CRM
Lead Scoring in a Small CRM: A Simple Model That Actually Works
Lead scoring helps you decide who to call first, and a small CRM only needs a handful of visible signals and simple points to do it well.
You have forty names in your CRM and about an hour before you need to be somewhere else. Some of those people asked you a real question last week. Some clicked a link once in the spring and never thought about you again. If you message them in alphabetical order, the warmest person might wait three days while you chat with someone who forgot your name.
That is the whole problem lead scoring solves. It is not a fancy analytics project. It is a small habit of writing down what you already notice, so you can see who to talk to first without rereading every conversation.
This guide walks through a simple model you can build in an afternoon and keep up in a few minutes a week.
What lead scoring is for
A lead score is a number that stands for one question: how likely is this person to want a real conversation with me soon? That is it. It does not measure whether someone is a good person, a good customer, or worth your kindness. Everyone gets treated well. The score just decides the order.
For a solo seller or a small team, that order matters more than it seems. Research published in Harvard Business Review on online sales leads found that companies responding quickly were far more likely to have a meaningful conversation than those who waited even a few hours. Interest fades. A score helps you catch it while it is still there.
The broader idea is well established. As the Wikipedia overview of lead scoring explains, the practice combines information about who a person is with what they have done. Big companies build that with dozens of data points. You need far fewer.
Pick signals you can actually see
The most common mistake is scoring things you cannot reliably track. If your model depends on knowing someone's job title, household size, or how many pages they scrolled, it will break the first week. Choose signals that show up in your CRM, your inbox, or your messages without extra detective work.
Think in two groups.
Interest signals are things a person did:
- Replied to a message
- Filled out a form or asked for more information
- Booked a call or a demo
- Asked about price, how to order, or how to get started
- Attended a webinar or a live event
- Clicked a link you sent
Fit signals are things that make a good match more likely:
- Was referred by a happy customer or teammate
- Already uses a similar product or has told you about the need
- Lives somewhere you can actually ship to or serve
That last one works better as a gate than as points. If you cannot serve someone, they do not belong in your active list at all, no matter how interested they are.
Notice what is missing: email opens. Many email apps now load messages automatically for privacy reasons, so an open often means nothing. If you count them, a lot of people will look warm when they never read a word.
If your CRM already has tags for things like "referral" or "asked about price," you are halfway there. The guide on CRM tags vs custom fields covers how to keep those tidy so your score can rely on them.
Assign simple point values
Now give each signal a number. Keep the math boring. Round numbers, a small range, and values you can remember without a cheat sheet.
Here is a starting model that works for most affiliate marketers, direct sellers and small sales teams:
- Asked about price, ordering, or getting started: 20 points
- Booked a call or demo: 15 points
- Filled out a form asking for information: 10 points
- Replied to a message: 10 points
- Attended a webinar or live event: 10 points
- Referred by a customer or teammate: 10 points
- Has told you about a clear need: 5 points
- Clicked a link you sent: 3 points
The strongest signals are the ones where a person tells you, in their own words, that they are thinking about buying. A question about price beats ten link clicks. Weight your model the same way.
Scores should also cool down, because interest does. HubSpot's guide to lead scoring describes using negative attributes alongside positive ones, and that idea scales down nicely:
- No reply after three friendly touches: subtract 10 points
- Every 30 days with no activity: subtract 5 points
- Said "not right now": reset to zero and set a reminder for when they suggested
That last rule is about respect as much as efficiency. When someone tells you the timing is wrong, believe them. Put a date on your calendar, check in kindly when it arrives, and let them start fresh.
A quick word for direct sellers. A person asking "how much can I make?" is showing real interest and deserves the 20 points. They also deserve an honest answer that does not promise any income. The FTC's business guidance on multi-level marketing is clear that earnings claims need to reflect what typical participants actually experience. A high score is a reason to have a careful, truthful conversation, not a reason to push.
Set a threshold for follow-up
Points only help if they trigger an action. Pick two cutoffs and attach a clear habit to each band.
