Know who to call first: lead scoring you can explain
Thirty new leads and one hour to call them back. Here's how HiloVoice scores every phone lead from 0 to 100 against signals you choose, shows the reason behind every point, keeps disqualified callers out of the ranking, and puts your strongest opportunities at the top.
Operations9 min read

It's Monday morning and there are thirty new leads from the weekend. Some are ready to book today. Some are shopping around for next spring. One is a homeowner whose furnace died on Saturday night. Most teams call them back in the order they came in, which means the best opportunity of the weekend may wait until after lunch.
Lead scoring fixes the order. HiloVoice scores every lead from 0 to 100 against what you've said makes a caller valuable, and puts the strongest at the top. This article covers how it works, how you set it up, and why every score comes with its reasons.
The score: 0 to 100, with a plain label
After each call, the lead gets a whole number from 0 to 100. The number decides the order. A short label makes it easy to scan:
Hot
- Score
- 80–100
- What it usually means
- Strong fit, ready to act. Call first.
Strong
- Score
- 60–79
- What it usually means
- A good opportunity with a gap or two, like a later timeline.
Potential
- Score
- 40–59
- What it usually means
- Worth following up, but not today's priority.
Low priority
- Score
- 0–39
- What it usually means
- Few of your signals were there.
Your Leads page sorts highest score first, and newest first among leads with the same score. You can also sort the other way, by date, or filter to one band, like just the Hot leads.
Qualify first, then score
Scoring answers "how good is this opportunity?" It doesn't answer "can we take this job at all?" That's a different question, and HiloVoice keeps them apart.
Your screening rules decide who qualifies. A caller outside your service area, asking for work you don't do, is disqualified, and their lead shows that, with the reason: "Disqualified: outside service area." They don't show up as a lead with a score of 22, and they're never ranked near the top because they happened to sound urgent.
Scoring then ranks the callers who do qualify. The line between the two is worth drawing carefully. If you can't serve a town, that's screening. If you can serve it but would rather not drive there, that's a scoring signal.
You decide what makes a lead valuable
A score is only as good as the signals behind it, and those are different for every business. You set them in the Lead Capture step of your agent. Each signal has a name, a sentence saying what to look for, and an impact:
- Positive signals add to the score: low +5, medium +10, high +20, or very high +30.
- Negative signals take points away, on the same scale: −5 to −30.
You don't have to start from a blank page. HiloVoice drafts signals from what your agent already knows: your industry and services, what the agent is for, its intake questions, your screening rules, and whether it books appointments. Then you read them, change what doesn't fit, and save.
HVAC company
- Positive signals
- No heat or cooling now (+20), homeowner (+10), system over 12 years old, a replacement opportunity (+20), wants a visit this week (+15)
- Negative signals
- Just collecting quotes for later (−15)
Bankruptcy firm
- Positive signals
- The kind of case the firm takes (+20), a garnishment or foreclosure date coming up (+20), asked for a consultation (+15)
- Negative signals
- Only wants general information (−15)
Real estate agent
- Positive signals
- Pre-approved for a mortgage (+20), buying within 3 months (+20), budget in the agent's range (+10)
- Negative signals
- Timeline over a year away (−15)
Every point has a reason
The quickest way to lose trust in a score is to show a number with nothing behind it. "87, because this sounds like a good lead" isn't something a team can act on or argue with.
So HiloVoice splits the work. After the call, AI reads the call and your intake answers and decides, one signal at a time, whether the evidence is there, with a short note saying what the caller said. It doesn't pick a number. Then the score is added up from the points you set, the same way every time. The AI can't give a signal more points than you did, or invent one you didn't ask for.
Tap any score to see the breakdown. For a bankruptcy caller, it might read:
- The kind of case the firm takes, +20: "Looking for help with a Chapter 7 filing."
- Deadline coming up, +20: "Wage garnishment starts next week."
- Asked for a consultation, +15: "Wanted the earliest available appointment."
When someone on your team disagrees with a score, the breakdown shows exactly which signal to change.
How the score adds up
The score is the share of your positive signals a lead earned, turned into a number out of 100, minus any negative signals that applied. Rounded, and never below 0 or above 100.
One detail matters a lot: "we don't know" isn't the same as "no". Each signal comes back as met, not met, or unknown. If the caller never said whether they own the home, that signal is set aside rather than counted against them, so a strong lead isn't dragged down by a question nobody asked.
Here's an HVAC example. The business has six positive signals worth 90 points in total:
- The caller has no heat (+20 met), wants a visit this week (+15 met), is in the main service area (+10 met), and the system is 15 years old (+20 met). That's 65 points earned.
- They're renting, so "homeowner" isn't met (+10 not met). It still counts toward the total available.
- Nobody asked about budget, so "budget fits" (+15) is unknown and set aside. The total available drops from 90 to 75.
- 65 out of 75 is 87. No negative signals applied, so the lead scores 87: Hot.
Had they said they were only collecting quotes for next year, the −15 would have brought them down to 72: Strong.
No extra questions on the call
Scoring works from what the call already gathers. It doesn't add questions to the call to fill in a score. That's why the best signals are ones your intake already covers, like timeline, the service, urgency, and whether they asked to book. If a signal keeps coming back unknown, either add the question to your intake, or accept that it won't count for much.
Two habits keep a scoring model honest:
- Don't count the same thing twice. "Ready to buy", "strong intent", and "wants to proceed" are one signal wearing three names. Together they'd outweigh everything else. Use distinct signals instead: fit, urgency, readiness, value, and whether they're the one who decides.
- Only reward what matters to you. Booking an appointment on the call can be a strong signal, if your business books. If it doesn't, leave it out.
A good lead stays a good lead
A lead's score measures how strong the opportunity is, not how old it is. A 94 doesn't become a 60 because it sat over the weekend. That's deliberate: the age of a lead tells you how fast to move, and the score tells you what you're moving on. Keeping them apart means a great lead from Friday still stands out on Monday.
Scores you can rely on
- Never a false zero. A lead that hasn't been scored yet shows "Not scored", and one whose scoring failed shows "Score unavailable". Zero only ever means a lead was scored and none of your signals were met.
- Scoring never holds up a call. It runs after the call ends. The caller and the lead are never affected by it.
- Changes apply going forward. When you edit your signals, new calls are scored with the new version. Earlier scores stay as they were, so last month's numbers don't quietly change.
- Existing leads stay as they are. Leads from before you turned scoring on show "Not scored" rather than a guess.
Getting started
- Open the Lead Capture step of your agent and turn on lead scoring.
- Let HiloVoice draft your signals, then read them as if you were training a new hire. Would you call these leads first?
- Remove overlaps, and move anything that's really a yes-or-no gate into Caller Screening.
- Make a few test calls as different kinds of callers, then open each score and check the reasons make sense.
- After a couple of weeks, compare scores with the leads that actually booked or bought, and adjust.
Lead scoring works best with a good intake, since it scores what the agent collects. See how intake agents gather it, and why structured outcomes make it possible.



