RevOps Playbook: Lead Routing, MQL/SQL Definitions, and Attribution Across HubSpot and Salesforce

Routing is where your pipeline either accelerates or stalls. Definitions are where trust is made or lost. Attribution is where budgets survive the next planning cycle. This playbook brings the three together for HubSpot and Salesforce: precise MQL/SQL semantics that sales accepts, routing logic that hits SLAs without collateral damage, and attribution methods that both CRMs can reproduce.

The alignment contract

Agree on a contract that binds definitions, routing SLAs, and attribution rules into a single, reviewable document. It prevents the common finger‑pointing loop where marketing claims MQL volume is strong, sales argues the leads are weak, and finance cannot reconcile influence with bookings. The contract states definitions, who is on the hook for response times, and how touches roll into opportunity influence.

Crisp, testable MQL and SQL definitions

Define MQL and SQL as milestones with input signals and acceptance criteria. MQL typically arises from behavior (intent) and fit (ICP). SQL is a sales acceptance event with documented next steps. Do not use fuzzy criteria. Encode the criteria in one system and mirror the resulting stage and timestamp in the other to avoid oscillation. Keep a change log so you can analyze performance across definition versions.

Routing as a product with SLAs

Design routing as a repeatable product. A request (new lead or qualified contact) triggers an evaluation against territory rules; an assignment event happens; and an SLA clock starts. Track time‑to‑owner at p50, p90, and p99 and publish it. Provide escalation routes when queues exceed thresholds and offer temporary overrides for campaigns without blowing up the rule set.

A pragmatic territory model

Territory logic should be deterministic and explainable: geography, industry, company size, named accounts, or a combination. Version this logic like code, with dates and reason codes. When rules change, capture a snapshot of the old logic, test on historical data, and deploy with a rollback path. Users should be able to answer “why did this record route to me?” without opening a black box.

Attribution that both CRMs can reproduce

Attribution fails when models are opaque or when the two CRMs do not agree on which interactions count. Use campaign membership and association labels to create a consistent catalog of touches. Decide whether you will use first‑touch, last‑touch, W‑shape, or data‑driven models; document each, including windowing rules and eligible events. If analytics lives in a warehouse, compute centrally and push results downstream. If not, reconcile reports periodically and explain differences in a shared appendix.

Guardrails to avoid brittle automations

Favor simple rules with strong observability over elaborate logic that only a few admins understand. Empirically test any new routing rule on a held‑out window. Shadow routing can reveal conflicts before go‑live. Keep cross‑object dependencies minimal and avoid having MQL or SQL depend on easily gamed fields that drift.

QA scenarios that matter

Hot lead from a high‑intent page converts to MQL with correct timestamp and routes to the right SDR within the SLA.
A re‑engaged customer contact is recognized as existing, moves to SQL properly, and does not create duplicate ownership.
A multi‑touch campaign shows consistent influence in both CRMs on the same opportunity without double counting.
A territory change does not retroactively break prior routing or attribution.

Rollout and measurement

Roll out routing and definition changes in phases: start with a region or a team and keep a tight feedback loop with sales leaders for two weeks. Publish a dashboard that tracks MQL→SQL conversion, time‑to‑owner, and influenced pipeline by channel. Hold a 30‑day review to decide which rules to keep and which to retire.

Publish the final definitions with examples and edge cases.
Train SDRs and AEs on escalation paths and ownership expectations.
Share a reporting cookbook so marketers and sales ops generate consistent views.
Document an exception process for named accounts and VIPs.

FAQ

Where should we compute MQL—HubSpot or Salesforce?

Compute in one place, usually HubSpot if marketing owns scoring and intent, and mirror the result with timestamps to Salesforce. The key is single authority with a mirrored read.

What SLA should we target for hot leads?

Sub‑15 minutes median to owner assignment is a good baseline for net‑new hot leads. Measure p90 and p99 and instrument alerts for breaches.

How do we stop routing flapping during territory changes?

Version territory logic with effective dates. Simulate changes on recent data before shipping. Freeze non‑essential schema updates during cutover.

Which attribution model should we start with?

Begin with first‑touch and last‑touch for clarity. Add W‑shape or data‑driven models only when stakeholders understand the trade‑offs and you can reproduce results in both CRMs.

Can we exclude low‑quality sources from MQL entirely?

Yes. Define exclusion lists and document them. Periodically review to ensure you do not over‑filter channels that later improve.