The L3 service manager who completed Lessons 1, 2, and 3 has the receptionist architecture, the lost-call recovery loop, and the CSR show-rate coaching workflow deployed. The three workflows together produce the $1.2M-$2.4M annualized revenue lift at a 7-truck residential shop — but only if the owner can see it and trust it. The dashboard is the artifact that makes the metric movement visible and defensible. This lesson is the build instruction for the L3 Ch2 weekly dashboard: the five headline numbers (booking %, after-hours capture rate, cost per booked call, abandoned-call %, AI-recovered revenue), the three diagnostic supplements, the owner email format, the underlying data pipeline that pulls from ServiceTitan / Sera / HCP / CallRail / Avoca / Jobber AI Receptionist / HCP AI Agents, and the weekly reporting cadence that lands the dashboard on the owner's screen every Friday at 5 p.m. with a 4-minute read. Target: the owner reads the dashboard in 4 minutes, asks zero clarification questions, and walks into Monday morning confident that the L3 Ch2 stack is producing the $20K-$45K of weekly revenue lift the platform CEO promised the PE partner. The dashboard is the deliverable that converts the L3 Ch2 workflows into board-defendable, coach-defendable, peer-group-defendable shop performance.

Why the Dashboard Is the Fourth Leg of the L3 Ch2 Stack

An L3 Ch2 stack with no dashboard is a workflow that nobody can defend. The architecture (Lesson 1) captures 75% Tier-One. The recovery loop (Lesson 2) lifts to 85-88% effective booking floor. The show-rate coaching (Lesson 3) takes show rate from 84% to 92%+. The combined revenue lift at a 7-truck residential shop is $1.2M-$2.4M annualized — architecture primary capture lift $400K-$600K, recovery loop $565K-$830K, show-rate coaching $280K-$365K. Without the dashboard, the owner reads these numbers as service-manager claims, the coach reads them as marketing copy, the PE partner reads them as unsubstantiated EBITDA gestures. The dashboard is what converts the workflows from claimed lift to verifiable performance.

The dashboard also closes the operating loop. The service manager runs four cadences (Lesson 1's weekly audit, Lesson 2's daily AI report and Friday recovery tuning, Lesson 3's Friday pull and Monday huddle, plus this lesson's Friday owner dashboard) and the owner reads the output of all four through the dashboard's metrics. The owner does not have time to attend the weekly audit or read the daily AI report; the dashboard summarizes those workflows' outputs into the five headline numbers that fit in a Friday 5 p.m. email. Without the dashboard, the owner asks questions the service manager already answered in the underlying workflows; with the dashboard, the owner trusts the workflows and the service manager protects their bandwidth for the next week's execution.

The dashboard is also the first ROI artifact the L3 service manager produces. Every owner conversation about budget for AI tooling (Avoca subscription, Hatch lead nurture, CallRail Conversation Intelligence add-on, Rilla expansion) starts with "show me the lift the current stack is producing." The dashboard answers that question directly with attributed revenue lines. Without the dashboard, every budget conversation devolves into vibes — "I think Avoca is working" versus "I'm not sure it's worth $3K/month." With the dashboard, the conversation is "Avoca produced $42K of recovered revenue last week, here's the cost-per-booked-call from Avoca-handled calls versus answering-service baseline, here's the trend over 12 weeks." The dashboard is the language for talking about AI ROI to the owner.

The Five Headline Numbers and What They Measure

The L3 Ch2 dashboard's top line is five numbers. Each one measures a distinct workflow surface; together they cover the architecture-plus-recovery-plus-show-rate stack. No number duplicates another; no number is decoration.

Booking %

The percentage of inbound calls that resulted in a booked appointment. Numerator: booked appointments this week. Denominator: total inbound calls this week (including AI-handled, human-handled, after-hours, in-hours). Source: FSM appointment count plus AI receptionist's call-handled count, joined on call-ID to avoid double-counting. 2026 baseline 65% at undertrained CSR floors; target 80-85% on human-handled calls plus AI receptionist's 80-87% on AI-handled calls; combined effective floor 75% Tier-One plus recovery 85-88%. The booking-% dashboard line is the L2 Ch2 plus L3 Ch2 stack's top-of-funnel metric and the first line an owner reads.

