Map
Operational Foundations
Real operating numbers, a KPI dictionary, named process ownership, clean data, and workflows documented as they actually run.
The buyer problem
You have numbers. You do not trust them.
The weekly operating report and the monthly close disagree by an amount nobody can explain. Two people define "closed won" differently and both are technically right. A figure on the scorecard is typed in by hand every Monday from somebody's best recollection of last week. When you ask how long a job takes, you get a confident answer and no evidence. When you ask who owns the handoff between sales and fulfillment, the room goes quiet, and the quiet is the answer.
Underneath that, the process itself is undocumented. It exists in the heads of four people, one of whom is leaving. What is written down is a flowchart made in a workshop two years ago. It draws the happy path, because the happy path is the version of the work a room can agree on in an afternoon. The exceptions -- the reason the work is expensive -- are not on the chart at all.
Everything you might want to automate depends on this layer, and this layer is the one nobody gets paid to fix.
Conditions for fit
This is for you if
- You are in one of the three workflow families we work on: lead to cash, request to resolution, or work to financial visibility.
- Fifteen or more people do repeatable process work, or you can point at leakage of comparable size.
- Your systems write timestamps. Almost everything does -- a CRM, a ticketing system, a dispatch tool, an ERP, a billing platform.
- Someone senior enough to settle a definition argument will be in the room. Disagreements about what a number means are decided by authority, not by analysis.
- You can live with the first month being reconciliation rather than anything that looks like AI.
This is not for you if
- The work is bespoke every time.
- The volume is low and the value per transaction is high.
- The real problem is a strategy question rather than an operating one. A firm that measures process will not tell you which market to enter.
We decline work that cannot clear its own benefit hurdle. If the arithmetic on your workflow does not show a return worth the fee, the answer is no, and you get that answer before you spend anything.
The questions we ask first
The diagnostic opens on these, in this order, before anything is proposed. If the answers do not exist, that is itself the finding.
- What is the most expensive thing your team does by hand?
- Walk me through it, start to finish. Where does it stop and wait?
- How many of those a month, how long does each take, and who does them?
- How would you know today if that process broke?
- What does your reporting say this costs you?
- Which system is the authority when two of them disagree?
- Who is the one human being accountable for this workflow end to end?
- What happens on an exception -- and what share of volume is an exception?
- Which numbers does your leadership team actually act on, as opposed to look at?
- If we changed this process next quarter, how would you prove it worked?
Deliverables
A measured baseline
Volumes, cycle times, handoff counts, wait time, exception rate, error and rework rate, and the distribution rather than just the average. Averages hide the tail, and the tail is where the cost is.
A KPI dictionary
Each number defined once and only once, with: name, the decision it supports, its owner, the formula, numerator and denominator, units and period, the authoritative source system, refresh frequency, current baseline, target, quality checks, known limitations, escalation threshold, and the person accountable when it is crossed.
A process map of the real process
Every step, handoff, decision rule, exception branch and approval, drawn where it actually happens rather than where it was supposed to.
A workflow specification
For each process in scope: purpose and customer, trigger, inputs and authoritative records, current steps, proposed steps, roles and handoffs, decision rules, exceptions, permissions and approvals, output and acceptance checks, service expectations, linked KPIs, failure handling and rollback, training, owner and reviewer, version and review date.
A data readiness assessment
Which fields are reliable, which are free text pretending to be structured, which are duplicated across systems, and what has to be cleaned before anything can be built on top.
A named owner for every process in scope
Agreed with you in writing. This one is not optional. Everything else rests on it, so we do not close the diagnostic until every process in scope has a name against it.
A quantified opportunity and a go/no-go
Sometimes the finding is that the work is not worth automating.
Implementation steps
- Week 1 -- pull the events
We start with the timestamps your software already writes: case identifier, event type, timestamp, source system, role identifier, status transition, exception flag, outcome. Read-only access. The event data is the bulk of the evidence, and nobody is observed at all.
