Decisions to force

Slide 1 · The close

Seven decisions. Not one of them can be closed by more desk research.

Each changes the answer to something else in this pack. Every one needs a person in the room, a paid seat, or a negotiation — not another search.

3×Disagreement in published mfg-ERP TAMConflict

$3.1bnSchneider–CogniteHas not closed

$7bnAviation MRO divergence by 2034Conflict

$2.13bnOctave Intelligence impairmentFirst public quarter

3 of 3Internal capability claims that do not stand upCredibility risk

Slide 2 · The seven

Decide these and the rest of the pack resolves itself.

Decision	What it changes downstream	Why it cannot wait	Owner
1 · Settle the manufacturing-ERP TAM definition	The headline number for IFS's core market. Every derived allocation in Part 1 rests on it.	Published figures disagree 3× — $5.9bn vs $15.2bn vs $17.6bn for 2025. Action: pull Gartner doc 7096933 and IT Key Metrics 5972771 on IFS's own seat before the session.	Strategy / Market Intelligence
2 · OT and semantic data layer: build, buy, or a written partnership boundary	Whether IFS owns the intelligence layer or rents it.	Schneider's $3.1bn Cognite acquisition has not closed; AVEVA's agentic CONNECT release is planned for Q1 2027. The window to negotiate from strength is now. No public statement defines overlap rules or co-sell economics in the AVEVA–IFS partnership.	CTO / Corp Dev
3 · Publish an AI commercial model and a model-dependency position	Pricing, RFP responses, and the equity story.	Six competitors have public positions; three are more buyer-friendly than anything IFS has stated. Already an RFP line item, and with an IPO signalled for 2027–28 it is also an equity-story asset.	CPO / CRO / CFO
4 · Life sciences: in or out, explicitly	Where industrial equipment and A&D capital goes.	Fastest-growing vertical (13%) with the best software-addressability ratio (~24% of IT spend) and the weakest IFS fit. The failure mode is drift — a half-funded push that neither wins regulated accounts nor frees capital. Recommended narrow entry: equipment qualification and validated-state maintenance.	Product Leadership
5 · Resolve the aviation MRO software CAGR: 2.57% or 8.0%	The entire A&D aftermarket business case.	A $7bn divergence by 2034 on the same ~$8bn base. 2.57% was verified arithmetically as a genuine analytical position, not a typo.	A&D Industry / Strategy
6 · Verify Octave Intelligence before the EAM competitive set is redrawn	Who we position against in EAM.	It replaces Hexagon as the EAM competitor, but its FY2025 US-GAAP revenue was not retrievable, it reports as a single unit with no EAM revenue disclosure, and its win-announcement URLs 404. Highest-value external verification item.	Competitive Intelligence
7 · Correct the internal capability record	What we are allowed to say in front of a customer or investor.	Falkonry was never acquired. The Gartner FSM Magic Quadrant no longer exists. "Custom Controls" has no trace in ifs.com's sitemap or newsroom. Each is a live credibility risk.	Product Marketing

Slide 3 · Timing

On decisions 1, 2 and 3 the clock is not ours to set.

Nothing on the left of this line gets cheaper by waiting. Decision 2 is the only one where the counterparty's clock runs out before ours.

Slide 4 · Decision 5

Two published CAGRs, one base, and a $7bn gap by 2034.

Same ~$8bn aviation MRO software base. 2.57% was verified arithmetically as a genuine analytical position, not a typo — which is why this has to be argued, not looked up again.

Higher forecast8.0%

Lower forecast2.57%

Conflict Range given, never an average. Picking one swings the entire A&D aftermarket business case.

Slide 5 · Decisions 6 and 7

Two of the seven are credibility risks, not strategy choices.

Octave Intelligence replaces Hexagon as the EAM competitor. Its first public quarter looks weak — and the parts we would quote are the parts we cannot verify.

−4%Revenue

−23%Licences

$2.13bnImpairment

Falkonry was never acquiredRemove it from any capability slide.

Gartner FSM Magic Quadrant no longer existsRetire the claim, including from partner sites.