After-Hours Capture Rate

The percentage of after-hours inbound calls (defined as inbound during the configured after-hours window, typically 6 p.m. - 7 a.m. weekdays plus all-day weekends) that produced a booked appointment. Numerator: after-hours booked appointments this week. Denominator: total after-hours inbound calls this week. Source: AI receptionist's post-call log filtered for after-hours timestamp plus FSM appointment-template completed bookings. 2026 baseline 0-15% at shops with answering services; target 80%+ at architecture-mature shops. The after-hours capture line is the workflow surface most owners under-track because pre-architecture there was no clean way to measure it; the dashboard makes it visible weekly.

Cost Per Booked Call

The total cost (marketing spend plus AI receptionist subscription plus answering-service residual if any) divided by booked appointments. Numerator: weekly marketing spend (GLSA, PPC, NiceJob, Hatch nurture, direct mail) plus AI tooling subscription cost (Avoca, Jobber AI Receptionist, HCP AI Agents, ServiceTitan Voice) divided to weekly. Denominator: booked appointments this week from all sources. Source: marketing-spend roll-up from CallRail or marketing-platform attribution plus accounting-system AI subscription pull plus FSM appointment count. Pre-architecture baseline at residential trades shops typically $280-$420; HL Bowman case study with Avoca drops it from $350 to $215 (39% reduction). Cost-per-booked-call is the L3 Ch2 dashboard's primary financial efficiency metric and the line the marketing manager and ops manager both look at.

Abandoned-Call %

The percentage of inbound calls where the homeowner hung up before a booking was attempted or completed. Numerator: abandoned calls this week (homeowner-initiated hang-ups during hold or before transfer completes). Denominator: total inbound calls this week. Source: phone-system or AI receptionist's call-completion log with abandonment flag. 2026 baseline 8-15% at shops with overflow during in-hours staffing gaps; target under 4% at architecture-mature shops. The abandoned-call line surfaces the in-hours overflow problem that Lessons 1 and 2 address through warm-transfer rules and recovery loop callback windows.

AI-Recovered Revenue

The dollar value of recovered bookings produced by the Lesson 2 recovery loop, measured at completed-job revenue. Numerator: recovered bookings × completion rate × average ticket. Calculation: sum across all kill-reason categories of (recovered calls in category × average ticket for that category × completion rate for recovered calls). Source: FSM completed-job revenue filtered to recovered bookings (flagged in the FSM appointment record by the recovery loop's workflow tags). Target $11K-$16K per week at a 7-truck shop, $565K-$830K annualized. AI-recovered revenue is the line that converts the recovery loop's effort into the owner's preferred metric (P&L revenue) and the line the PE partner reads as the AI-ROI proof point.

The Three Diagnostic Supplements

The five headline numbers tell the story; three diagnostic supplements explain when the story breaks. Each supplement is a single number that surfaces a specific workflow surface that the owner doesn't need to see weekly but the service manager monitors continuously.

CSR Show Rate

The percentage of booked appointments where the truck ran and the homeowner was there. Numerator: completed appointments. Denominator: booked appointments scheduled to run this week. The Lesson 3 workflow's primary metric, supplemented to the dashboard as a diagnostic rather than headline because booking-% and AI-recovered-revenue already incorporate show-rate effects. Target 92%+ at architecture-mature shops; sub-90% triggers Lesson 3 review.

Morning Confirmation Completion

The percentage of overnight-booked appointments where the morning CSR completed the confirmation call or text before the truck rolled. Numerator: morning-confirmed bookings. Denominator: total overnight bookings from the AI receptionist. Target 95%+; sub-90% triggers Lesson 1's architecture review (morning CSR overnight-review discipline failing) and Lesson 3's after-hours-confusion no-show pattern climbs.

