Engineer room · P&L only

Dynatrace × Kell

Public DT scale · if → then → \$ · no mill math · pnl.txt · not affiliated · not a forecast

Public scoreboard

FY26 ARR≈ \$2.05B

FY26 revenue≈ \$2.02B

Non-GAAP op margin≈ 29% (≈ \$592M)

GAAP op margin≈ 12% (≈ \$245M)

FY26 FCF≈ \$529M

FY27 ARR guide≈ \$2.38–2.40B

FY27 rev guide≈ \$2.32–2.34B

Cloud GM note~100 bps headwind (consumption > ARR lag)

Source: Dynatrace FY26 results / FY27 guide (IR, May 2026). Illustrative math uses that public scale.

If now → then Kell =

IF cloud hosting is the ~100 bps gross-margin headwind on ~\$2.3B revenue

THEN Kell on the hot land / retain path

NOW weather ≈ \$23M/yr P&L pressure class
IF claw back 10% → ≈ \$2.3M/yr gross profit
IF claw back 25% → ≈ \$5.8M/yr
IF claw back 50% → ≈ \$11.6M/yr
% = measured on one lane — not promised

IF consumption grows faster than ARR and COGS hits before revenue catches

THEN denser land · less drop · less reprocess

= protect the ~29.5% non-GAAP margin guide from further COGS creep
= shorter lag between cost and expansion revenue

IF a regulated / private buyer (Japan·APAC·gov·finance) stalls on egress

THEN zero-egress / on-prem door beside the stack

IF keep/unlock 0.05% of \$2.05B ARR → ≈ \$1.0M ARR
IF 0.10% → ≈ \$2.1M ARR · IF 0.25% → ≈ \$5.1M ARR
one enterprise logo can clear that band

IF reopen / backfill burns SE and eng hours

THEN receipt-closed land

IF 5 eng × 4 hr/wk × \$150/hr × 48 wk → ≈ \$144k/yr opex
IF 20 eng → ≈ \$576k/yr · IF 50 eng → ≈ \$1.4M/yr

IF Tokyo / director ear wants one clean ask

THEN 90-day one-lane pilot · P&L scorecard only

Pilot cost class \$50–150k (NDA) vs unlock ≥ \$1M ARR or ≥ \$2M COGS-path → ROI clear on paper

Say this in the room

DT is a \$2B+ ARR public company. Earnings called out about 100 bps of cloud gross-margin pressure — call it \$20M+ a year of P&L weather on a \$2.3B revenue base. Kell is not a science demo. If we claw back even 10–25% of that weather on the hot land path, that’s a few million of gross profit. Or if one regulated logo needs a private path, that’s million-class ARR. We measure drop, \$/GB hot, and deal stage — not mill math.

90-day scorecard

1. Cost	\$/GB hot (or lane COGS) — before / after
2. Waste	drop / defer % — before / after
3. People	reopen hours / week — before / after
4. Growth	regulated deals advanced × \$ARR
Kill	nothing moves by day 60 → stop

Net ≈ Δ gross profit (COGS) + Δ ARR × contrib. margin + Δ opex hours − pilot cost. Replace illustrative \$ with DT internal actuals before any external claim.
IR: FY26 results