"Custom Controls" has no traceAbsent from ifs.com's sitemap and newsroom. Verify or drop it.

Before the EAM set is redrawn: Octave's FY2025 US-GAAP revenue was not retrievable, it reports as a single unit with no EAM revenue disclosure, and its win-announcement URLs 404. Highest-value external verification item in the pack.

Slide 6 · Commission, don't search

Six gaps that no amount of web research will close.

1 · Named industrial OEM service-revenue-mix targetsABB, Siemens, Atlas Copco, Sandvik, Konecranes, Wärtsilä, Caterpillar, Rolls-Royce TotalCare penetration. Needs investor-relations documents and capital-markets-day decks, not search. Highest-value missing input for the servitization GTM narrative in bet 01.

2 · Field service KPI benchmarks beyond first-time fixTechnician utilisation, cost per truck roll, scheduling-optimisation gains. BCG measured these for top performers but publishes no values.

3 · A defensible data-readiness statistic for manufacturingS&P found cost and privacy topping the obstacle list, not data quality. Probe it, do not assume it.

4 · Forecast accuracy (MAPE/WMAPE), stockout rates, OTIF for 2025–26Nothing credible exists publicly. This is the missing quantitative anchor for the planning business case in challenge #2.

5 · A real revenue-band cut of manufacturer countsFrom SUSB or D&B. No source segments manufacturers by revenue band.

6 · Named manufacturing AI case studies with hard, independently verifiable ROIConverts challenge #10 from a posture into proof.

Slide 7 · Do not use

Roughly a third of the numbers in this market do not survive a primary-source check.

Dating a statistic honestly consistently increases credibility against competitors quoting the same number naked.

Most likely single trap in a strategy deck. Panorama "189% average budget overrun", "215% in manufacturing", "73% of discrete manufacturing projects fail", "68% overall failure rate" — these are NOT in the actual Panorama 2026 ERP Report. The primary PDF was read in full: it reports only that over a quarter exceeded budget and almost a quarter exceeded schedule, on a 9-month median timeline. The inflated figures originate on a third-party blog.

"68–75% of ERP projects fail" is not traceable to any primary Panorama or Gartner publication.