- 30 points or more: personal follow-up within one business day. A real message from you, written for them. Reference what they asked. Offer a time to talk.
- 15 to 29 points: steady nurture. Check in weekly or every other week with something useful, like an answer to a common question or a short customer tip.
- Under 15 points: light touch. Newsletter, occasional content, and no pressure.
With this model, someone who replied to you and then asked about price lands at 30 and goes straight to the top of today's list. Someone who clicked a link and attended a webinar sits at 13 and gets your helpful content until they show more interest.
The threshold should match your capacity. If 30 points gives you more people than you can personally reach in a day, raise it to 35. If it gives you almost nobody, lower it. The goal is a short daily list you can finish, not a long one that makes you feel behind.
These bands also map well onto pipeline stages. If you have not set those up yet, the guide on CRM pipeline stages for a small team pairs nicely with this, and a written follow-up system tells you exactly what to send once someone crosses the line.
Test the score against real outcomes
Any scoring model is a guess until you check it. The good news is that checking takes about twenty minutes once a month.
Pull up everyone who reached a real outcome in the last month. That might be a first order, a booked call that actually happened, or a new team member who enrolled. Write down the score each person had right before that moment. Then look at two groups.
High scorers who went nowhere. If lots of people hit 30 and then vanished, one of your signals is probably overweighted. Webinar attendance is a frequent culprit. Plenty of people watch out of curiosity.
Low scorers who said yes. If people with 10 points keep buying, you are missing a signal. Often it is something you see but never recorded, like a person who comments on your posts every week or a customer who reorders without being asked.
Change one thing at a time. If you adjust five values at once, you will never know which change helped. Make one tweak, run it for a month, and look again.
Also check the boring stuff. A score is only as good as the data behind it. Duplicate contacts, missing tags and old leads that should have been archived all throw the numbers off. A quick CRM cleanup before your first review makes the results far easier to trust.
When to leave scoring alone
Lead scoring is a tool, not a requirement. There are times it does more harm than good.
When your list is small. If you have twenty active conversations, you probably know each person well enough to prioritize by memory. A score adds busywork without adding insight. Come back to it when the list outgrows your head.
When you keep fiddling. If you change point values every few days, the score never settles long enough to mean anything. Pick a model, commit for a month, and review on a schedule.
When the number starts replacing the person. This is the one to watch. A score of 8 does not mean someone is unimportant. It means they are not ready yet. Keep replying to every message, warmly and promptly, regardless of the number. Scoring decides who you reach out to first. It never decides who deserves a reply.
When your offer changes a lot. A new product, a new audience, or a big change in how people find you can make old signals stale. Treat it as a fresh start and rebuild from your next month of results.
Common questions
How many signals should a small lead scoring model start with?
Start with four to six. That is enough to separate warm people from cool ones, and few enough that you can explain every score from memory. Add a new signal only when testing shows you are missing something.
Should email opens count toward a lead score?
Give them very little weight or skip them. Privacy features in many email apps load messages automatically, so an open often does not mean a person read anything. Replies, clicks, bookings and direct questions are far more reliable.
How often should I change my point values?
Review once a month, and change one value at a time. If you adjust weekly, you never collect enough results to know whether a change helped, and the score stops meaning anything.
The bottom line
A good lead scoring model for a small business fits on an index card. Pick a handful of signals you can actually see, give the strongest ones the most points, let scores cool over time, and attach a clear follow-up habit to each band. Then check it monthly against who really said yes, and adjust one thing at a time.
The point is not the number. The point is making sure the person who asked you a real question this morning hears back from you today, while everyone else still gets your attention at a pace that suits them. That small habit is what makes follow-up feel natural instead of frantic.
If you would like a place to keep tags, scores, pipeline stages and reminders together, Prala is one option built for sellers like you, and you can see the plans on the pricing page. Whatever tool you use, start simple, stay kind, and let the results tell you what to change.
