Lead scoring: how to qualify leads when you do the selling

Marcel GarciaUpdated Oct 4, 2026

Lead scoring rates each contact on how well it fits what you sell and how close it is to buying, so you know who gets your time first. A qualified lead fits, can afford it and comes from the person who decides. If you do the selling, you don't need points: three questions and one rule are enough.

If you sell something expensive and you do the selling, every hour in your calendar is worth money: every call with someone who can't buy is an hour you don't spend with someone who can. The lead scoring manuals assume a marketing team, a sales team and reps who qualify leads and book calls. This guide doesn't: it's written for a business where the person who qualifies, the person who sells and the person who decides are usually the same one.

It is signed by Marcel Garcia, founder of HAKI. The form you'll see is ours, the one that decides today who reaches our calendar. If you have a sales team and want to know how to set up scoring in your CRM, its documentation will serve you better than this guide.

What lead scoring is and how a points model works

Lead scoring gives each contact a score built from two kinds of signal. Profile signals say whether it fits: job title, industry, company size. Behaviour signals say how close it is to buying: visiting the pricing page, opening emails, downloading a guide, coming back to the website. Signals that rule a lead out subtract: a student, a competitor, someone outside the area you serve. When the total crosses a threshold, the contact goes to sales.

An invented example with round numbers: managing director, 20 points; an industry you serve, 15; a pricing page visit, 10; each opened email, 2; a student, minus 50. From 50 up, someone in sales calls. Good weights aren't set by eye: they come from comparing the leads that ended up buying with the ones that didn't. That's why a points model needs hundreds of leads and plenty of closed sales before it tells you anything useful.

What a qualified lead is

A lead is a contact who has raised their hand: filled in a form, written to you or called. It is qualified when it meets three conditions at once: what they need fits what you sell, they can afford it, and the person writing to you is the one who decides, alone or with their partners. If one fails, it isn't a qualified lead, however interested it looks. Qualifying a lead means checking those three things before you give it an hour.

The manuals use two acronyms. An MQL (marketing qualified lead) is a contact marketing considers interested, mostly because of what it did: downloaded a guide, opened emails, came back to the website. An SQL (sales qualified lead) is one that sales has validated and that deserves a sales conversation. The split is useful when there are two teams and someone has to decide when a contact moves from one to the other. When one or two people sell, both are the same inbox, and the only real question is this: does it deserve a call or an email?

The BANT method when you do the selling

The classic qualification method is BANT: Budget, Authority, Need and Timeline. Need and authority work as they are; the timeline comes out in the conversation anyway. Budget, in an expensive sale, is of little use: "what's your budget?" usually gets you a made-up figure, or one meant to talk your price down. It works better to ask for something the buyer knows without working it out and that points to the same thing, their ability to pay. If you sell to companies, what an average client spends with them: someone who charges 5,000 euros per client understands what it costs to win one. If you sell to individuals, the size of their project.

Our form is BANT, adapted. Fit, the N for need, comes from "What do you sell?", which a person reads. Authority is "Who decides the marketing budget?". Budget is replaced by "What does an average client spend with you?". And the timeline is left for the call.

Lead scoring by the manual, and lead scoring when you do the selling

The manuals are written for teams with SDRs (reps who find and qualify leads and book calls for whoever closes the sale). If you do the selling, you're in the right-hand column.

	A team with dedicated sales reps	One or two people selling
What gets scored	Email opens, downloads, pricing page visits, job title and industry	What the contact answers: what they need, how big their business or project is, who decides
Where the data comes from	A platform that records every click	The form and your CRM
Tool	A CRM with automatic scoring, such as HubSpot or Salesforce	A form, one rule and the CRM you already have
The threshold	A points figure, say 50, that someone recalibrates	Yes or no: it meets the conditions or it doesn't
What happens when it's crossed	A rep picks it up and calls	Your calendar opens, with the person who decides on the other side
What happens to the rest	Nurturing until it scores enough points	A reply by email: nobody is left unanswered
When it's reviewed	Whenever someone recalibrates the weights	Every month, against real sales
How many leads it needs	It needs volume and history before the weights mean anything	It works from the first one

How we do it at HAKI: three questions and one rule

Besides the contact details and a free-text message, our form asks three questions. One decides where each request goes; a person reads the other two. It's a real example, not a template: the same form you'll find at the end of this page.

