Lead scoring assigns a simple, actionable priority to each prospect so sales and marketing know who to contact first. It works by combining fit signals, like job title and company size, with engagement signals, like email clicks and website visits, into a single number or band that tells your team where to spend their time. This guide walks through the models, the build process, and the Canadian compliance rules that govern automated follow-up.

TL;DR:

• Using a 0 to 10 scoring scale improves clarity and decision speed for beginners, instead of a more complex 0 to 100 range.
• Prioritize active behaviours like demo requests and pricing page visits over passive content engagement for higher weighted scores.
• Regularly review conversion rates, sales acceptance, and time-to-close metrics to keep the scoring model aligned with real outcomes.
• Ensure strict compliance with Canada’s anti-spam legislation by confirming explicit consent and maintaining proper record-keeping before automating email follow-ups.
• Keep the scoring model simple with 4 to 6 explicit attributes and behaviors, and conduct quarterly audits to prevent data quality issues from skewing results.

Table of Contents

• What is lead scoring, and why do fit and engagement both matter?
• Explicit versus implicit signals: what to collect and how to subtract points
• Lead scoring models and criteria you can adapt today
• How do you build a lead scoring model from scratch?
• Putting scores to work: CRM, automation, routing, and compliance
• Keeping a model accurate: what to measure and how often to review it
• Common pitfalls and best practices worth building in early
• How we approach lead scoring in practice
• Want help building or refining your lead scoring model?
• FAQ
• Sources

What is lead scoring, and why do fit and engagement both matter?

Lead scoring is a method for ranking prospects by how likely they are to become customers, based on who they are and how they behave. Instead of treating every form fill the same way, we assign points to the traits and actions that correlate with a closed deal, then total those points into a score that sales can act on without guessing.

A simple example makes this concrete. Imagine a marketing manager at a mid-sized company who downloads a pricing guide and visits the demo page twice in one week. Her job title and company size might earn her a moderate number of points on the fit side, while the demo page visits and content download add more points on the engagement side. A combined score well within the model’s range puts her solidly in “warm” territory, worth a sales call this week rather than a slow-drip newsletter.

That example illustrates the two dimensions every scoring model rests on:

• Fit (explicit): who the prospect is, drawn from form data, firmographic enrichment, or sales notes, such as industry, company size, or role.
• Engagement (implicit): what the prospect does, drawn from tracked behaviour like page visits, email opens, or webinar attendance.

Fit tells you whether someone matches your ideal customer profile; engagement tells you whether they are actually interested right now. A prospect can score high on fit and still be a poor sales priority if they have shown no recent activity, and the reverse is just as common: a smaller, less obviously qualified company that engages heavily with your content can outperform a textbook-perfect account that never opens an email.

Once fit and engagement are combined, most teams reduce the result to a plain classification: hot, warm, or cold. A hot lead typically scores high on both dimensions and goes straight to a sales rep. A warm lead might have strong fit but light engagement, so it enters a more targeted nurture sequence. A cold lead sits in general nurture until its score improves or it ages out. This three-tier structure keeps the system usable for everyone on the team, even before any automation is in place.

Explicit versus implicit signals: what to collect and how to subtract points

Building a useful score starts with knowing which signals to track and where to find them. Explicit signals come directly from the prospect, usually through form fields, sales conversations, or data enrichment tools. Implicit signals come from observed behaviour, usually tracked through your website, email platform, or CRM.

Explicit attributes worth collecting include:

• Job title or seniority, since a director has more buying authority than an intern.
• Company size or revenue band, which indicates budget and likely deal size.
• Industry, when your offering performs better in certain verticals.
• BANT details (budget, authority, need, timeline), typically captured during a discovery call or qualification form.

Implicit behaviours worth tracking include:

• Pricing or demo page visits, a strong signal of active buying intent.
• Webinar attendance or content downloads, showing topical interest.
• Email click-throughs on key offers, distinguishing engaged subscribers from passive ones.
• Visits to comparison or review pages, which often signal late-stage evaluation against alternatives.

Not every signal should add points. Negative scoring deducts points for behaviours or attributes that make a lead less likely to convert or harder to reach, and it is just as important as the positive side of the model. A prospect using a personal email address instead of a work one, a title like “student” or “retired,” or a company size far below your minimum deal size are common candidates for deductions. Unsubscribing from emails, bouncing repeatedly, or going cold for 90 days should also pull the score down, since an outdated score is worse than no score at all.

Data quality deductions matter too. A form filled with placeholder text, a clearly fake phone number, or a mismatched company domain should trigger a penalty or a flag for manual review rather than a free pass into the hot lead queue.

Pro Tip: Audit your form fields once a quarter; a single mistyped required field can quietly corrupt months of scoring data.

