigned Dec 2024. A distribution-heavy operator now runs a distribution-heavy retailer — sensible fit — but Paulsen is unproven at FND and inherits a demand trough. [FACT]

Tariffs—the headline reversed, the exposure did not. The Q2 IEEPA refund created a $45.2M net pretax benefit and roughly $80M receivable, reversing the immediate accounting hit that dominated the Q1 narrative. But the refund is neither recurring earnings nor a blanket removal of import risk. FND still sources a majority of products internationally and faces product-specific duties, freight, customs and policy uncertainty. About $22M of reduced inventory cost should flow through H2, temporarily helping adjusted gross margin before the benefit expires. The analytical change is from “tariff cost shock” to “refund-supported 2026 margin with persistent structural import exposure.” [FACT / INTERPRETATION]

Demand and guidance—better execution, unchanged demand range. Q2 sales rose 3.0% to $1.250B, comp improved to −2.1%, adjusted EBITDA rose 1.2% to $152.0M, and adjusted EPS held at $0.58. Management raised FY2026 adjusted EBITDA to $550M–$585M and adjusted EPS to $1.88–$2.13, but left sales at $4.77B–$4.99B and comps at −4% to flat. A 53rd week contributes about $11M EBITDA and $0.08 EPS. The profit raise reflects execution, buyback and refund-related dynamics more than a raised demand view. [FACT]

Competitive capacity: contraction, not elimination. The LL bankruptcy removed 211 stores but F9 retained 219 under the Lumber Liquidators brand. This is still a durable reduction in specialty capacity, though not the complete competitor exit described previously. [FACT]

Product deflation emerged as a clearer headwind. Laminate/vinyl—the second-largest category—was the only major surface area under pronounced downward pressure in Q2. Excess industry supply has “devalued” the category, allowing better-quality products to move into opening price points. Management is responding with opportunity buys, sharper prices and new SKUs and expects pressure potentially through H1 2027. This can increase units and customer value while reducing ticket and obscuring share gains in reported revenue. [FACT / INTERPRETATION]

The smaller box and digital transformation are new execution tests. The 2026 store class averages about 55,000 square feet, intended to preserve assortment productivity with lower build and occupancy capital. At the same time, FND is spending through an 18–24 month digital transformation; online penetration reached 20.3%, and a Pro app is planned for 2027. Both can improve incremental returns, but neither has enough cohort disclosure to be underwritten as proven self-help.

Litigation and accounting. The 2020 derivative case settled in FY2024 and produced a $6.8M net recovery. A wrongful-death case settled in Q4 FY2025 entirely within insurance coverage; Q2 2026 disclosed only ordinary-course matters not expected to be material. The five-year filing review found no restatement or material accounting-policy change. [FACT]

Verdict: Incrementally better execution, not a fundamental turn. The refund reduced 2026 tariff pressure, Q2 comps improved sequentially, Pro mix rose and the buyback was used. Offsetting those gains, Q3-to-date comps stayed negative, LVP oversupply emerged, the housing/remodel outlook remains soft, and the smaller-store economics are still undisclosed. The thesis still hinges on whether current comps and returns are the bottom of a housing cycle or the new normal of a maturing concept.

9. Risk Analysis

Overall: The dominant risks are cyclical and correlated — all roads lead back to housing turnover and rates. The tail risks (solvency, fraud, obsolescence) are low: FND is profitable through the trough, net-cash ex-leases, with clean accounting. The genuinely thesis-relevant risk is the “cyclical vs. structural” ambiguity on comps — if FND is losing share as the market recovers, the whole normalization case fails.

10. Valuation Discussion — Embedded Expectations

No price target; no recommendation. This section frames what the current price implies.