Recovery Rate by Kill-Reason Roll-Up

The single number that aggregates recovery rate across the 12 kill-reason categories from Lesson 2, weighted by category volume. Numerator: total recovered calls this week. Denominator: total lost calls this week. Target 60-80% at recovery-loop-mature shops. Sub-50% triggers Lesson 2's quarterly recovery review and the per-kill-reason distribution audit.

The five headline numbers plus three diagnostic supplements fit on one dashboard page. No more, no fewer. Adding a ninth number dilutes attention; removing one breaks workflow surface coverage. The discipline is non-negotiable.

The Data Pipeline and the Attribution Rules

The dashboard reads from five data sources daily, joins on call-ID and appointment-ID, and produces the weekly aggregate at Friday 4 p.m. for the service manager's review before the 5 p.m. owner email. The pipeline is mature in 2026 across the named vendor stack but requires the service manager's quarterly verification to maintain attribution accuracy.

Data Source One: The AI Receptionist

Avoca, Jobber AI Receptionist, HCP AI Agents, or ServiceTitan Voice produces the post-call audit log with call-ID, timestamp, AI-handled/warm-transferred/emergency-paged classification, kill-reason tag for non-booked calls, FSM appointment-ID for booked calls. Nightly pull. Quarterly audit: all 11 FSM handoff fields populate correctly.

Data Source Two: The FSM Platform

ServiceTitan, Sera, HCP, FieldEdge, or BuildOps produces the appointment record with status, completed-job revenue, marketing-source attribution, recovery-loop workflow tag. Nightly pull. Quarterly audit: completed-job revenue rolls up to AI-recovered-revenue line without double-counting.

Data Source Three: CallRail Conversation Intelligence

CallRail produces inbound call-ID, marketing-source attribution, transcript link, sentiment-classifier output, missed-opportunity flag. Nightly pull. Quarterly audit: marketing-source-to-appointment mapping holds (no orphaned calls without source, no orphaned appointments without inbound-call reference).

Data Source Four: The Marketing Spend System

GLSA, Google Ads, NiceJob, Podium AI Employee, Birdeye AI Employee, Hatch nurture, direct-mail, sponsorship lines collectively produce weekly spend by channel. Weekly Monday pull for prior week. Quarterly audit: spend total reconciles with accounting and channel attribution maps cleanly.

Data Source Five: The Accounting System

QuickBooks, Xero, or FSM-integrated accounting produces the AI tooling subscription cost (Avoca, Jobber Plus, HCP plan, ServiceTitan Voice add-on, Rilla, Hatch, CallRail) and answering-service residual if any. Monthly pull, weekly-average roll. Quarterly audit: subscriptions align with vendor invoices.

The joined view runs nightly via a managed query (typically built in the FSM's reporting layer or a separate BI tool like Looker, Metabase, or Sigma). The service manager opens the joined view Friday at 4 p.m., spot-checks the weekly aggregate against intuition (does the booking-% match the floor's experience this week, does the AI-recovered revenue match the recovery loop's daily report aggregates), and finalizes the dashboard for the 5 p.m. owner email.

The Friday 5 p.m. Owner Email and the Monday Morning Cadence

The dashboard ships every Friday at 5 p.m. as a one-page owner email. Subject line: "L3 Ch2 dashboard — week of [date]." Body: the five headline numbers as a single block at top, the three diagnostic supplements as a second block, a 4-6 sentence commentary paragraph from the service manager, and a one-line forward priority for next week. Total reading time: 4 minutes. The owner reads on Friday evening or Monday morning before standup. No attachments, no spreadsheet, no PowerPoint — the discipline of the format is what makes it stick.

The Five-Number Block

Each number presented as: name, this-week value, trailing-4-week value, delta. Example for a healthy mature shop: "Booking % — 86% (4-wk avg 85%, +1pt). After-hours capture — 78% (4-wk avg 76%, +2pt). Cost per booked call — $228 (4-wk avg $235, -$7). Abandoned-call % — 3.4% (4-wk avg 3.8%, -0.4pt). AI-recovered revenue — $13,400 (4-wk avg $12,800, +$600)." The delta column tells the owner which direction the workflow is moving without requiring them to compute it.

