Program Measurement · Finance

Measuring impact across a product portfolio

Three products, none live yet. Figures are estimates from the people who do the work, and should grow rather than settle.

What this is

A sanitized sample · structure, not a result

Figures

Replaced or rounded · illustrative only

Unit of work

One row per metric

Owner

A named human per row

For the people filling this in

Measure from where we are today, not from before the ERP migration. That migration moved these teams to a less efficient state. Getting back to a steady state counts. Take it from the peak of the problem to the future state, and do not net off what we used to have.

Send back two things per row, and do not multiply them together. First, time saved in minutes or hours, on a stated interval (per day, per month, per cycle). Second, the pay rate or salary band you would assume for the people doing that work, for example ~$120K for managers or a different figure for finance SMEs. Danielle applies the rate card, so the total can be traced back to its inputs.

Not every row converts hours into dollars. Some are about how often something goes wrong and what it costs when it does, rather than time spent. Where the two columns do not fit the row, say so rather than forcing a number into them.

Everything here is a swag. It is meant to be defensible, not exact, and it will be replaced with real data once the tools are live.

01 · Finance portfolio

Highlighted outstanding items are what to chase first. A grey ? means the size is genuinely unknown. Usage metrics will be added once the products are fully live and in use in production.

Product and scale	Category	Metric and its components	Where this stands	How we got therecomputation and sources	Outstanding itemshighlighted = start here	Owner	Time saved hours or minutes, per interval · if relevant
Cash ManagementNot in market · Stage 3 Estimates cover all LN accounts — roughly 60 payment accounts globally plus the US disbursement account. Volume is concentrated, with 70–75% running through the US disbursement account. Colloquially "cash app." Value realisable at Stage 4b · matching + file load	Cat 2 · Indirect	1Fraud risk mitigationCash that cannot be reconciled in time is cash nobody is watching. Unauthorised debits can run for months before anyone notices. Fraudulent debits detected sooner Est Auditor fees — described, not quantified Reputational and investor exposure — described, not quantified	$600K annually Annualised risk expectation, not recoverable savings. Individual events run to millions. Some losses are later recovered.	Computation$600K a year, agreed as an annualised risk expectation. Not a per-event figure. When an event does occur it runs to several million, and a sister division lost that order of magnitude to unauthorised direct debits that went unnoticed because cash could not be reconciled in time. The $600K spreads that across the years in which nothing happens. Some losses are later recovered. It also excludes the costs that follow a material misstatement — additional auditor fees, remediation effort, and exposure on published financials — none of which are quantified here.Agreed with Finance in session. Not derived from inputs.	Auditor fee impact of a material misstatement Settled in session: $600K annually is the agreed figure, no further confirmation needed. Do not add this to the labour figures — it is an avoided loss, not a recovered cost. A sister division has real historical examples. Things to consider Do weak reconciliation controls raise our audit fees, and by roughly how much?	Finance leader Controller	Already answered: $600K a year. Nothing to fill in here.
	Cat 1 · Direct	2Reconciliation labourMatching bank activity to the ledger by hand, every day, plus the back-and-forth with AP when something does not tie. Line-by-line reconciliation Est AP back-and-forth — 12.5 hrs a month Est	~$390K / yr 13,150 hours, one onshore analyst at $80/hr and the rest of the team offshore at $20/hr ↗ ~$750K group-wide Estimated	Computation250 hrs/week × 52 = 13,000 hrs/yr AP back-and-forth: 50 emails/month × 15 min = 12.5 hrs/month = 150 hrs/yr One onshore 2,080 hrs × $80 = $166.4K Offshore rest 11,070 hrs × $20 = $221.4K Total ~$390K The 250 hrs is total reconciliation effort today, and auto-reconciliation removes essentially all of it. Where a journal entry is still required, as with FX wires, that entry has to be made whether the tool exists or not, so it does not reduce the saving. The tool reconciles the wire itself and generates the journal template. Rate assumption. The reconciliation team sits in an offshore shared-services centre apart from one onshore analyst, both locally and group-wide. Treating the whole team as offshore is a conservative read and we believe a realistic one. The onshore analyst is costed at a full 2,080 hours. Rolled out across all four divisions, the same mix against 15 FTE gives roughly $750K / yr.Agreed with Finance in session.	AP ↔ cash management back-and-forth: "the AP back and forth is significant." Possible input from the AP lead FX wire variance handling is out of scope for quantification (agreed in session, 2026-09-03): the time saved is not material enough to be worth estimating. This is a measurement decision only. The FX journal-entry export itself stays: the cash team asked for it and it is built. Things to consider	Controller Finance SME