Start with internally consistent enterprise value. At $47.95 and 106.886M quarter-end shares, equity value is approximately $5.13B. Add $200.0M gross funded debt and subtract $320.6M cash to obtain about $5.00B of enterprise value excluding operating leases. TTM sales are approximately $4.712B, and TTM adjusted EBITDA is approximately $531.7M, producing 1.06× EV/sales and 9.4× EV/adjusted EBITDA. Both denominators are after rent, so excluding the $1.817B operating-lease liability is the consistent primary convention. Adding lease liabilities without adding rent back to EBITDA mixes accounting bases and overstates the multiple. Lease-inclusive enterprise value is useful only with a matched EBITDAR denominator; the lease liability instead appears in this memo’s fixed-charge and lease-adjusted ROIC analysis.

That corrects a material error in the July memo, which showed roughly $8.3B of EV at $59 and divided it by after-rent sales and EBITDA. The correction does not make leases disappear. It separates two questions: What is the market paying for after-rent operating profit? And can that operating profit service long-dated store commitments? The first is answered by lease-excluded EV multiples. For the second, $1.714B of net funded debt plus recognized lease liabilities equals 3.22× after-rent TTM adjusted EBITDA—a deliberately conservative fixed-commitment diagnostic, not a matched lease-adjusted valuation multiple—alongside roughly 5.4% lease-inclusive FY2025 ROIC.

Earnings require a second normalization. Reported trailing EPS of about $2.15 includes a $0.31 GAAP-to-adjusted Q2 difference comprising the $45.2M pretax tariff adjustment, a $1.3M debt-extinguishment add-back and the combined tax line. A simple clean bridge—FY2025 EPS of $1.92 plus H1 2026 adjusted EPS of $0.95 less H1 2025 EPS of $1.03—gives approximately $1.84 of clean TTM earnings, or 26.1× at the current price. The FY2026 adjusted guide midpoint is $2.005, or 23.9×, but includes a 53rd week worth about $0.08. The 52-week midpoint is $1.925 and the multiple approximately 24.9×. FND is cheap on sales and book, not yet conventionally cheap on trough earnings.

Own-history context. On September 2 data, P/S ranked at approximately the 0.55th percentile, P/B at the 0.33rd percentile, P/E at the 4.87th percentile, and a composite measure at the 1.92nd percentile of FND’s public history (AZI fundamentals). The headline P/E percentile is flattered by the refund, but the sales/book message survives. FND once traded as a high-teens unit-growth concept with double-digit comps; today investors pay roughly one turn of sales for negative comps, slowing unit growth and sub-WACC returns. The multiple has compressed for rational reasons, yet it also embeds unusually little value for the nominal 219-store runway.

Scenario framework—what today’s price asks operations to deliver. These are operating cases, not price targets.

The bear case says today’s quote is expensive for a no-growth, sub-WACC retailer: the P/E stays in the mid-20s and the low sales multiple never converts to owner earnings. The base operating case makes current P/E look more reasonable, but a cash-flow model still has to fund the store estate and terminal reinvestment. The bull case requires the store pipeline to mature and housing turnover to normalize. The equity quote has fallen, yet it still capitalizes a meaningful recovery because current earnings and assets alone do not cover market value.

Greenwald EPV cross-check. Earnings power value asks what the existing franchise is worth without credit for growth. Start with the 52-week FY2026 adjusted EBITDA midpoint of about $556.5M, subtract roughly $245M of D&A and $0M–$30M of economic SBC, tax the result at 23%, capitalize it at 9%, and add about $125M net cash. The result is approximately $2.5B–$2.8B of equity EPV. This method treats depreciation as the maintenance-capital proxy and current adjusted margin as sustainable; it neither credits a cyclical margin recovery nor charges incremental growth capex. Market capitalization of $5.13B is about 1.8×–2.0× that value. A material franchise recovery and/or profitable growth is therefore embedded despite trough conditions.