The Three-Diagnostic Block

Same format: name, this-week, 4-week avg, delta. Example: "CSR show rate — 92.4% (4-wk avg 91.8%, +0.6pt). Morning confirmation — 96% (4-wk avg 95%, +1pt). Recovery rate roll-up — 68% (4-wk avg 65%, +3pt)." The diagnostics confirm the headline numbers are not hiding workflow breaks. If a headline number improves while a diagnostic regresses, the service manager's commentary explains the underlying cause.

The Commentary Paragraph

4-6 sentences from the service manager. Names the week's wins, the gaps, and the next-week priority. Example: "Booking % held above 85% for the fourth straight week; after-hours capture climbed 2 points as the Avoca system prompt's Variant B reframe deployed for price-shopper kill-reason last Monday. CSR show rate compounded another half-point as the Lesson 3 Monday huddle focused on no-window-commitment. AI-recovered revenue $13,400 this week, $90 above run rate. Next week's focus shifts to the warm-transfer SLA — pickup time drifted to 92 seconds vs. 90-second target and Tier Two booking-% is showing first signs of regression. Service manager reruns warm-transfer audit at Friday's cadence." Commentary is the trust-building artifact across quarterly cycles.

The Monday Cadence

The owner reads the dashboard, decides whether to follow up with any clarification, and the Monday morning operations standup opens with the service manager confirming the next-week priority is being executed. The dashboard does not replace the standup; it replaces the owner's need to ask "how are things going" because the answer is already in the dashboard. The standup becomes a forward-looking 5-minute alignment rather than a backward-looking 20-minute status check.

The Quarterly Attribution Audit and the ROI Defense

The weekly dashboard is the operational artifact; the quarterly attribution audit is the defensive artifact. The audit answers the question "is the L3 Ch2 stack actually producing the lift the dashboard claims" against three challenges: (1) double-counting (is the same booking attributed to both architecture primary capture and recovery loop?), (2) baseline integrity (is the 84% show-rate baseline measured pre-deployment or assumed?), (3) counterfactual rigor (would these bookings have happened without the AI tooling, or are they attributable to the tooling specifically?). The quarterly audit takes 4-6 hours and produces a 6-page audit memo the owner can hand to a coach, peer group, PE partner, or franchisor quarterly review without further explanation.

The Double-Counting Audit

Pull 100 random bookings from the prior 90 days. Verify each booking is attributed to exactly one workflow (architecture primary or recovery loop, never both). Calculate the double-count error rate. Healthy: under 2%. Above 5% means the FSM workflow tags are mis-applied and the AI-recovered-revenue line is inflated. Fix: tighten the workflow-tag application logic in the FSM appointment-template.

The Baseline Integrity Audit

Compare the pre-deployment baseline metrics (booking %, show rate, after-hours capture, abandoned-call %, cost per booked call) to the trailing-4-week post-deployment metrics. Validate the baseline was measured (not assumed) — pull the pre-deployment FSM and CallRail records, recompute the baseline, and confirm the dashboard's claimed baseline matches. Document any baseline corrections in the audit memo. Owners and PE partners specifically ask about baseline rigor in quarterly reviews; the audit memo must answer.

The Counterfactual Rigor Audit

For the AI-recovered-revenue line, sample 30 recovered calls from the prior quarter. For each, validate the counterfactual: would this booking have happened without the recovery loop? Calls where the homeowner explicitly said "I was going to call back myself" count as low-counterfactual. Calls where the homeowner said "I was going to call [competitor]" count as high-counterfactual. Healthy distribution: 70%+ high-counterfactual. Below 50% means the recovery loop is taking credit for bookings that would have happened anyway. Fix: tighten attribution criteria to count only calls with clear competitor-alternative or permanent-loss-without-intervention signals.

The quarterly audit's 6-page memo appends to the L3 service manager's quarterly architecture memo from Lesson 1. Combined documentation — architecture, recovery, show-rate coaching, measurement — survives any owner, coach, peer group, or PE partner challenge.

