Measure It Like a Deal
Applies to both verticals. Distribution examples are worked below; the same five lines apply to an agency — substitute COI cycle time and submissions-per-hour for cost-per-order, and mark those figures [VERIFY] until measured against your own baseline.
The problem, in your words
Someone pitches you automation. It'll "save time" and "reduce errors." Maybe it will. But how much, against what, by when?
Without numbers, you're buying a promise. And after it's installed, "it feels faster" is not a result you can take to the bank. You run your business in cash, hours, and turns. Your automation should be measured in the same units — or it shouldn't be funded.
Why it persists
Vendors sell features, because features demo well. Few stop to build the baseline, because the baseline is real work and it isn't their cash on the line. So the project gets judged on enthusiasm in the room, and the number — the only thing that should decide it — never gets written down. A measurement nobody took is a number nobody can argue with later. That's convenient for everyone except you.
The teach — measure it like a deal, in five lines
- Baseline (the "before"). The hard number today: cost-per-order, cost-per-invoice, COI cycle time, hours per task. Measured over two weeks of real data — not estimated, not remembered.
- Target (the "after"). The defensible number you expect, with the attributed benchmark standing behind it — "up to ~80% faster processing — Conexiom" — not a hope dressed as a forecast.
- The bridge. Baseline → target, in dollars and hours, with every assumption shown. Every number shows its work, or it doesn't count.
- Payback. Cost of the build ÷ monthly savings = months to payback. Owners fund a four-month payback far faster than a four-year vision. Lead with it.
- Sensitivity. What if volume runs lower, accuracy lands worse, adoption comes slower? A number that survives the pessimistic case is a number you can actually trust.
The discipline underneath all five: one metric, one definition, measured the same way before and after. No moving goalposts. No black box.
How to measure it
- Payback period — months to recover the build cost.
- Realized savings vs. baseline — the after, in the same units as the before.
- Forecast accuracy — projected vs. realized. Your own forecasting is a number worth tracking; it's how you earn the right to underwrite (see Play 10).
The number it moves
It moves all of them — because measuring it like a deal is how you find out which number actually moved. The point isn't a prettier report. It's that a measured payback is the entire difference between an expense and an investment, and owners fund those two things very differently.
Distribution: cost-per-order, touchless rate, AP cost-per-invoice. Insurance: COI cycle time, submissions per hour — [VERIFY] against your own baseline.
[PLACEHOLDER: MarginArc baseline-to-realized accuracy — to be added]
If a firm can't show you the number before you pay, they're guessing. Make them show their work — then hold them to it.