Lead scoring models and criteria you can adapt today

Most teams choose from a small set of model types rather than inventing one from scratch. A rule-based points model, where each attribute or action carries a fixed point value that totals into a threshold, is the easiest to build and explain, which is why it remains the default starting point for teams new to scoring. A fit-and-engagement matrix, where fit is scored on a letter scale and engagement on a number scale, then combined into a grid like A1 or B3, gives a clearer visual sense of where a lead sits relative to your ideal customer. Oracle’s lead scoring guidance describes exactly this approach: profile fit scored A through D and engagement scored 1 through 4, with combinations like A1 representing the best-matched, most-engaged leads ready for immediate follow-up.

Predictive models sit a step further along. These use historical won and lost data to let software weight signals automatically rather than relying on a human-assigned point value. They can outperform manual rules once you have enough closed deals to train on, but they are rarely the right starting point for a team scoring its first few hundred leads, since there is no track record yet to learn from.

Whichever model you choose, the underlying criteria tend to repeat across industries, with relative weight adjusted to your business:

• Role and seniority, usually weighted heavily since it affects buying authority.
• Company size or budget indicators, weighted according to your typical deal size.
• Active buying behaviours, such as demo requests or pricing page visits, weighted higher than passive content consumption. Adobe’s lead scoring guide makes this distinction explicit, recommending that active behaviours like requesting a demo or submitting an RFP carry more weight than passive repeat visits to a blog.
• Recency of engagement, since a click from yesterday means more than one from three months ago.
• Source of the lead, since a referral or a direct demo request often converts better than a generic content download.

For the scale itself, simpler tends to work better. Many practitioners recommend a compact 0 to 10 scale over a 0 to 100 scale, because a smaller range forces clearer differentiation between leads and speeds up decisions. A classification layer on top, hot (8 to 10), warm (4 to 7), cold (0 to 3), gives sales a plain-language label to act on without doing mental math on every lead that lands in their queue.

How do you build a lead scoring model from scratch?

Building a scoring model is less about finding the perfect formula and more about getting the right people aligned before a single point value is assigned. Pipedrive’s guidance on lead scoring for small and mid-sized businesses puts it simply: start with a small, manageable set of criteria, assign weighted points, pilot against historical data, and refine from there rather than trying to perfect the model on paper first.

1. Align sales and marketing on a shared definition of a qualified lead. Before choosing criteria, agree on what an MQL (marketing qualified lead) and your ideal customer profile actually look like, since a scoring model built without this agreement will be second-guessed the moment it reaches sales. Research on CRM and lead scoring readiness found that most organizations already own a CRM, but far fewer have built shared dashboards between departments, and treats that alignment gap as the main barrier to useful scoring, well ahead of any technology choice.
2. Pick a small, focused set of criteria. Choose 4 to 6 explicit attributes and 4 to 6 implicit behaviours to start. Resist the urge to track everything your CRM can capture; a model with 20 variables is harder to explain, harder to debug, and no more accurate in most cases than one with 10 well-chosen ones.
3. Assign weights and set thresholds. Put each criterion on your chosen scale (0 to 10 works well for most teams), decide which signals deserve more weight, and set the score bands that separate hot from warm from cold.
4. Map each score band to a specific action. A hot lead might trigger an immediate call from a sales rep, a warm lead might enter a targeted three-email sequence, and a cold lead might simply continue in general nurture content.
5. Pilot the model against historical data before going live. Run your proposed scoring rules against deals you already know the outcome of. If your highest-scoring historical leads mostly turned into closed-lost deals, something in the weighting needs adjusting before the model ever touches a live prospect.
6. Implement in your CRM or marketing automation platform, and agree on SLAs. Once the model is validated, build it into your workflow tools and set a clear service-level agreement for follow-up, such as a 24-hour response window for hot leads, so the score actually changes behaviour rather than sitting unused in a database field.

Pro Tip: Build your ideal customer profile before you touch the scoring spreadsheet; explicit criteria only work when they are grounded in a clear picture of who actually buys from you.

Putting scores to work: CRM, automation, routing, and compliance

A score only matters if it changes what happens next. Most CRM and marketing automation platforms let you store the score as a custom field on the contact or lead record, then trigger workflows automatically when that field crosses a threshold. A lead crossing into the hot band might automatically create a task for a sales rep, send a Slack or email alert, or move the record into a different pipeline stage. A lead sliding into the cold band might get removed from active sales outreach and routed back into a general nurture sequence instead. Marketing automation built around these triggers is what turns a static score into an operational system rather than a number nobody checks.

Typical automation triggers include:

• Score crosses the hot threshold: notify the assigned sales rep and log a follow-up task.
• Score drops after 60 to 90 days of inactivity: move the lead back into nurture and reduce sales priority.
• Negative signal detected (bounced email, unsubscribe, disqualifying title): flag for review or remove from active sequences.
• Score hits the warm band: enrol in a targeted three-to-five email sequence tied to the prospect’s declared interest.