	Cat 2 · Indirect	3Commercial risk mitigationPayments that fail without anyone noticing, until the vendor chases us. The damage runs from support effort and reissue work through to a vendor stopping service. Failed payments caught sooner — an estimated 10% fail undetected Est Double payments avoided — only a handful known Vendor relationship and reputational exposure — described, not quantified	5 failed payments a week roughly 250 a year Counted as vendors contacting us after a payment failed and we had not realised. Cost per incident not yet established. Not quantified: the tail risk is a vendor stopping service, with the product impact and reputational damage that follows.Each failure caught early also avoids a poor experience for the vendors and technology providers we pay, a manual support interaction and reissue work. Estimated	Computation5 vendor escalations a week × 52 = 250 a year. No cost per incident given, so the row is not costed. Finance separately estimates ~10% of payments fail undetected. That is a rate without a volume base, so the row is built on the weekly escalations instead.Agreed with Finance in session.	Cost per incident — the one missing input The stated tail risk is a vendor stopping service and products being affected. Real, but no frequency attached, so not carried into the number. Confirmed in session as worth keeping, and as the framing that lands with executives: "when we talk about commercial risk, we're having issues paying vendors and that's affecting the business." Quantify by estimated reduction in event frequency (2026-09-03). Things to consider If we reduced those by 40 to 60 per cent, what would that be worth?	Controller · Finance SME	How often does this happen in a month, and what does each one cost us?
Capital Workflow/PPM HubNot in market · Stage 3 over 10,000 active projects in the ERP, ~10 new/month. Confirmed with Finance Value realisable at Stage 4b · closed loop Additional value drivers, not in any computations to the right One system of record, instead of spreadsheets and email chains Immutable records and audit history, so SOX evidence exists rather than being assembled Visibility into where coding practice needs correcting Manager and resource-leader value, from people who had no hand in these estimates The figures opposite came from the finance operations side. Finance’s own view is that they are conservative for that reason.	Cat 1 · Direct	4Daily data entry and manual work for projects, allocations, and forecastsThe operations analysts keep project assignments, allocations and forecasts accurate in the ERP by hand, every day, through a documented six-step process. The overhead is there even when nothing goes wrong. Reclass and record changes, 25 hrs a week Est Forecast maintenance and the annual build Est Manager time: spreadsheets, email, working out what to action Est	~$315K / yr 40 hours a week from the two operations analysts, plus manager time on 250 changes a week Estimated	Computation250 record changes a week at 5 min each Reclass and record changes: 25 hrs/week Forecast maintenance: 10 hrs/week Annual forecast build, 50% of 10 hrs/week: 5 hrs/week Team’s stated total: 40 hrs/week 40 × 52 = 2,080 hrs, +10% for other groups = 2,288 hrs 2,288 × $60/hr = $140K Manager time, not previously counted. 250 changes a week at ~10 min of manager time each = 2,167 hrs/yr. At a placeholder $160K a year ($80/hr) that is ~$175K. The 10 minutes is ours, not Finance’s: managers see the sheet the analysts send and generally know what needs changing, so this is a few minutes of judgement rather than a research task. The $160K is a placeholder pending a manager rate from the Innovation Office. The analysts stay at the $120K figure the team supplied. $140K + $175K = ~$315K Still uncosted: managers seeing where a team is fully allocated and where the gaps are, faster onboarding, and one standard view across every manager in the business. A resource leader gave strongly positive feedback on an early dashboard view of this work. That view carried none of the workflow this metric costs, which is a reason to read these figures as a floor.The product team, in session.	Operations analysts: hours per day, and per month-end Manager time saved, potentially ~1 hr/month per manager, to be quantified Gap volume per cycle and cycle frequency Manager time covers spreadsheet navigation, the email back-and-forth with the analysts, and working out which sections to action. It excludes approval routing, which does not happen today at all, so it is an enhancement rather than a saving. The 1 hr/month figure is illustrative, not an estimate. Things to consider How many hours a day do the analysts spend on this today? How many managers are in scope, and what would you put on their time per month? How many gaps come up per cycle, and how often does the cycle run?	Product team lead · Delivery lead
	Cat 1 · Direct	5Month-end and year-end processingMonth-end corrections and reconciliations, plus creating new projects, concentrated in the October to November bulk load. Month-end reconciliation, roster against ERP forecasts Est Project creation: the year-end bulk load Est Project creation: ongoing through the year Est	~$12K / yr 200 hours a year at $60/hr ↗ project creation saving expected to exceed this Estimated	ComputationMonth-end reconciliation of roster against ERP forecasts takes 2 days a month; half is saved, so 12 days = ~100 hrs/yr Project creation takes 200 hrs/yr today; half saved = 100 hrs/yr 200 hrs × $60/hr = $12K The reconciliation itself is not being replaced. The analyst still checks that everyone on the roster is allocated at 100%; the tool removes the manual comparison, not the control. The team expects the project-creation saving to run past 100 hrs/yr once the cheat sheets go and dropdowns are pre-populated, but has not sized it.The product team, in session.	Month-end correction and reconciliation hours Project creation hours, both the year-end bulk load and ongoing through the year Year-end here means project creation specifically, not year-end processing generally. The Oct–Nov bulk load is the highest-volume moment and the primary focus; creation continues through the year and benefits too (no cheat sheets, dependent dropdowns pre-populated, lower error risk), but that is the smaller half. Things to consider How many hours go into month-end corrections and reclasses? How long does the October to November project-creation bulk load take today, and how much creation happens across the rest of the year? What does removing the cheat sheets and pre-populating dropdowns save per project?	Product team lead · Delivery lead
