Example — fictional company
The Action Map: Tidewater Millwork
The Map, plus a prioritized, priced list of improvements — with the top three already done.
Summary
We spent three hours with three of Tidewater's people — Dana, the estimator; Luis, a shop foreman; and Priya, who runs install. Each hour's findings shaped the next hour's questions, which is why the picture below looks a fair bit different from the X-ray's guesses. We found 52 improvements worth naming, sized each by estimated yearly effect, and built the top three before handing this back.
The interviews ran back to back, one Tuesday afternoon, with both of us — the two consultants — in the room and our agents listening to a live transcript in the background. That's the part that makes this quicker than a normal consulting engagement: by the time we sat down with Luis in hour two, we already had Dana's answer to "what breaks most often" in front of us, and could ask him directly whether the shop floor's version of that story matched hers. It didn't, entirely, and that mismatch is most of what's interesting in this document.
People interviewed
3
Hours of interview
3
Improvements found
52
Est. yearly effect, top 10
~$118,000
The map
How work actually moves through Tidewater, by lane, with the five hot spots we found. Numbers correspond to the findings below. The dashed line is the change-order loop — the path a client's post-approval change is supposed to take back into production, and the one most likely to drop it.
What the map shows
The overall shape held up: quote, drawings, shop, finishing, install, invoice. But three things surprised us, and we'd rather tell you that than pretend the X-ray nailed it.
First surprise: the change-order problem isn't a shop-floor communication gap, it's a sales-to-shop gap. Dana said the change almost always gets written down correctly — it just gets written down in an email to the client, not anywhere the shop looks. The fix isn't "get the shop to check email more," it's routing the change into what the shop already reads.
Second surprise: finishing, not the shop floor, is where the worst schedule slip actually compounds. Luis's shop hits its dates more often than we assumed; the finishing crew is smaller and shared across jobs, so a one-day shop delay becomes a three-day finishing delay because they've moved on to the next job's queue. The X-ray didn't call this out at all.
Third surprise: install scheduling by text is a real cost, but it's smaller than we guessed and mostly self-correcting — Priya already keeps a personal backup calendar. The bigger install-side issue is that punch-list items get resolved verbally on site and never make it back to the office, so the same small defect sometimes gets "fixed" twice on two different callbacks.
None of these are dramatic corrections — the X-ray's overall shape of the business held up fine. But each one changes where we'd point a fix. "The shop doesn't check email" would have led us to build a shop-floor notification system; "the change gets written down in the wrong place" led us to build a form that writes to a place the shop already reads, which is a smaller, cheaper, more durable fix. That's the value of the hour with Dana: not a new fact so much as the right-sized version of a fact we already had.
We'd also flag one thing that surprised us in the other direction — something the X-ray worried about that turned out to be a non-issue. The estimating spreadsheet's bus-factor risk, which the X-ray called out as a medium-confidence concern, is real but less urgent than we assumed: Dana has, on her own initiative, already talked a junior estimator through the basics twice. It's not documented anywhere, which is still worth fixing, but it's not the ticking clock the X-ray implied.
Done already: the top three
1. A change-order form that posts straight to the shop's daily sheet
Built a short form (five fields: job, item, what changed, who approved it, date) that, on submit, appends a line to the shared daily production sheet both shops already open every morning — no new login, no separate tracker.
Time saved: an estimated 3–5 hours/month of rework and re-explaining, plus the miscuts this was meant to prevent — roughly $8,000–$12,000/yr.
Maintaining it: no upkeep beyond keeping the daily sheet where it is; the form is a thin layer on top of what already exists.
We built this one first because Dana and Luis agreed on it independently, in separate hours, without seeing each other's answers — that kind of agreement across two people who don't normally compare notes is about as strong a signal as we get in three hours.
2. A script that turns the estimating spreadsheet into an accounting import
Built a small script that reads Dana's spreadsheet export and writes a file in the exact format your accounting system's import wants, matching line items to the right categories automatically.
Time saved: roughly 2–3 hours/week of manual re-entry, call it $5,000–$9,000/yr, plus fewer transposition errors on invoices.
Maintaining it: if either spreadsheet's column layout or the accounting system's import format changes, the script needs a small update — flagged in the script itself with a comment showing exactly what to check.
This one is deliberately boring. It doesn't change how Dana works, doesn't ask the office to learn a new tool, and doesn't touch anything client-facing. That's on purpose: the fastest fixes to actually land are the ones that remove a step nobody liked doing rather than the ones that add a step somebody's supposed to remember.
3. A written, agent-assisted procedure capturing Dana's estimating rules of thumb
Spent forty minutes with Dana walking through how the spreadsheet actually prices a job — the shortcuts, the rounding habits, the "if it's got a curve, add 15%" rules that live in her head, not the formulas. Wrote it up as a plain-language procedure a second person could follow, with the spreadsheet as reference, not replacement.