- Week 1-2 -- reconcile the numbers
We take the figures on your existing report and trace each one back to the system that should produce it. Where they do not match, we find out why. This is where the definition arguments surface, and it is why we ask for someone who can settle them.
- Week 2 -- sit with the work
Walkthroughs with the people who actually do the process, following a small number of real cases end to end, including bad ones. We are looking for the steps that never made it onto the flowchart.
- Week 2-3 -- fill the gaps
If the event logs cannot answer a specific question, we may propose scoped screen observation. It is optional, aimed at one question, visible to the employee, pausable by them, and time-boxed. If you would rather not, we run the whole diagnostic without it. Anything touching recorded calls, SMS or email is a separate scope with its own permissions and its own legal review.
- Week 3 -- write it down and hand it over
Baseline, dictionary, map, specification, opportunity, recommendation. Delivered as documents you own, in formats you can edit, usable by any firm or by nobody.
Example measures
Examples of what gets instrumented. Each row is a definition: what gets counted, the unit and the period it is counted over, and the system that holds the authoritative figure. A number without those three is not a measure. The figures come from your baseline, measured on your data.
| Measure | Unit and period | Authoritative source |
|---|---|---|
| Cycle time from trigger to completion | Median and 90th percentile | Status-transition timestamps in the system of record |
| Touch time versus wait time | Share of cycle time | Event timestamps, per step |
| Handoffs per case, and the queue depth at each one | Count per case; cases waiting per handoff | Status transitions and role identifiers in the event data |
| Exception rate, and the cost of an exception against the cost of a clean case | Share of volume; dollars per case | Exception flag in the event data |
| Rework hours, and the step that generates them | Hours, per period | Rework transitions in the event data |
| First-response time on inbound, and the share breaching your own stated target | Median; percent breaching | Inbound timestamp to first outbound on the same case |
| Percentage of KPI values sourced automatically rather than typed in | Percent of KPI values | Authoritative source field in the KPI dictionary |
| Reconciliation gap between the operating report and the system of record | Dollars and percent | Operating report against the system of record |
| Days sales outstanding, and the share of that caused by process delay rather than by the customer | Days; share of days | Billing platform against the ledger |
Dependencies
What has to be true on your side. If one of these is missing, the engagement stalls and we will say so before it starts.
- An executive sponsor who can settle a definition dispute in a single meeting.
- Read-only access to the relevant systems, arranged in the first few days. A delay here pushes the whole diagnostic, so it is the first thing we ask for.
- Your controller or finance lead in the room for the financial baseline. Not copied on the output -- in the room.
- Time from the people who do the work. Roughly two to four hours each across three weeks.
- Honesty about the exceptions. If the documented process is aspirational, tell us on day one. We will find out in week two anyway, and week one is cheaper.
- A decision-maker who will accept a no. The go/no-go is only worth producing if someone on your side has the authority to act on a no.
Limitations
What this work does not do. On the page, not in a footnote.
- It does not fix anything on its own. A baseline and a map are measurement, not change. Acting on them is the Build stage or your own team's work, and we will say which we think is better value for you.
- It is management reporting, not audit or assurance. We do not give tax or accounting opinions, we do not issue an opinion on your financial statements, and we do not take custody of your funds. Judgments in those areas stay with your qualified professionals, and we would expect to coordinate with them.
- It cannot measure what your systems never recorded. If the process runs on phone calls and memory, we can tell you that is the finding, but we cannot reconstruct a year of history that was never written down.
- Three weeks is short enough to be affordable and short enough to miss seasonality. A baseline taken in your busiest month is not your average month, and we label it that way rather than quietly annualising it.
- Permission from you is not permission from your employees and is not permission from a third party on a recorded call. Before any production recording we want state- and industry-specific legal review, done by your counsel or one you approve.
The engagement path
- One 25-minute call. Six questions, then the value arithmetic on your worst workflow, out loud, with your numbers.
- If it clears, the next step is the Operating Diagnostic at $12,500 over three weeks, credited in full against implementation if we go further.
- If it doesn't clear, we tell you on the call and neither of us spends another hour on it.