Asset/reproduction-value cross-check. Book equity is $2.496B, or about $23.35 per share. A conservative historical-cost reproduction proxy subtracts $257.9M goodwill and $144.7M other intangibles, then adds $680.0M accumulated fixed-asset depreciation, leaving approximately $2.77B. Operating right-of-use assets and lease liabilities remain matched inside book equity. This is not an appraisal: it omits construction inflation, vendor relationships, brand, training and pre-opening know-how, while adding back some potentially obsolete depreciated assets. Its purpose is diagnostic. EPV and reproduction value both cluster near $2.5B–$2.8B, while market capitalization is roughly 1.85× reproduction. About half the quote is payment for future improvement, not current earning power or replicable assets. That is consistent with a narrow moat whose excess returns are presently unproved.

FCF and capital-cycle cross-check. Filing-rebuilt TTM OCF is about $505.0M and capex $293.6M, or $211.4M FCF and a 4.1% yield. Only about $7M of IEEPA refunds had been received by June 25; the remaining roughly $80M sat in receivables and was collected after quarter-end. Thus the reported TTM FCF is not inflated by the full refund, although working-capital timing still matters. At the other extreme, a pre-growth owner-FCF estimate uses roughly $206M of 52-week adjusted net income plus $245M D&A less $100M–$110M non-new-store/relocation capex, producing $341M–$351M or a 6.7%–6.9% yield. That cash capacity exists only if growth spending is curtailed. While FND opens 20 stores, it is not distributable owner cash.

Reverse DCF—the demanding part. Using $4.712B starting sales, a 5.5% normalized starting EBIT margin after SBC/tariff normalization, 23% tax, a five-year explicit period, 9% WACC, 2.5% terminal growth and no net dilution, current EV is hard to reproduce with a routine partial recovery. A bear case of 1% revenue CAGR, 5.0% terminal margin and 20% NOPAT reinvestment covers about 0.43× current EV. A base case of 5% CAGR, 7.5% margin and 25% reinvestment covers about 0.70×. A bull case of 8% CAGR, 9.5% margin and 25% reinvestment reaches roughly 0.99×. Terminal value contributes 72%–79% of modeled value, so these are diagnostics, not false precision.

The reverse sensitivity is equally stark. At 3%, 5%, 7% and 8% five-year revenue CAGR, the terminal EBIT margins required to reproduce current EV are roughly 12.1%, 11.0%, 10.1% and 9.6%, respectively, holding the other assumptions constant. The corrected filed FY2022 operating-margin peak was 9.3%, not the 10.2% previously reported in this memo series. Thus slower growth requires a terminal margin above any filed level in the five-year history. The price looks depressed relative to its own multiple history, but the cash-flow hurdle is closer to a bull case than the sales multiple suggests.

Relative-value cross-check. At September 2 prices and filing-rebuilt values, Home Depot traded around 2.15× EV/sales, 14.8× EV/EBITDA, 22.3× earnings and 21.1× FCF; Lowe’s around 1.62×, 11.5×, 16.9× and 16.0×, respectively. FND’s 1.06× sales and 9.4× EBITDA look discounted, but its 26.1× clean earnings and 24.2× reported FCF are premiums. The reversal occurs because D&A is 46% of FND adjusted EBITDA and continued expansion consumes cash. HD/LOW are imperfect—larger, diversified, mature and service-rich—but the cross-check warns against calling FND cheap on EBITDA alone. Unit-growth premium is deserved only after incremental stores clear WACC.

Verdict: Sales/book valuation is exceptionally depressed and conventional EV/EBITDA looks modest, but earnings, cash flow, EPV and reverse DCF remain demanding. Current EV embeds something close to high-single-digit growth plus near-peak margins under the stated assumptions. The market grants little value to a failed 500-store plan, but still pays materially for a successful one. Return-on-capital evidence—not the headline multiple—is the arbiter.

11. Variant Perception

Consensus view. FND is a high-quality category-killer temporarily depressed by a housing trough; own the compounder through the cycle and ride the 500-store runway. The de-rate, Pro growth and buyback identify the entry, while housing normalization restores margin.

