Where they started
Summit Plumbing Partners is a composite illustration of a PE-backed plumbing services roll-up operating five companies across Arizona, Texas, Colorado, Nevada, and Utah. At Day 0, the five entities — Lone Star Plumbing (DFW), Desert Flow Plumbing (Phoenix), Mile High Plumbing (Denver), Sunstate Mechanical (Las Vegas), and Wasatch Service Group (Salt Lake City) — had been acquired over 18 months with no financial integration plan, no shared chart of accounts, and reporting that arrived as emailed PDF snapshots from each company's bookkeeper.
The Day 0 crisis wasn't operational — it was compliance. During diligence on the sixth potential acquisition, the platform CFO discovered that two of the five existing portcos had let their backflow prevention device certification registrations lapse. Those certifications were required for their commercial plumbing contracts. The gap represented $4.2M in contract revenue at risk and a covenant breach risk on the SBA 7(a) debt backing two of the acquisitions.
The integration project launched under pressure. The CFO needed a compliance dashboard before the next LP call, a financial integration before the refinancing conversation with their SBA lender, and a unified reporting cadence before adding a sixth acquisition to a platform that still couldn't produce a consolidated P&L.
The 100-day plan
Summit followed the RollForge 100-Day Integration Playbook with one modification: they front-loaded the compliance work before the data integrations, because a covenant breach was a worse outcome than a delayed reporting timeline.
Compliance Triage — fix the backflow gap first
The two lapsed portcos filed renewal applications within 72 hours of discovery — one same-day, one requiring a technician recertification that took 11 days. RollForge anomaly alert configured specifically to track license/certification expiry dates across all five entities. Day 14 checkpoint: all five portcos have current backflow certifications and a 90-day advance-warning alert live in the system. The SBA lender received written confirmation before Day 14.
Data Connections — three systems, one view
Lone Star and Sunstate on QuickBooks Online — OAuth connected in under 10 minutes each. Desert Flow on Xero — connected via PKCE OAuth, nightly sync active Day 17. Mile High and Wasatch on QuickBooks Desktop — CSV export via the RollForge upload wizard, 25 minutes per entity. All five bank feeds live via Plaid by Day 35. Chart of accounts normalization began Day 30 once data was flowing — plumbing revenue split into residential service, commercial service, new construction, and emergency call-out. Labor separated into licensed plumbers, apprentices, and back office.
KPI Standardization — the hard conversations happen here
Six KPIs locked across all five portcos: gross margin (normalized COA), average revenue per service call, technician utilization rate, material markup percentage, AR days outstanding, and cash runway. The material markup conversation was the hardest — three of the five companies had been pricing materials at cost plus 10–15%, while industry standard is 30–40%. That gap alone explained 4 of the 7.2 EBITDA improvement points that materialized by Day 100. Anomaly alerts configured per company with plumbing-specific thresholds. Monday Pulse digest started Day 45.
Board-Ready + Lender Package
Banker-ready loan packages for all five entities. The refinancing conversation with the SBA lender that had been delayed by the compliance gap happened on Day 88 — the CFO generated a combined 5-entity package with normalized DSCR calculations, updated backflow certification status, add-back schedule (owner salaries, personal vehicles, one-time equipment purchases), AR aging, and AI executive narrative in 23 minutes. The lender extended terms without requiring a field visit.
What actually changed (before / after)
These are the eight operational metrics Summit tracked at Day 0 and Day 100:
The hard parts
The material markup repricing was the most operationally sensitive intervention. The three GMs running sub-market markups weren't making an accounting error — they'd priced that way deliberately to win bids in competitive residential markets. The CFO's approach was to model the margin impact for each GM separately: at 32% markup instead of 12%, what did that mean for their annual EBITDA? For Lone Star, it was an additional $680K annually on existing call volume without adding a single truck. That framing converted all three GMs inside a 90-minute working session.
AR days outstanding was a hidden cash flow problem nobody had quantified. Desert Flow was invoicing commercial accounts net-60 by default — a legacy of the prior owner's relationships with property management firms. Shifting to net-30 with a 2% early payment discount converted $1.1M of slow receivables to cash within 45 days. That change came directly from the AR aging report in RollForge — without live visibility, it would have taken another year to surface.
The QuickBooks Desktop migration for Mile High and Wasatch was slower than expected — not because of the export process, but because three years of historical data had been entered inconsistently and required line-by-line review on the COA mapping. The RollForge upload wizard flagged 47 accounts that didn't map cleanly to the target COA, each requiring a manual classification decision. That took one person two full days per entity. The CFO would do it the same way again — the alternative was guessing at the mapping, which would have produced a garbage consolidated P&L.
What Day 100 actually looked like
The refinancing conversation with the SBA lender was the clearest proof point. The lender had three concerns coming in: the compliance gap from Day 0, the variability in gross margins across the five companies, and the AR quality at Desert Flow. The CFO opened the RollForge loan package builder, selected all five entities, and generated the combined package in 23 minutes. The package showed: current certification status for all entities, normalized gross margin using the new platform COA (margin variances narrowed significantly after COA standardization), and Desert Flow's updated AR aging showing the net-30 conversion already underway.
The lender asked two follow-up questions — both answered live off the dashboard without pulling a single spreadsheet. The term extension was approved without a field visit, which the lender said was unusual for a portfolio with a prior compliance issue.
Day 100 Results
- 5 portcos unified in a single portfolio view — live P&L, no manual aggregation
- EBITDA margin improved 7.2 points — 11.4% to 18.6% blended across the platform
- $112M combined revenue with standardized KPIs — first time the GP had a real platform view
- Loan package generated in 23 minutes — SBA refinancing closed without a field visit
- AR days outstanding reduced from 52 to 31 days — $1.1M in freed cash at Desert Flow
- Compliance tracking live — 90-day advance certification alerts across all five portcos
- Sixth acquisition onboarded in 17 days using the playbook built during integration
What comes next
Summit's next 100 days are focused on expansion. They have three acquisition targets in the pipeline — all in markets adjacent to existing portcos. The playbook from Days 1–100 has become the acquisition onboarding template: compliance audit on Day 0, data connections in the first two weeks, COA migration by Day 30, anomaly alerts live by Day 45. The CFO estimates they can onboard a new acquisition into the platform in under 20 days, compared to 90-day timelines before the integration work.
The material markup repricing conversation is now part of their first 30-day operating plan for every acquisition — not an optional optimization, but a standard intervention. At five companies, the markup normalization alone accounted for 4 of 7.2 EBITDA improvement points. At 10 or 15, it's a predictable, repeatable value creation play.
The playbook is documented. The platform is ready.
If your roll-up has fragmented reporting, unmapped compliance exposure, or a reporting lag you've stopped counting — the path Summit followed is replicable.