Circulating statistic	What is actually true	Use instead
McKinsey "aftermarket 25% vs 10% EBIT margin"	Published 27 July 2017 — nine years old	BCG Feb 2025: services ~2× the 15–25% equipment margin; top performers ~42% gross / ~20% EBIT
"Aftermarket margins up to ten times higher (Bain)"	No Bain primary source reachable. Appears to originate in vendor blogs. "McKinsey: aftermarket margins 25–35% higher" is not in any McKinsey source	The BCG 2025 figures above
Siemens "$2.3m/hour automotive downtime" and "$1.4trn / 11% of Fortune 500 revenue"	Published 2024 but data covers Apr 2019 – Mar 2023, n=181; the $1.4trn is an extrapolation and losses are down 6% since 2022. No 2025/26 update exists	Always date it, and pair with ABB's $125k/hour median, n=3,215
"65% of manufacturers say workforce is their primary challenge"	NAM Q1 2024. Current figure is 46.95%, ranked 6th (NAM Q2 2026)	46.95% / 6th, noting raw material costs are now #1 at 83.1%
Deloitte/MI "1.9m unfilled manufacturing jobs by 2033"	April 2024, fielded Dec 2023, base >200 firms. No 2025/26 refresh	Cite as "a 2024 study projecting to 2033" and pair with live JOLTS
Plant Engineering "88% preventive / 51% run-to-failure / 52% CMMS"	A 2020 study published June 2021 — six years old, widely mislabelled as 2025	Plant Engineering 2026 (67% name predictive maintenance most-critical) and ABB's 21% run-to-fail
"67% of manufacturers still rely on reactive maintenance"	No primary study exists. Appears only in aggregator blogs	Challenge #7's water-utility figure — governmental and unarguable
"23% of unplanned downtime caused by unavailable spare parts (Aberdeen)"	Attributed to Aberdeen with no report title, number or year; not retrievable; Aberdeen's service research is 2010–2015 vintage	Do not use. No substitute found.
OEE "~60% typical for discrete manufacturing"	No locatable primary survey	The 85% world-class figure IS defensible — TPM arithmetic (90% × 95% × 99%), not survey data
Warranty fraud / leakage "3–5% of warranty spend"	No credible source found at all. Widely circulated in warranty-software marketing	The SEC-audited reserve and catch-up-accrual figures in bet 04 — stronger anyway
"Cost of knowledge loss" (IDC "$31.5bn lost by Fortune 500")	Traces to unretrievable 2000s-era white papers	Siemens' 49→81 minute recovery degradation, and the 86%-vs-53% first-time-fix spread
"25% of utility workers retire within 5 years" (CEWD)	Traces to 2012 framing, and CEWD later reported the opposite (9.7% eligible within 5 years, lowest since 2016)	ASCE/EPA: ~one-third of the water workforce retirement-eligible over the decade, median age 48
"Only 5.5% of companies are AI high performers"	McKinsey's actual published wording is "about 6 percent"	6%
BCI Supply Chain Resilience Report 2025 or 2026	Does not exist. Latest edition is 2024	Cite the 2024 data as 2024 — including that Excel is still the #1 tool for recording disruptions
PwC "$5.5trn tied up in working capital"	2021/22 vintage; the global study page is a 404 and no 2025/26 edition exists	Hackett's $1.94trn (2026)
Transformer / switchgear lead times ("128 weeks", "+77% since 2019")	All vendor or SEO content	IEA: "wait times for transformers and cables have doubled in the past three years"; or ASCE 2025's 80–210 weeks with its June 2024 date
Aggregate 2026 hyperscaler capex ($650bn / $690bn / $725bn / $760bn)	Mutually inconsistent aggregator figures	Only Meta's $130–145bn guidance was verified from a primary filing
"~10,000 entities in scope for California SB 261"	CARB's own working list is 4,160 rows / 3,127 unique names	Do not attribute 10,000 to CARB
"Gartner Magic Quadrant Leader for Field Service Management"	The MQ was retired after the 2022 edition and does not exist	2026 IDC MarketScape AI-Enabled Asset-Intensive EAM Leader; 4th consecutive Gartner Peer Insights Customers' Choice for Cloud ERP
Zimmer Biomet v Deloitte "$172m" and Birmingham City Council "£216m"	Consultancy round-ups only; no court filing or council report verified	Do not use externally without reading the primary document
"Industrial AI $280bn by 2035 at 46% CAGR"	Implausible against every peer estimate	Mordor industrial AI software: $23.52bn (2026) → $52.97bn (2031), 17.62%

Slide 8 · Still open

Five figures remain unresolved as of 12 August 2026. Check before external use.

SAP extended-maintenance surchargeThe "+2 percentage points" figure appears only on third-party sites. SAP's own maintenance pages returned empty bodies.

CBAM HICP-indexed penalty per tonne, 2026The formula is verified (€100/t indexed) but the indexed value is not published. Quote it as "€100/t indexed", never as a figure.

Expected CBAM cost per tonne of steelNo sourced figure exists. Any number would be our own arithmetic.

Average S/4HANA migration cost and durationNot found from Gartner, IDC or Forrester under a named analyst.

NIS2 transposition status country by countryOnly Germany's ~29,500-entity figure is documented by a regulator. Do not put a 2026 date on NIS2 for any specific country without checking it.

Slide 9 · How to leave the room

Seven owners, seven dates, and one paid Gartner pull.

If the session ends with named owners against decisions 1–7 and the two Gartner documents ordered, the pack has done its job. If it ends with a request for more research, it has not.

IFS Core Product Strategy — Manufacturing Evidence Pack. Prepared for the R&D and GTM core product strategy session. Cited figures carry a named source and URL in the source pack. Derived and Estimate figures are labelled with their arithmetic. Gaps are stated as gaps rather than filled with plausible numbers.

Internal · Strategy Input · 12 August 2026 · Horizon FY2027–FY2029