Wrench Group's approach to managing 100+ home services brands offers a blueprint for how PE-backed platforms can unify backend operations without destroying local brand autonomy.
The Wrench Group model: 100+ brands, one operational backbone
Wrench Group operates one of the largest home services platforms in the United States. Their portfolio includes national franchise brands, regional independents, and local specialists. Each brand has its own identity, customer base, and market positioning. But behind the brands, the platform needs operational consistency to scale.
The challenge is familiar to any PE-backed platform: how do you maintain brand differentiation while creating operational leverage? Wrench Group's answer is architectural separation. They do not force every brand onto the same CRM or dispatch tool. They force every brand to produce the same operational data in the same format. The tools can vary. The data cannot.
This approach is different from the typical roll-up strategy, which usually migrates every acquired company onto the platform's chosen tools within 90 days. Wrench Group understood early that brand value lives in customer trust, and customer trust lives in local execution. Disrupt the local execution and you damage the brand you just paid to acquire. That is why their architecture has lasted while other platforms cycle through integration projects every two years.
The three-layer architecture
Wrench Group's architecture separates concerns into three layers. Each layer has different ownership, different constraints, and different success criteria.
Layer 1 — Brand-facing customer experience
This layer includes everything the customer sees and interacts with: the brand website, the call center script, the scheduling experience, the technician uniform, and the follow-up communication. Local brands own this layer. They know their market, their customers, and their competitive positioning.
The platform's role in Layer 1 is governance, not control. Set brand standards. Enforce quality minimums. Provide shared services like call center overflow or marketing support. But let the brand manage the customer relationship in the way that works for their market.
Some platforms get this wrong by centralizing customer experience too aggressively. They create a single website, a single phone number, and a single scheduling system. It feels efficient. But customers who trusted 'Joe's HVAC' now find themselves talking to a generic call center. The brand value erodes. Wrench Group avoids this by keeping brand identity intact while standardizing the operational handoffs behind the scenes.
Layer 2 — Operational workflow engine
This layer includes dispatch, job management, technician tools, inventory tracking, and field operations. It is the engine that turns a customer request into a completed job. The platform owns the architecture of this layer, but brands can have flexibility in implementation.
Wrench Group's insight was to standardize the data that flows through this layer, not the tools that produce it. A job must have a standard set of fields: customer ID, service line, status, technician, start time, end time, materials, and outcome. But the tool that captures those fields can be ServiceTitan at one brand, Jobber at another, or a custom system at a third — as long as the data exports to the platform standard.
This approach requires a strong integration layer. But it preserves local operational knowledge. A dispatcher who has spent ten years on one tool does not need to relearn their job because the platform bought their company. And the platform still gets the operational data it needs for consolidation. The dispatcher keeps their workflow. The platform gets their data. Both win.
Layer 3 — Consolidated reporting and finance
This layer is the platform's primary value add. It includes consolidated dashboards, platform-level financial reporting, cross-brand analytics, and investor-facing metrics. The platform owns this layer completely. Brands consume reports from it but do not build their own parallel reporting.
The key to Layer 3 is that it depends entirely on Layers 1 and 2 producing clean, standardized data. If Layer 2 allows every brand to define 'job complete' differently, Layer 3 cannot produce reliable technician productivity metrics. If Layer 1 allows every brand to collect customer data in different formats, Layer 3 cannot produce accurate customer lifetime value analysis.
Wrench Group invested heavily in data standards before building advanced analytics. They understood that a beautiful dashboard on dirty data is worse than a simple spreadsheet on clean data. Most platforms try to build Layer 3 before Layer 2 is stable. That is why their reporting fails. Wrench Group built the foundation first, then the house.
What Wrench Group got right (and what others copy wrong)
What Wrench Group got right was patience. They built the data layer first. They accepted operational complexity in exchange for brand preservation. And they treated integration as a long-term capability, not a 90-day project.
What others copy wrong is the surface-level architecture without the underlying discipline. They see that Wrench Group allows different tools and conclude that 'we do not need to standardize.' They miss that Wrench Group standardizes data ruthlessly. The flexibility is in tools, not in data definitions.
Another common mistake is trying to replicate Wrench Group's three-layer model without Wrench Group's scale. A 10-brand platform does not need the same integration infrastructure as a 100-brand platform. Start with data standards. Add integration layers as you grow. Do not over-architect before you have the volume to justify it. I have seen 5-brand platforms build enterprise integration platforms that sit unused because there is not enough data volume to make them valuable.
How to apply this architecture to smaller platforms
If you are running a 5-brand or 10-brand platform, you can apply the same principles with lighter tooling. Start by documenting your data standards: what fields must every brand produce, what do they mean, and what format should they be in? Then audit each brand against those standards.
Build lightweight integrations that pull data from each brand's existing tools into a shared reporting database. This does not need to be expensive. A scheduled export to a shared data warehouse, combined with simple standardization rules, is often enough for the first stage.
Resist the urge to force every brand onto the same tool immediately. The operational disruption usually exceeds the reporting benefit. Standardize data first. Consolidate tools later, if at all, when the business case is clear. The goal is platform visibility, not tool uniformity.
When this model is overkill
The three-layer model is overkill for platforms that are acquiring companies with the intent to fully absorb them into a single brand. If your strategy is to buy local HVAC companies and rebrand them all under one name, you do not need brand-facing autonomy. You need operational consolidation.
It is also overkill for platforms with fewer than five brands that all operate in the same market with the same tools. If everyone is already on ServiceTitan, the integration challenge is configuration, not architecture.
The model shines when you are managing a portfolio of distinct brands with distinct customer relationships, distinct local operations, and a need for platform-level visibility. That is exactly the situation most PE-backed home services platforms find themselves in. If that describes your platform, Wrench Group's approach is worth studying closely.
Building the integration layer incrementally
You do not need to build the full three-layer architecture on Day 1. Start with data standards for one metric — revenue, or job count, or technician utilization. Prove that you can collect clean data from every brand for that one metric. Then add metrics one at a time.
This incremental approach reduces risk and creates early wins. Leadership sees progress. Brands are not overwhelmed with compliance requirements. And the technical team can iterate on the integration layer without betting everything on a big-bang launch.
If the problem is recurring, treat it as a systems problem before adding more manual process around it.