HD · Forward model · McPhail case

The McPhail case, 20 quarters out

Model as of 2026-08-27

Each segment is projected from its own operating driver, rolled up into consolidated cash flow, and discounted back to a fair value per share. The assumptions below are editable — change them and every number on this page moves with them.

Two verticals, two disclosed lines, and nothing apportioned. Home Depot has reported a two-way segment split since it adopted ASU 2023-07: a Primary reportable segment aggregating the U.S., Canada and Mexico operating segments together with the online business and HD Supply, and an 'Other' line holding five non-reportable SRS operating segments. In all fourteen quarters carried here they sum to reported consolidated net sales to the dollar - $42,806M plus $5,055M equals $47,861M at the basis quarter - and the earnings bridge reconciles every one of the fourteen at zero difference. No sub-split is invented: HD Supply inside Primary, Pro versus DIY, organic versus acquired growth inside Other, and the three merchandising product lines are all either undisclosed or too short and too uniform a series to earn a vertical, and none of them appears here. THE QUARTER LABELS ARE FISCAL, AND THE SEASONAL FACTORS FOLLOW THEM. Quarters carry Home Depot's own fiscal labels, the ones its earnings calendar uses, so the basis quarter - the 13 weeks ended 2 August 2026 - is '2026 Q2' here, and both seasonality arrays are written in the fiscal order the engine indexes. Primary carries [0.9753, 1.1161, 0.9885, 0.9201] - the spring peak sits at fiscal Q2 - derived over three complete fiscal-year windows with a worst window-to-window spread of 0.0148 against 0.196 of amplitude. Other carries [0.9794, 1.2137, 1.0328, 0.7740] from two rolling windows. The fiscal 2024 fourth quarter was 14 weeks and is rescaled 13/14 in every factor. Because seasonality is present, each driver rate here is a DESEASONALISED TREND, not a printed sequential step: the engine deseasonalises the basis quarter, runs the driver, then reapplies the factor. Average ticket is left aseasonal deliberately, and that is a finding rather than a gap. Deriving factors on customer transactions alone gives 0.9707 / 1.1232 / 0.9870 / 0.9191 in fiscal order, while average ticket gives 1.0123 / 0.9963 / 0.9941 / 0.9973 - a whole seasonal amplitude of 0.018 index points against window spreads of up to 0.021. The noise exceeds the signal, so all of Home Depot's seasonality lives in traffic and none of it in price. The unit driver is fed deseasonalised inputs for the same mechanical reason: runUnitCurve compounds from units and price and ignores the deseasonalised base, while the projection still multiplies by the factor, so 394.58M and $97.20 go in rather than the printed 443.2M and the printed $92.50 ticket. Where the model sits against the company's own guidance, on the model's own output. Fiscal 2026 net sales come out at $170.8bn, +3.7% on fiscal 2025's $164,683M, inside the reaffirmed 2.5%-4.5% guide and above its midpoint. The second half is $81.17bn against the $79.17bn-$82.47bn the guide implies after a first half that ran +5.3%. The next quarter - fiscal 2026 Q3 - is $42.38bn, +2.48% year over year, and the January quarter is $38.80bn, +1.56%. Projected second-half capital expenditure is 2.52% of projected second-half sales against a guide of approximately 2.5%. Carrying the first half's $2,161M of D&A into the second half and the guided $2.3bn of net interest and 24.3% tax rate, the model implies about $14.80 of fiscal 2026 GAAP diluted EPS on 996M diluted shares - the top of the guided flat-to-+4.0% range off $14.23 - and about 12.7% of GAAP operating margin against a guided 12.4%-12.6%. The model therefore runs at the optimistic edge of the profit guide and in the upper half of the sales guide; it is not tuned to the midpoint of either. Three limitations to hold in view. First, fiscal 2027 lands at $173.0bn, only +1.3%, because the deseasonalised Primary trend is +0.27% a quarter and the fiscal 2026 first half was above that line; consensus adjusted EPS of $16.07 for the year ending January 2028 implies a faster top line than this, and the drivers are deliberately NOT calibrated onto sell-side estimates. No revenue consensus was verifiable from an openable source and none is used. Second, Other grows at organic-plus-branches only. Home Depot has spent roughly $14bn on SRS, GMS and Mingledorff's in three years, gives no forward acquisition guidance, and another tuck-in would put this model below reported revenue without the model being wrong; no acquisition allowance is added. Third, the capital expenditure split - 2.65% of Primary revenue against 1.50% of Other - is judgement, not disclosure. Home Depot does not split capex by segment; the two were chosen to blend to the guided ~2.5% of total sales, so consolidated free cash flow is sound and no per-vertical free-cash-flow claim should be built on the split. Share count is fixed at 998,000,000 and that is deliberate. Treasury stock stood unchanged at 806 million shares between 1 February and 2 August 2026, shares outstanding rose from 995M to 998M year over year, and neither the financing section of the cash-flow statement nor the call's capital-allocation discussion contains a repurchase. Deleveraging - $3,040M of long-term debt repaid in the first half - and the $2.33 quarterly dividend are the uses of cash, so no buyback-driven per-share accretion is modelled. Corporate overhead is zero because the segment reconciliation has no unallocated line: Primary $6,592M plus Other $247M equals consolidated operating income of $6,839M, with only interest and tax below it.

Where each case comes from

From cash flow to fair value

The projected path

The published model's 20 quarters. The horizon and sliders above do not change this table.

Every row is projected. A year-over-year change is shown only where the quarter it compares against exists — an em dash means there is no comparable quarter, not a flat year.

Model revisions

Assumptions are marked to reality as each quarter prints. Every change is appended here, with the fair value the model produced at the time, so the model's own history stays visible.