"What do you sell?"

Free text, with examples to guide the answer: dental implants, consulting, software for accounting firms, full renovations. It isn't part of the automatic rule: a person reads it, because it says more about whether we fit than any checkbox.

"What does an average client spend with you?"

Four options: under 1,000 €, 1,000 to 5,000 €, 5,000 to 25,000 € and over 25,000 €. It's what a client is worth to you, and it says whether you sell on value and whether there's margin to reinvest in winning clients. Under 1,000 €, the calendar doesn't open.

"Who decides the marketing budget?"

Three options: me, me with my partners, or someone else. It doesn't close the calendar: at a mid-sized or large company, the person filling in the form is rarely the one who signs, and that doesn't make it a worse client. It tells us who will need convincing and how to prepare the call.

The rule

If an average client spends 1,000 € or more with you, the form takes you to the calendar to pick a day and time for a 30-minute discovery call with Marcel, with your three answers already filled in. If not, you see a thank-you message on the same page and get a reply by email. The calendar isn't open to everyone, and that's the point.

Everything reaches the CRM

Every request lands in the CRM with its verdict, qualified or not, and with its origin: the page it arrived on and the channel. That's how you can check whether the rule is right: how many of the people who opened the calendar became clients, and whether any of those left out ended up buying anyway. It's growth marketing applied to the front door: one rule, measured against sales.

How to qualify leads in six steps

None of them needs a tool you don't already have.

1. Look at your last 20 clients and your last 20 "no"s

What the buyers had in common: how much they spent with you, who decided, how they found you. And why the others didn't buy. If you don't have twenty of each, use what you have: the pattern usually shows up sooner.

2. Pick two conditions that rule a lead out, not ten that add points

Points models add for every sign of interest. What saves you more is the part that subtracts: whatever, if missing, makes the sale impossible. A ticket that's too low, someone who doesn't decide, a service you don't offer.

3. Turn them into questions people answer off the top of their head

Closed options in bands, not open figures: a band takes no working out and leaves less room to dress up the answer. A few examples: a full-renovation company can ask the size of the home; software for accounting firms, how many clients the firm handles; a consultancy, how many people work on the enquirer's team.

4. Decide what happens on each side

Whoever qualifies goes to your calendar, with their answers already filled in so you don't ask twice. Whoever doesn't gets an email reply: a guide, the name of someone who fits better, or a polite "not now".

5. Send the verdict to your CRM and review it every month

Every lead with its yes or no and its origin. Once a month, compare it with sales: if good clients are left out, the rule is too strict; if the calendar fills with pointless calls, too loose. It's the first piece of a marketing automation that works: without it, you automate follow-up for people who were never going to buy.

6. Move to points when one person can't keep up

Points-based scoring makes sense when more qualified requests arrive than one person can handle, and you have enough closed sales to know which behaviour predicts a purchase. Before that, it's complexity with no data to hold it up.

The mistakes that make lead scoring lie

Scoring interest instead of the ability to buy

An opened email pays no invoice. A contact can read your whole website and have neither the money nor the signature.

Too many questions

Every question that doesn't change the decision is a reason to close the tab. Three are enough to decide whether there's a call; context fits in one free-text field, and the rest gets asked on the call.

Filtering without replying

Today's "no" refers someone tomorrow, or comes back in a year with another business and another ticket. A filter that leaves people without a reply saves you a call and costs you your reputation.

Never checking it against sales

Without comparing the verdict with who ended up buying, scoring is an opinion with decimals. And every call that shouldn't have happened pushes up your customer acquisition cost.

Copying the model of a 200-person software company

Their weights come from thousands of leads and a team that recalibrates them. Yours would come from twenty. A model like that, at your volume, only looks precise.

Telling Google Ads that any form is a conversion

With automated bidding, Google optimises for whatever you tell it is a conversion. If that's any form, it looks for people who fill in forms. If it's the appointment with someone who fits, and better still the sale imported from your CRM, it looks for people who resemble your buyers.

What it looks like when it works

Two published cases and two ways to make sure your time goes to the people who matter: filter every request before replying, and count the appointment as the conversion, not the click. Neither publishes a qualification rate, and we don't publish a client's.