	Cat 1 + 2 · Both	6Audit and complianceOne analyst sends hundreds of sheets to auditors with extensive back-and-forth, and the team lead spends more time again chasing managers on Teams and email before getting them on a call to explain why they rejected the data. Direct: The analyst's time on the audit sheets and the back-and-forth Est Direct: The lead chasing managers for the reasoning behind a rejection Est Indirect: stronger SOX control assurance Unsrc	~$2K / yr 30 of the 150 hours in scope, at $60/hr ↗ ~$9K if contractors and similar groups come into scope Estimated	ComputationSOX manager-review emails: twice a year, ~62 hrs each = 125 hrs/yr SOX audit preparation: 25 hrs/yr, 80% saved = 20 hrs (analyst) Lead chasing managers: 20 hrs/yr, halved = 10 hrs 150 hrs in scope. The team applied a 20% factor, giving ~30 hrs/yr 30 × $60/hr = $2K That 20% is a coverage factor, not a view on the mechanism. Their note says the full 150 hours is reachable once contractors and similar groups are brought into scope, which at $60/hr is ~$9K. The direct saving here is small. The value the team names is control assurance: more complete and consistent SOX documentation, evidence gaps found before an auditor finds them, and less manual effort for control owners. None of that is costed.The product team, in session.	The analyst's hours on the sheets and the back-and-forth — the easiest number here, since the time is known precisely The lead's hours chasing managers before a conversation can even happen SOX scope support, as an indirect benefit This row carries both direct and indirect value, which is why it is not labelled as one or the other. Routing changed in session: this sits with whoever coordinates audit deliverables with functional management, rather than with the audit relationship itself. Things to consider How many hours go into sending the audit sheets and the back-and-forth each cycle? How much time goes into chasing managers on Teams and email before you can get their reasoning on a rejection? Where does this help SOX scope, beyond the time saved?	Product team lead
Month-endNot in market · Stage 3 CER = Phase 1, RF = Phase 2, one pipeline. App deployed on Databricks Value realisable at Stage 4	Cat 1 · Direct	7Time saved getting a trusted answerOps and decision-support people trying to pin down a financial fact from the last month-end review: finding the right deck, opening it, reading through it. Asking is faster, if the answer can be trusted. Time saved answering finance queries Est Answer accuracy against source, the quality gate on all of it Est Meeting hours reduced as people arrive already briefed — countable once it is in production and meetings actually shorten	~$1.54M / yr 11,340 hours, about 5.5 FTE-equivalent ↗ 250 users is a floor, not a ceiling	Computation250 users × 60 min per task, at segmented rates: Exec heavy 5 × 10/mo × 12 × 1.0h × $400 = $240.0K Exec rest 20 × 1/mo × 12 × 1.0h × $400 = $96.0K Senior heavy 15 × 10/mo × 12 × 1.0h × $200 = $360.0K Senior rest 100 × 1/mo × 12 × 1.0h × $200 = $240.0K Staff 125 × 5/mo × 12 × 1.0h × $80 = $600.0K Total $1.54M · 11,340 hrs · 5.5 FTE The 250 covers finance staff plus named leaders in the senior leadership and executive tiers. It does not include others across the wider management reporting chain who would likely use it, so this population is a floor rather than a ceiling.Headcounts and rates agreed with Finance in session. Time per task from the Innovation Office survey. Volume is our own assumption.	Volume is our own assumption, and it is the biggest lever on this number Downstream work a question displaces, which the method does not count Merges what were two separate rows, answer trustworthiness and time to reach a decision. Answer accuracy is not a separate row: it is the precondition, since a CFO cannot act on an answer they cannot trust. Headcount from HR, rates and method from the Innovation Office rather than rebuilt. Things to consider How long does it take to answer one of these questions today?	Programme owner Innovation Office (method)
	Cat 2 · Indirect	8Strategic decision qualityAcquisition and cash-position decisions carry real cost when they are made on incomplete or stale information. This is about getting them right, not about getting to them faster. Acquisition decisions made on complete information Cash position decisions Cost of a decision made on stale or partial information	Not yet quantified P3 · after 15 Sept		What a strategic decision made on incomplete information has cost us before Different in kind from metric 7, not a smaller version of it. That row is ops and decision-support people retrieving a fact. This is the small number of consequential decisions where being wrong is expensive. Kept visible because it is where the strategic value sits, even though it cannot be measured yet. Things to consider Which strategic decisions would have gone differently with better information to hand? What did getting one of those wrong cost?	Not yet assigned	How often has a decision gone wrong for want of better information, and what did it cost?