Strongest bull case. Specialty capacity has contracted, FND’s gross margin and Pro growth demonstrate that its merchandising engine is intact, and the store estate contains a large immature cohort. A modest turnover recovery pushes comps positive; fixed-cost absorption then restores margin faster than sales grow, while smaller boxes lower incremental capital. The 219-location whitespace and digital/Pro tools extend the runway. The initial buyback below $50 retires shares before earnings recover. In this version, current lease-inclusive ROIC is a cyclical snapshot, not normal economics, and mid-cycle EPS power of roughly $3–$4 makes today’s trough sales valuation unusually attractive.

Strongest bear case. The de-rate is correct. Store count doubled from FY2020 to FY2025 while the five-year cash surplus after capex and SBC was negative; average-unit-volume proxies fell, adjusted H1 EBITDA declined, and lease-inclusive ROIC is roughly 5%–6%. The “500-store runway” describes physical openings, not profitable ones. LVP oversupply removes pricing, mature-market openings cannibalize the fleet, Home Depot and Lowe’s defend Pro relationships, and commercial-acquisition margins deteriorate. Even after the share-price fall, normalized earnings still cost roughly 25×. In this version the business is a maturing, capital-hungry retailer whose peak returns will not recur.

Variant conclusion. The market debate is usually framed as “housing turns” versus “housing stays frozen.” The more useful variant is incremental returns. FND can gain share and grow revenue while destroying value if the next store cohorts merely redistribute local sales and never cover occupancy and capital. Conversely, it does not need a return to pandemic demand if the 55,000-square-foot format, lower capex and Pro density restore payback. The differentiated work is therefore cohort-level, not macro forecasting. The absent data—format-level invested capital, four-wall EBITDA, cannibalization and payback—matter more than another mortgage-rate estimate.

The 3–5 assumptions that matter most:

1. Housing turnover normalizes (existing-home sales rise off ~4.06M SAAR) within a 2–3-year horizon. Falsified by: another leg down in existing-home sales / higher-for-longer rates.
2. The margin collapse is cyclical operating-leverage, not structural. Confirmed by: gross margin at a record high; falsified by: gross-margin erosion or SG&A that fails to re-leverage as comps recover.
3. New-store four-wall economics still clear WACC (consolidated ROIC is dragged by immature units, not broken unit economics). Falsified by: new-store productivity / mature-store four-wall returns disclosed below cost of capital.
4. FND is gaining, not losing, share. Falsified by: comps that stay negative while the broader flooring market recovers (management’s own “lagging the industry” admission is the yellow flag).
5. The ~500-store runway is real and value-accretive. Falsified by: accelerating cannibalization in developed markets / further cuts to the store target.

The factor-positioning read. FND is a high-beta, rate-sensitive housing cyclical. A September 1 base model shows Market +1.24, SmallSize +1.21, InterestRate −0.85, Momentum −0.58 and Quality +0.20, with 44.3% explanatory power; a July 31 broader model shows Home Construction +0.94, Housing Builders +0.62 and Retail +0.57 with 63.6% explanatory power (FactorsToday methodology). Raw price returns through September 2 were −3.2% over three months, −25.2% over six months, −39.3% over 12 months and −60.1% over five years. Annualized idiosyncratic volatility was 30.2%. The stock is below every tracked EMA and only 10.3% above its five-year low: a renewed falling knife, not the July “basing” setup. The tape reduces timing confidence; it does not decide intrinsic value.

