Lead Scoring and Routing: Getting the Right Leads to the Right Reps
If a demo request comes in at 2 p.m. on a Tuesday, who gets it, and how long until they call?
In the companies I’ve worked with, lead routing was duct tape: rules added one at a time by different people, each one sensible when it went in. Then someone changed the scoring and forgot the routing. Leads sit in limbo and MQLs go untouched, and nobody notices until a rep says the leads are bad.
A better MQL definition won’t fix that, and neither will a smarter score. What decides whether a lead becomes pipeline is the hour after it qualifies: who gets it, how fast, and where it goes if it isn’t ready. So I’d fix routing and recycling before I spent a week tuning points. This guide covers scoring, routing and lifecycle stages in HubSpot and Salesforce, but if you only have time for two sections, read 4 and 5.
What is lead scoring? Lead scoring ranks contacts and accounts by fit (could they buy?) and intent (are they in-market now?) so reps work the best leads first. It's one part of lead lifecycle management, the system that defines each stage, routes qualified leads to the right rep, and recycles the ones that don't convert.
1. MQL is a concept until you build the workflow
Every team has an MQL definition. Far fewer have an MQL workflow. The gap between "we define an MQL as X" and "an MQL automatically gets scored, stamped, routed, and recycled if it stalls" is where pipeline leaks. The build matters more than the definition. Your CRM is not a strategy, but it is where the strategy either happens or doesn’t.
A working MQL workflow does five things the moment a contact qualifies:
- Updates the lifecycle stage to MQL in both HubSpot and Salesforce.
- Stamps the date in a custom property, so you can measure how long leads sit at each stage.
- Assigns an owner using your routing rules.
- Notifies the rep with context: why they qualified and what they did.
- Starts an SLA clock, and escalates or recycles if nobody acts in time.
Before any of that, sales has to agree to the definition. If reps don't trust the criteria, they won't work the leads, no matter how clean the workflow is.
Read the full breakdown: MQL Is a Concept, Not a Workflow →
2. Build your lead scoring model from closed-won, not gut feel
Work backward from your closed-won deals: who were those buyers, and what did they do before they bought? (The method, step by step: building a lead scoring model from closed-won data.)
A model that holds up scores two things separately:
| Dimension | What it measures | Examples |
|---|---|---|
| Fit | Could this be a customer? | Company size, industry, job title, tech stack |
| Intent | Are they in-market now? | Pricing page visits, demo requests, high-intent content, repeat sessions |
Three rules keep it working. Add decay, so engagement from six months ago stops counting. Add caps, so one person reading every blog post can't score their way to MQL. And don't count email opens as engagement, since privacy features and bots inflate them.
Read the full breakdown: a lead scoring model your SDRs will use →
3. Score the account too
In committee sales, one qualified contact tells you very little about whether the company is buying. Scoring at the account level (MQA) catches intent the lead model misses, and it works alongside the MQL.
Common MQA triggers:
- Several contacts engaged at the same company within 30 days.
- A combined account score crossing a threshold, built from visits, downloads, clicks and ad engagement across everyone at the account.
- A high-intent action plus a second contact from the same account visiting pricing in the same week.
Treat an MQL from an MQA account as your highest-priority lead. That combination is where your best pipeline comes from.
Read the full breakdown: Stop Qualifying Leads. Start Qualifying Accounts. →
4. Lead routing in minutes
Write the routing rules down in one place. If they only live in workflow settings, nobody can check them.
A qualified lead that sits in a queue goes cold fast. Whether you route in HubSpot or Salesforce, the logic is the same. Routing should be automatic, predictable, and fast enough that the rep reaches out while the buyer still remembers why they raised their hand.
- Match to existing owners first. If the account already has an owner, the lead goes to them. Reps hate finding out a colleague called their account.
- Then route by rule. Territory, segment, company size or product line, whatever matches how your sales team is organized.
- Round robin only as the fallback.
- Put a time on it. Demo requests deserve a response within the hour. Other MQLs should be worked within an agreed SLA, with alerts when it's missed.
Map every way a lead can enter, including every form, every ad platform, every integration and every list import. The entry points nobody mapped are the ones that leak.
