Your sales team sees dozens of enquiries a day and answers them in arrival order. Meanwhile the customer most ready to buy waits at the back of the queue.

Lead scoring gives each prospect a score based on how likely they are to buy. The point is not to discard anyone but to point the team's energy at the highest odds first. You do not need an AI model or a data scientist. The first version is built from a spreadsheet and an honest look at your history.

Two axes: fit and intent

• Fit: does this person or company resemble your ideal customer? Sector, size, region, budget, role.
• Intent: how close are they to buying now? Visiting the pricing page, asking for a quote or demo, saying "we want to start this month".

Mixing the two is the most common mistake. A highly suitable company in no hurry and a person with an urgent need but no budget can earn the same total score. One needs nurturing content; the other needs a quick, clear "we're not the right fit".

Where the weights come from

From your own history. List the deals you won and lost over the last 6 to 12 months. For each, note where it came from, sector, size, what the first message asked for and how long the decision took. Traits common among wins earn high points.

An example starting point, out of 100:

• In one of your ideal sectors: +20
• Headcount in the target range: +15
• Decision-maker role (founder, director): +15
• Asked for a price or quote: +20
• Start time "this month" or "next month": +20
• Came through a referral: +10
• Outside your region or asking for a service you do not offer: -30

Your numbers will differ. What matters is that each point has a reason found in your own data.

The value of a scoring model is not its accuracy. It is that the whole team looks at the same enquiry the same way.

Tie the score to an action

• 70 and above: called the same day, assigned to a senior person.
• 40 to 69: answered by email or message within 24 hours, with a short call offered.
• Below 40: moved into an automatic information and content flow, reassessed when the need becomes clear.

Why speed matters so much for high scores is covered in our piece on response time.

Mistakes that break the model

• Too many criteria: nobody understands or maintains a twenty-factor model. Five to eight is enough.
• Manual scoring: if reps type the score, it turns subjective. Inputs should flow from the form and CRM.
• Never updating: every three months, review won and lost deals and adjust the weights.
• Not asking in the form: you cannot score intent you never measured. Adding a single question such as "when would you like to start" changes a lot.

Frequently asked questions

Does a small sales team need lead scoring?

With a few enquiries a day you can answer them all quickly, so scoring is optional. Once volume exceeds the team's capacity, or ad leads vary widely in quality, a simple score makes it easier to decide which enquiry can wait.

Is AI scoring better?

With enough history, such as hundreds of won and lost deals, machine learning can find patterns you miss. With little data, a simple rule-based model is more reliable and easier to understand. Where AI helps most today is extracting intent signals from free-text messages.

Should customers see their score?

No. The score is an internal prioritisation tool. Every customer should get a polite, fast first reply; the score only decides who is called the same day and who enters an automatic information flow.

Scoring lives in the CRM; the basics are in what a CRM is. To turn capture, scoring and routing into a system that runs on its own, see our CRM and lead systems service.