12. Fact vs. Interpretation Table

13. Open Questions

1. Four-wall economics: What are sales, four-wall EBITDA, gross invested capital, lease-adjusted ROIC and payback for the 2022–26 cohorts versus mature stores? The filing acknowledges lower post-2022 first-year sales and returns but does not quantify the gap. A cohort table would distinguish temporary maturation from broken unit economics.
2. Cannibalization: What percentage-point drag separates reported comp from ex-cannibalization comp by district? Management disclosed eight of 16 districts and both regions positive excluding cannibalization, but not the system bridge. The answer determines whether 500 stores expand the profit pool or merely redistribute demand.
3. Share: Is FND gaining surface-product units and gross profit relative to the flooring market, rather than simply growing because it opens stores? Management cites private sources that investors cannot reproduce. A consistent retail sell-through share series, adjusted for channel markup, is needed.
4. Buyback execution: Will remaining repurchases be funded from normalized post-maintenance FCF or incremental debt? At what lease-adjusted leverage and liquidity level would the board stop? The first tranche below $50 looks rational; exhausting the authorization during a prolonged trough may not.
5. Smaller-format pivot: Can 55,000-square-foot boxes preserve 4,200-SKU breadth, same-day job-lot depth and Pro convenience while reducing capital? Disclose opening cost, inventory, rent, sales ramp and cannibalization versus the legacy 76,000-square-foot box.
6. Normalized margin: Where does operating margin settle—near 7%–8%, closer to the corrected 9.3% FY2022 peak, or near the current 5.8%? Gross margin has held, but SG&A absorption and product deflation determine the answer.
7. Commercial returns: Why did commercial segment operating margin fall from 9.8% in FY2023 to 5.1% in FY2025 and 1.0% in H1 2026? What returns have Spartan and follow-on acquisitions earned on approximately $102M of disclosed purchase consideration?
8. Incentive design: Why was the former minimum ROIC gate removed? With ROIC now only 10% of total annual LTI value, how will the board prevent EBIT growth bought through low-return stores or leverage?

14. What Must Be True (Bull and Bear, with Falsification Tests)

Bull thesis—“cyclical trough in a capable consolidator.”

• Must be true: Existing-home turnover stabilizes and eventually improves; reported comps become positive; adjusted gross margin holds around 43%–44%; SG&A re-leverages; smaller new stores preserve productivity with less capital; Pro growth stays above total growth; and lease-adjusted ROIC ultimately clears a roughly 9%–10% hurdle.
• Falsification test: Once existing-home sales show a sustained year-over-year recovery, FND fails to report positive comps for two consecutive quarters; or Pro growth falls below total growth; or adjusted gross margin erodes beyond temporary product mix; or disclosed mature/small-format returns remain below the cost of capital. Any one would materially weaken the bull; persistent comp and return failure would break it.
• Current score: Not met. FY2025 comp was −1.8%, H1 2026 −2.9%, Q2 −2.1%, and early Q3 −2.2%. Q2/H1 SG&A rate increased 120bp/130bp year over year, and lease-adjusted return remains roughly 5%–6%. Sequential monthly comps are encouraging but insufficient.

Bear thesis—“maturing, capital-hungry concept at a full trough-earnings multiple.”

• Must be true: Housing remains slow or FND fails to participate in its recovery; new stores keep diluting reported comps and AUV; LVP deflation persists; commercial margins remain weak; and normalized EPS stays around $2 while the market withdraws the growth premium.
• Falsification test: FND reports at least two positive comp quarters that outperform a defensible flooring-market benchmark, adjusted operating margin rises on that growth, and management discloses new-store paybacks at or better than pre-2022 cohorts. That combination demonstrates cyclical weakness plus value-creating share gain and breaks the bear.
• Current score: Also not proven. The housing trough supplies a plausible external explanation; gross margin resilience, Pro growth and lower capex are favorable; and the current valuation is much less demanding than it was. Bears still need to show that the weak cohorts remain weak after demand normalizes, rather than extrapolating trough productivity forever.

The single fact that resolves both: the reported comp line relative to the flooring market, paired with disclosed cohort returns, over the next two to four quarters. Positive, peer-beating comps with rising lease-adjusted returns confirm the bull. Negative comps while end demand recovers, alongside sub-WACC small-format cohorts, confirm the bear. Ex-cannibalization commentary alone cannot settle it.

15. Source Appendix

Floor & Decor Holdings, Inc. (NYSE: FND) · Research current through 2026-09-03. Primary sources take precedence; third-party calculations were reconciled where possible.