Time saved: not a weekly number — this is risk reduction. It turns a full-stop bus-factor risk into a "slower, but survivable" one.
Maintaining it: review it with Dana every quarter or whenever pricing assumptions change; it's a living document, not a one-time capture.
Dana's own reaction, roughly quoted: "I didn't realize how much of that was just... in my head." That's common. Most experts don't experience their own judgment as a set of rules — it feels like common sense to them, which is exactly why it's never been written down, and exactly why it disappears the day they leave.
The Action Map
We sized every one of the 52 items the same way: what it probably costs you not to fix it, over a year, in either hard dollars or hours we converted to dollars at a rough blended labor rate. Items under about $10K/yr individually still made the list if they were cheap or fast to fix — a $3,000/yr problem with a $0 procedure fix is still worth doing, it just doesn't carry the map on its own. The list below is the top 15, ranked by that estimated effect; "procedure only" items are free — write-ups you can act on yourselves, no build required from us.
| # | Improvement | Area | Est. yearly effect | Effort | Our price | Kind |
|---|---|---|---|---|---|---|
| 1 | Change-order form posts to shop daily sheet | Sales → Shop | $10,000 | S | Done | fix |
| 2 | Estimating-to-accounting import script | Office | $7,000 | S | Done | script |
| 3 | Estimator's rules-of-thumb procedure | Sales / Est. | Risk reduction | S | Done | procedure |
| 4 | Shared finishing queue visible to shop & office | Finishing | $14,000 | M | $4,500 | connector |
| 5 | Single-link current shop drawing per job | Drawings | $9,000 | M | $4,000 | connector |
| 6 | Punch-list capture on a phone, synced to office | Install | $8,500 | M | $3,500 | fix |
| 7 | Shared install calendar, one owner per date | Install | $6,000 | S | $0 — procedure only | procedure |
| 8 | Weekly two-shop sync on in-flight changes | Shop | $5,500 | S | $0 — procedure only | procedure |
| 9 | Standard job-naming convention across systems | Office | $5,000 | S | $1,500 | fix |
| 10 | Auto-flag quotes open >10 days with no response | Sales / Est. | $4,500 | S | $2,000 | script |
| 11 | Material-order reminder tied to approved drawing date | Shop | $4,000 | M | $3,000 | connector |
| 12 | Callback log shared between install & office | Install | $3,800 | S | $1,800 | fix |
| 13 | Finishing crew capacity forecast, 2 weeks out | Finishing | $3,500 | M | $4,200 | fix |
| 14 | Written checklist for shop-to-finishing handoff | Finishing | $3,200 | S | $0 — procedure only | procedure |
| 15 | Client-facing status page, read-only link per job | Sales / Est. | $3,000 | L | $7,500 | fix |
Items 16–52 are in the appendix.
A note on effort: S/M/L reflects our time to build and hand off, not your time to adopt. Most S items are a form, a script, or a short procedure — a day or two of our work, little to no change to how your people already do things. M items usually touch two systems or two teams and need a short check-in partway through to make sure we're building the right thing. L items, like the client-facing status page, are closer to a small piece of custom software; worth doing, but they're the ones we'd sequence last and confirm with you before starting.
Keep going
The fastest path from here isn't more consulting hours, it's putting this loop in one of your own people's hands. Step 4, Keep Going, teaches whoever you'd pick — Dana would be a natural fit, given how much of this touches estimating — to run a lightweight version of this cycle monthly with their own agent: a short interview, an updated map, picking the next two or three items, and shipping a fix. About two hours a week for the first month while it beds in, then lighter after that.
Concretely, that means Dana (or whoever) spends an hour a month talking to a colleague the way we talked to her, with an agent listening and proposing follow-up questions the way ours did. She picks the two or three items from the growing backlog that matter most that month, and either builds the small fix herself with the agent's help or hands it to whoever on her team is best positioned to. We stay reachable for the things that are genuinely hard — the L-effort items, the ones that touch more than one system — but the day-to-day loop doesn't need us in the room anymore. That's the point: the first Action Map is us doing it for you: fifteen items, three fixed the first time. By month three or four of Keep Going, it's mostly them doing it themselves, with us as backup rather than as the bottleneck.
Items 16 through 52 are exactly the kind of backlog that loop is built to work through — none of them urgent enough to justify a rushed fix today, several of them worth ten or fifteen minutes of Dana's time with her agent next month.
About this example
Tidewater Millwork, Dana, Luis, and Priya are invented for this page, so you can see the shape of an Action Map before committing to one. The findings, the map, the prices, and the 52-item structure follow the same format we'd use for your business — the specifics would come from your people, not ours. We haven't padded this example to look more impressive than a real one would; a real Action Map for a business your size might turn up 35 items or 70, and the top three fixes we build might be smaller or larger than the ones here, depending on what your experts actually say in the room.
Start with a free X-ray, or read the ladder in full on the offer page. The first few Action Maps are free right now, in exchange for a frank debrief afterward.