Lead Scoring: How to Set Up a Model That Tells You Who to Call First

Lead scoring gives every contact a number that answers one question for your team: who should we talk to first? Done properly, it means the closer spends the morning on the five people most likely to buy instead of the fifty who filled in a form once. Done badly, it is a number nobody trusts, because it went up every time someone opened an email.

This guide shows how to build a lead scoring model that reflects how your customers actually buy, how to score positive and negative signals, how to make scores fade when leads go quiet, how to route leads on their score, and how to set it all up in Inflowave.

What lead scoring is

A lead score is a single number, usually 0 to 100, that rises when a contact does things buyers do and falls when they show they are not a fit or have gone cold. It combines two kinds of signal:

• Fit: who they are. Are they in the service area, the right type of business, a budget you can work with, from a source that tends to buy?
• Engagement: what they do. Did they reply, book a call, fill in the form, click the pricing link, pick up the phone?

Fit tells you whether they could buy. Engagement tells you whether they are about to. A good model uses both.

Step 1: Decide what "converted" means

Before you assign a single point, write down what a conversion is for this business: a won deal, a paid invoice, a booked and attended appointment. Everything else in the model exists to predict that event. If you skip this, you end up scoring activity for its own sake.

Step 2: List the touch points that come before it

Look at the last twenty or thirty customers who converted and list what they did on the way: the message they first replied to, the form they filled in, the call they booked, the link they clicked. You will usually find a handful of steps that almost every buyer took and a lot of activity that made no difference. Score the first group, ignore the second.

Common touch points worth scoring:

• Replied to a message on any channel
• Submitted a form, or started one and left
• Booked an appointment, and attended it
• Clicked a link in an email or a message
• Called in, or completed a call
• Came from an ad or a source that tends to convert

Step 3: Assign points

Give bigger points to actions closer to buying. The numbers below are an example to show the shape of a model, not a recommendation for your business:

Keep the model small at first. Five to eight rules you understand beat thirty you do not.

Step 4: Score the negatives too

A model that only adds points ends up treating persistence as interest. Subtract points for signals that a lead is not going to buy:

• They are outside the area or the business type you serve
• They missed a booked call
• They asked to stop receiving messages
• They said the price is not something they can do

Negative scoring is what stops a contact who has been around for a year from looking like your hottest lead.

Step 5: Make scores fade

Interest goes stale. A lead who was very engaged in March and silent since is not a hot lead in October. Build in decay: when a conversation has been idle for a set time, take points off, for example 10 after two weeks of silence. Scores then describe how close someone is to buying now, not how active they ever were.

Step 6: Set thresholds and route on them

A score is only useful if it changes what happens next. Pick a few bands and give each one an action:

• High (for example 70 and above): assign to a closer, notify them straight away, move the deal to a "sales ready" stage.
• Middle (for example 40 to 69): keep nurturing with useful messages and reminders.
• Low (for example under 40): slow down or pause nurture; do not spend sales time here.

Again, the cut-off points are examples. The right threshold is wherever your leads actually tend to convert, which you find in the next step.

Step 7: Check the model against real conversions

Once the model has been running for a few weeks, look at the people who converted and note what their score was shortly before they did. If most buyers converted somewhere around a particular score, that is your real threshold, and it is often quite different from the number you guessed at the start. Do the same for leads who went quiet: if plenty of them reached your "high" band and never bought, a rule is giving away too many points.

Review it monthly at first, then quarterly. A scoring model is a hypothesis about how your customers buy, and it should change when the evidence does.

Setting up lead scoring in Inflowave

Every contact in Inflowave has a lead score from 0 to 100, stored on the contact and visible in the inbox and contact lists. You can edit it by hand in the contact panel, and workflows change it automatically.

Scoring rules are workflows. Each rule is a trigger plus the Adjust lead score action, which either sets the score to a value or adds or subtracts points. Triggers you can score on include:

• Customer replied, on any channel
• Form submitted and form started
• Appointment booked, completed and no-show
• Email opened and email link clicked
• Inbound call and call completed
• Instagram ad click
• Tag added and deal stage changed
• Conversation idle, which is how you build decay

Routing is a condition on the score. In any workflow, a condition can check whether the contact's score is at least, above, below or at most a number, and branch from there: assign the contact to a closer, send an internal notification, move the deal to a stage in the pipeline, or stop a nurture sequence.

Every score change is recorded. The contact panel shows Lead score (90 days), a bar with the current score and a chart of how it moved over the last 90 days when you hover over it. Scores are banded as high (70 and above), middle (40 to 69) and low (under 40), so a list of contacts can be read at a glance.

Reading the 90-day chart:

• Climbing steadily: a lead working their way towards buying. This is who the closer should call.
• A sharp jump: a single big action, such as booking a call. Act on it the same day.
• Flat for weeks: nothing is happening. Check whether nurture is still running.
• Falling: decay or negative signals. Look at why before spending more time.

Because each client has its own workspace, each client can have its own model: a dentist's buyers behave nothing like a roofer's. Build the rules once per client, check them against that client's conversions, and adjust, all inside the same agency CRM where the conversations and deals already live.

Common lead scoring mistakes

• Scoring opens and page views heavily. They are weak signals. A reply, a booking or a call says far more.
• No negative scoring. Without it, long-time non-buyers float to the top.
• No decay. Old interest looks like new interest.
• Guessing the threshold and never checking it. Compare scores with real conversions and move the cut-off.
• Too many rules. If nobody can explain why a lead has 74 points, nobody will trust the number.
• A score that changes nothing. If the score does not route leads, assign owners or change follow-up, it is decoration.

FAQ

What is a good lead score?
There is no universal number. A "good" score is wherever your own leads tend to convert. Look at the scores of recent customers shortly before they bought, and use that as your threshold.

How many points should each action be worth?
Give the most points to actions closest to buying, such as booking or attending a call, and the fewest to weak signals like email opens. Start simple, then adjust against real conversions.

What is negative lead scoring?
Subtracting points for signals that a lead will not buy, such as being outside the service area or missing a booked call. It keeps poor-fit contacts from looking hot.

What is lead score decay?
Reducing a score when a lead goes quiet, for example after two weeks without a reply, so the score reflects interest now rather than interest months ago.

Can I change a lead's score by hand?
Yes. In Inflowave you can edit the score in the contact panel, and workflows keep adjusting it from there.

How often should I review my lead scoring model?
Monthly while you are setting it up, then every quarter, or whenever you change what you sell or how you sell it.