Nuria Marco: qualifying without a single call

Nuria didn't want calls of any kind. Whoever asks for an appointment fills in a prequalification form and she sees whether she can help; if she can, they get an automated message with the method, the offer and the payment options, and follow-ups if they go quiet. In her words: "a longer waiting list, a more consolidated schedule and higher-quality clients".

Laura Jové: the appointment, not the click, as the conversion

Her whole website has one call to action, a free 20-minute assessment session, and in Google Ads the conversion that counts is the booked appointment, not the click: 246 appointments from ads in 2025 and 339 from January to September 2026, with a 15x ROAS in 2026.

Keep going with the other guides

CAC

What it costs to win a client, how to calculate it per channel and what a good LTV:CAC ratio is.

Go-to-market

How to take an offer to market: who, which problem, at what price, through which channel and with what proof.

Sales pipeline

The stages of a long, expensive sale, and the numbers that show where it gets stuck.

Selected projects

Embat

Fintech, Tech & SaaS

Clínica Albareda

Health

Laura Jové

Health

View all projects

A growth partner is paid partly on new clients: that's why it filters

It charges a base plus a variable part tied to new clients, so the model only works with the two conditions in our form: a client worth 1,000 euros or more, because below that the variable part doesn't pay for the work, and someone who decides. What it is, what it costs and when it fits.

What a growth partner is

An opinion, signed

Points-based lead scoring solves the problem of teams with more leads than hours. Almost no business that sells something expensive with one or two people is there: it gets fewer requests than it would like and still has a calendar full of calls that should never have happened. What it lacks isn't a hundred-point model, it's a door with judgement. I gave ours three questions and one rule. The hard part isn't writing them: it's accepting that your calendar isn't for everyone who wants to talk to you.

Marcel Garcia, founder of HAKI. Growth partner to businesses that sell on value, from Barcelona.

Contact

This is the form the guide talks about. Tell us what you sell.

hello@haki.marketing

Name

Surname

Phone

Company

Role

What do you sell?

What does an average client spend with you?

Who decides the marketing budget?

Message

Lead scoring questions

It's rating each contact on how well it fits what you sell and how close it is to buying, so you know who gets your time first. One that clears the bar is a qualified lead: it fits, can afford it and comes from the person who decides.

An MQL is a contact marketing considers interested, mostly because of what it did (downloads, visits, opened emails); an SQL is one sales has validated for a sales conversation. The line matters when there are two teams. If you do the selling, what counts is whether the contact deserves a call or an email reply.

By temperature: cold, warm and hot. By stage: MQL (marketing considers it interested) and SQL (sales has validated it). If you do the selling, only two matter: the one that deserves a call and the one that deserves an email.

Partly. Who decides and what the client needs are still the key questions, but budget is better replaced with a figure the buyer knows without working it out: for a company, what an average client spends with them; for an individual, the size of their project. You get a signal of their ability to pay without inviting a negotiation.

Just enough to decide whether there's a call: in our case, three, on top of the contact details and a free-text message. Anything that doesn't change that decision gets asked later, in the conversation.

No, while one or two people do the selling. Both offer scoring with rules that add or subtract points (job title, industry, visits, opened emails) and a threshold, and they work out the score themselves; their predictive, AI versions learn the weights from your closed sales. Both pay off with plenty of leads a month and a solid sales history. Until then, a "qualified: yes or no" field filled in by your form does the same job in any CRM.

Yes, the logic is the same: fit, ability to pay and who decides, and it's worth it whenever one call too many costs you money. The questions change: you ask a company what an average client spends with them and who signs; an individual, what treatment or project they want and whether they decide alone or with a partner.

Always reply, by email, and record in the CRM why they didn't qualify. A useful guide, the name of someone who fits better or an honest "not now" keeps the door open for when their business changes, or for a referral. And if many fail on the same condition, either the rule is too strict or your marketing is attracting the wrong people.

Lead scoring decides who is ready to talk to you; lead nurturing stays with everyone else, with emails and content, until they are or until it's clear they don't fit. One filters; the other warms leads up. If you do the selling, the filter comes first, then the automated follow-up for people who fit but aren't in a hurry: we build it as part of marketing automation.