SEC filings (EDGAR, CIK 0001507079)

• FY2025 Form 10-K, filed 2026-02-19—business model, categories, stores, supplier network, cohorts, five-year financials, leases, acquisitions and risk factors.
• Historical Forms 10-K: FY2021, FY2022, FY2023, and FY2024—reconciled operating, cash-flow, store and acquisition history.
• Q2 FY2026 Form 10-Q, filed 2026-07-30—current operations, categories, segments, buyback, leases, debt, tariff accounting and cash flow.
• Q2 FY2026 earnings exhibit, dated 2026-07-30—adjusted results, reconciliation and FY2026 outlook.
• June refinancing 8-K, term agreement, and ABL agreement—maturity, pricing, capacity and covenant analysis.
• 2026 proxy statement, filed 2026-03-23—board, ownership, compensation metrics, related parties and revised PSU design.
• Taylor Form 4, filed 2026-08-10—non-10b5-1 option exercise and sale; Paulsen purchase and Langley purchase, filed 2026-05-04; Adamson tax withholding, filed 2026-08-06.
• Leadership/event filings: Trevor Lang retirement, Luke Olson resignation, Paulsen appointment, and CEO/Executive Chair succession.
• Derivative settlement 8-K, filed 2024-09-24, plus the FY2025 10-K and Q2 2026 10-Q for resolution and current litigation status.
• The 60-month filing review covered 5 Forms 10-K; 15 Forms 10-Q; 39 Forms 8-K plus one amendment; 5 DEF 14A, 5 DEFA14A and 1 PRE 14A; 9 Forms 3 plus one amendment; 196 Forms 4 plus one amendment; 5 Forms 5; and 1 Form S-8. No restatement, material accounting-policy change or unresolved material litigation was identified. FND consistently disclosed two operating segments aggregated into one reportable segment.

Earnings calls and company materials

• Q2 FY2026 call transcript, 2026-07-30—monthly comp cadence, Pro/digital/commercial strategy, cannibalization, LVP conditions, refund flow-through and outlook assumptions.
• Q2 FY2026 results page, 2026-07-30—official company release and links.
• Q1 FY2026 earnings exhibit, 2026-04-30—weather, negative comps, smaller stores, initial guide, tariffs and capital priorities. The FY2025 results discussion was reconciled to the FY2025 Form 10-K and official year-end earnings materials.

Quantitative market data

• AZI adjusted daily price history, through 2026-09-02—price path, return windows, ranges and moving averages.
• AZI fundamentals, dated 2026-09-02—own-history valuation percentiles. Headline earnings were normalized against filings.
• FactorsToday FND loadings, dated 2026-07-31/09-01; risk leaderboard and methodology—factor sensitivity, volatility, drawdown and risk-adjusted history. Factor labels are statistical exposures, not causal claims.

Industry, housing and peers

• NAR July existing-home sales, 2026-08-11, and July pending-home sales, 2026-08-18—turnover and forward demand.
• Freddie Mac PMMS, 2026-08-27—30-year fixed mortgage rate. Harvard JCHS LIRA, 2026-07-23—remodeling spending outlook.
• Floor Covering News 2025 market statistics, 2026-07-01—flooring TAM, hard-surface SAM and product capacity. Laminate channel analysis, 2026-07-06—directional channel-share evidence.
• US Commerce ceramic-tile determination, 2025-04-17—product-specific duties that remain separate from IEEPA treatment.
• LL/F9 acquisition and store-closure announcement, 2024-09-06; Lumber Liquidators current company page; and Bed Bath & Beyond Q2 update, 2026-08-04—capacity contraction, surviving chain, and agreed brand-asset transaction.
• Home Depot Q2 FY2026 results, 2026-08-18; Lowe’s Q2 FY2026 results, 2026-08-19; and Tile Shop FY2025 Form 10-K, filed 2026-02-26—peer demand and specialty-economics cross-checks.

Third-party valuation, price, industry and factor data are non-primary and definition-sensitive. Filing-derived facts take precedence. Management commentary is treated as a hypothesis and tested against reported comps, margins, cash flow, industry data and observable peer results.