5. Recycle what doesn't convert
An MQL that goes nowhere shouldn’t sit in MQL forever, out of nurture and still getting sales emails it stopped answering in April.
Keep two cases apart. If nobody worked it inside the SLA, that’s a follow-up miss, so escalate it to the manager. If a rep worked it and it isn’t ready, recycle it with a reason, take it out of sales sequences, put it in nurture and let it re-score. The HubSpot setup, including how to move a lifecycle stage backward, is in the MQL post.
Recycle by reason. Bad timing, the wrong person, a bad fit and a dropped lead each need a different path.
Read the full breakdown: What to Do With MQLs That Never Convert →
6. Know where the gaps are before you build
You can’t fix routing you haven’t mapped. Before rebuilding lifecycle and scoring, audit the current state: every entry point, the path each lead takes, and the dead ends where contacts get lost. That diagnosis tells you where to start.
What I find most often: forms that don't trigger any workflow, contacts with no lifecycle stage, MQLs with no owner, scores inflated by old or low-value activity, and a sync between HubSpot and Salesforce that drops the stage on the way across.
Read the full breakdown: What a Demand Gen Audit Looks Like →
A ten-minute check
Answer these before you rebuild anything:
- Can sales and marketing both say the MQL definition in one sentence, and is it the same sentence?
- Submit your own demo form. Who gets it, and how long until someone reaches out?
- How many MQLs from last quarter have no owner, or an owner who left?
- Does a score ever go down?
- What happens to an MQL that a rep works and it isn’t ready? Write down the actual steps.
Whichever answer took longest to find is where to start.
Where to start
See where your leads are falling through.
The free B2B Attribution Audit includes a lead flow and routing check. It takes about five minutes, and you finish knowing the three things to fix first, plus a 30/60/90 plan.
Run the free audit →Want someone to map it with you?
The Demand Gen Health Check
Lifecycle and routing problems depend on how your team works day to day, which is rarely how the process doc says it works, which makes them hard to fix from a template. A Health Check maps every entry point, the scoring logic and the handoff, then hands you a ranked list of fixes. Two weeks, fixed price.
Book a 20-minute fit call →Common questions
What's the difference between an MQL and an SQL?
An MQL is a lead marketing has qualified based on fit and engagement. In the convention I use, an SAL (sales accepted lead) is an MQL a rep has accepted and started working, and an SQL is a lead that has had a meeting and qualified. Your team can draw the lines differently. What matters is that each stage is an event with a date, written down and agreed with sales. The handoff between MQL and SAL is where most lead flow breaks. See how to build the handoff.
How do I set up lead scoring in HubSpot?
Build fit and intent as separate scores, using property criteria for fit (title, company size, industry) and activity criteria for intent (pricing visits, demo requests). Set a threshold that triggers a workflow to set MQL. Check what your HubSpot tier includes first, since scoring features vary by plan. See a full scoring model.
How does lead routing work in Salesforce?
Salesforce routes leads with assignment rules, usually by territory, segment or owner, with round robin as a fallback. The common failure is a lead that should match an existing account and doesn't, so it goes to the wrong rep. Lead-to-account matching fixes that.
What lifecycle stages should I use in HubSpot?
HubSpot's defaults work for most teams: Subscriber, Lead, MQL, SQL, Opportunity, Customer. What matters more is that every stage has a written definition, an owner and an automated trigger, and that each one maps cleanly to a lead status or stage in Salesforce.
How do I build a B2B lead scoring model?
Start with your closed-won deals from the last year. Look at what those buyers had in common and what they did before buying. Score fit and intent separately, add decay and caps, and check the model against conversion rates after a few months.
What's the difference between MQA and MQL?
An MQL is one person showing enough interest. An MQA is a whole account showing enough collective engagement to suggest the company is in-market. They work together. Read the case for MQA.
Why does sales ignore marketing leads?
In my experience, because the leads arrived without context, too late, or by criteria sales never agreed to. Fix the definition with sales, route fast, and give reps the reason a lead qualified. Then ask for specific feedback instead of general complaints.
How fast should leads be routed?
Routing itself should be instant and automatic. Demo and contact requests deserve a human response within the hour. For other MQLs, agree on an SLA with sales and track it. A missed SLA is a follow-up problem to escalate; a worked lead that isn't ready is the one to recycle.