Appendix evidence tiers, and the rates and assumptions behind every figure

A1 · Evidence tiers

Every component inside a metric carries a tier. Estimated is SME judgement, dated and attributed. Unsourced is a placeholder with a named owner. Nothing here is measured, because nothing is in market yet.

A2 · Methodology and assumptions

Method reused, not rebuilt. Time saved per task at a segmented labour rate, following the approach the Innovation Office already presents to the executive committee. Its three inputs come from Finance, not from us:
- Time per task: 60 minutes. The Innovation Office survey figure for Claude Enterprise Chat, the closest analogue to a purpose-built agent. Their alternative-tool figure of 45 minutes is the conservative floor.
- Labour rates. Executive tier $400/hr, senior leadership $200/hr, staff $80/hr, all from the Innovation Office. $120K a year (about $60/hr) for the product team, per their own estimate. That $60/hr is the team’s own figure and is used for its three named members throughout. Managers are a separate population, held at $160K a year ($80/hr) as a placeholder pending a manager rate from the Innovation Office. The reconciliation team sits in an offshore shared-services centre at $20/hr apart from one onshore analyst, confirmed by the Innovation Office. Rates exclude share-based payments, so the senior figures understate.
- Population: 250 across the executive tier, senior leadership, decision support and the finance operations team, excluding AP. Scoped by the two function heads, who provided the list of names.
Volume is our own assumption and the biggest single lever. Conservative basis: 10 tasks a month for heavy users, down to 1 a month for occasional ones. At 20 tasks a month the month-end figure roughly doubles.
Measured from today, not from before the ERP migration. Getting back to a steady state counts.
Reusable beyond Finance. Category 2 generalises as the business quantity the tool's domain is accountable for: revenue influenced for a sales tool, cash integrity and decision latency for a finance one.

Questions? Email Danielle Brown →

LexisNexis Embedded Innovation · Danielle Brown · For internal review only · Sanitized sample, illustrative figures · Last updated September 2026
Questions or want to talk through applying it: danielle.brown1@lexisnexis.com