The Failure Modes the Dashboard Defends Against

Three failure modes kill the dashboard's value to the owner. First, the "no Friday email" failure: the service manager skips the Friday 5 p.m. cadence when busy, the owner stops receiving the dashboard, and the workflow loses its visibility. Within 4-6 weeks the owner concludes "I don't know if the AI is working" and starts asking questions in standups that the dashboard would have answered. Fix: the Friday cadence is non-negotiable; if the service manager cannot send, the ops manager or assistant sends the auto-generated version with a brief commentary placeholder. Second, the "metric proliferation" failure: the service manager adds a ninth, tenth, twelfth number to the dashboard because each one seems important. Owner attention dilutes, the 4-minute read becomes 12 minutes, and the owner stops engaging with the dashboard. Fix: discipline the five-plus-three structure; new metrics go to the quarterly attribution memo, not the weekly dashboard. Third, the "no commentary" failure: the dashboard ships as numbers-only without the 4-6 sentence service manager commentary. Owner reads the numbers and infers their own interpretation — usually less favorable than the actual workflow state — and trust erodes. Fix: every Friday email includes the commentary paragraph, even if it is the standard "all metrics on target, next-week focus is X."

The dashboard also defends against an upstream failure mode the service manager has to watch: workflow drift detected in the dashboard before it's detected in the underlying workflow logs. A 3-point drop in CSR show rate one week followed by a 4-point drop the next is the dashboard signaling a Lesson 3 cadence break before the service manager notices in the Friday show-rate pull. The dashboard's diagnostic supplements catch this early because the diagnostic numbers move 1-2 weeks before the headline numbers; the service manager's responsibility is acting on the diagnostic signal in the same week it surfaces, not waiting for the headline number to confirm.

Key Takeaways

• The dashboard is the fourth leg of the L3 Ch2 stack. Architecture (Lesson 1), recovery loop (Lesson 2), show-rate coaching (Lesson 3) produce the $1.2M-$2.4M annualized lift at a 7-truck shop. The dashboard (Lesson 4) is what makes that lift visible, defensible, and owner-trustworthy.
• Five headline numbers cover the L3 Ch2 stack: booking % (top-of-funnel), after-hours capture rate (architecture surface), cost per booked call (financial efficiency), abandoned-call % (in-hours overflow), AI-recovered revenue (recovery loop dollar attribution).
• Three diagnostic supplements explain when the headline numbers break: CSR show rate, morning confirmation completion, recovery rate by kill-reason roll-up. They move 1-2 weeks before the headlines and surface workflow drift early.
• Five data sources feed the dashboard: AI receptionist (Avoca/Jobber/HCP/ServiceTitan Voice), FSM (ServiceTitan/Sera/HCP/FieldEdge/BuildOps), CallRail Conversation Intelligence, marketing spend system, accounting system. Nightly join on call-ID and appointment-ID; quarterly attribution audit.
• The Friday 5 p.m. owner email is the discipline. One page, five-number block, three-diagnostic block, 4-6 sentence commentary, one-line next-week priority. 4-minute read. No attachments. The format is the message.
• The quarterly attribution audit is the defensive artifact. Double-counting audit (target under 2% error), baseline integrity audit (validate against pre-deployment FSM/CallRail records), counterfactual rigor audit (target 70%+ high-counterfactual on AI-recovered-revenue sample). 6-page memo appends to the L3 architecture memo.
• Three failure modes to defend against: no Friday email (owner loses visibility, asks standup questions the dashboard would have answered), metric proliferation (5+3 structure dilutes to 12+ numbers, owner disengages), no commentary (numbers without interpretation erode trust).
• The dashboard converts AI tooling from cost line to ROI line in every budget conversation. Avoca subscription discussion becomes "Avoca produced $42K of recovered revenue last week, cost-per-booked-call $215 vs. $350 answering-service baseline." Hatch budget becomes "Hatch reactivated $18K of stale leads, ROI 6.2x on subscription." The dashboard is the language for AI ROI discussions with owners and PE partners.