Any automated commercial electronic message triggered by a score change has to comply with Canada’s anti-spam legislation. Under CASL, businesses need express or implied consent before sending a commercial electronic message, must identify the sender clearly, and must include a working unsubscribe mechanism. Proof of that consent needs to be kept on file, contact information must stay valid for at least 60 days after a message is sent, and any unsubscribe request has to be actioned promptly. Before wiring a scoring band to an automated email sequence, run through this short checklist:

• Confirm consent (express or implied) exists and is documented for each contact in the sequence.
• Display a clear, accurate sender identity in every automated message.
• Include a functional unsubscribe link that processes requests without delay.
• Retain consent and unsubscribe records in case of a compliance review.

The Competition Bureau’s guidance on anti-spam enforcement also flags that false or misleading electronic representations carry their own penalty risk, separate from CASL’s consent rules, which is worth keeping in mind if your nurture copy ever overstates what a product does.

Keeping a model accurate: what to measure and how often to review it

A lead scoring model is never finished. Scores drift out of alignment with reality as your product, market, and buyer behaviour change, so ongoing measurement is what keeps a scoring programme useful rather than decorative.

Four metrics tell you most of what you need to know:

• Conversion rate by score band, to confirm hot leads actually close more often than cold ones.
• Sales acceptance rate, to check how often sales actually acts on the leads marketing hands off as qualified.
• Lead velocity, to see how quickly leads move from one score band to the next.
• Time-to-close, broken out by score band, to confirm higher scores correlate with shorter sales cycles.

When those numbers drift, won and lost deal analysis is the most reliable way to recalibrate. Pulling a sample of recently closed deals and checking whether the winners scored higher than the losers at the time of handoff tells you quickly whether a particular criterion is pulling its weight or quietly skewing the model. Practitioner guidance on lead scoring treats this kind of ongoing testing against real revenue outcomes as essential, since a model built on assumptions drifts away from market reality without it. A/B testing different weightings on a subset of leads before rolling a change out to the whole pipeline reduces the risk of a bad adjustment doing more harm than good.

A practical cadence works well for most teams: pilot the model for 30 to 60 days against historical and live data, run quick monthly checks on conversion by band, and schedule a deeper quarterly review that revisits weightings, thresholds, and the criteria list itself.

Common pitfalls and best practices worth building in early

Most scoring programmes fail for a small number of repeatable reasons, and most of them show up before the model ever reaches sales.

• Overcomplication, where a model with too many weighted criteria becomes impossible to explain or maintain.
• Ignoring data quality, since a score built on incomplete or fake form data is worse than no score at all.
• Lack of alignment, where sales never agreed to the MQL definition marketing used to build the model, so handoffs get ignored.
• False reliance on the score, treating a high number as a guarantee rather than a prioritization signal that still needs human judgment on complex or unusual deals.

A few rules of thumb protect against most of this. Keep the model simple enough to explain in one sentence. Review it on a fixed schedule rather than waiting for a complaint from sales. Build in a human check for edge cases, since an odd combination of signals can produce a misleadingly high or low score that a quick look would catch immediately.

Before any lead is handed to sales, a short pre-handoff checklist helps: confirm the score crossed the agreed threshold, confirm no negative signal or data-quality flag is attached, and confirm the lead has been active within the recency window your model assumes.

Pro Tip: Keep a human-in-the-loop for any deal large enough to matter; a score is a prioritization cue, not a verdict.

How we approach lead scoring in practice

We treat lead scoring as a discipline that lives or dies on alignment, not on the sophistication of the model itself. Our process starts the same way for every project: agree on a shared ICP and MQL definition with both marketing and sales stakeholders in the room, pilot a simple scoring structure against real historical data, and only then build out the automation and reporting layer around it. That sequencing matters more than most teams expect, since a technically elegant model built before alignment tends to get ignored the first time sales disagrees with a handoff.

Because every engagement is custom-built around each project’s own data and sales process, we refine weightings and thresholds on an ongoing basis rather than treating the initial model as finished work. If you want a closer look at how that discipline connects to broader lead quality, our note on lead quality versus lead quantity walks through why a tighter scoring model often outperforms a bigger top-of-funnel.

— Stephen Ramkissoon

Want help building or refining your lead scoring model?

We work with small business owners and growing teams who need a reliable scoring model but do not want to hire a full-time analyst to build and maintain it. Our analytics and reporting, marketing automation, and digital marketing strategy services exist specifically to turn a spreadsheet full of explicit and implicit signals into a working system your sales team actually trusts.

Because our subscription plans run without long-term contracts, a pilot scoring project can start small, prove itself against your own won and lost data, and expand once it earns sales’ confidence. If that sounds like the kind of collaborative, data-driven support your team needs, our subscription-based marketing model is built for exactly this kind of ongoing, measurable work. Visit Marketingguardians to talk through an audit or pilot for your own lead scoring programme.

FAQ

How is a lead score calculated?

A lead score is calculated by assigning point values to explicit attributes, like job title or company size, and implicit behaviours, like email clicks or page visits, then adding those points into a single total. That total is usually placed on a scale, such as 0 to 10, and mapped to a classification like hot, warm, or cold so sales knows how to prioritize follow-up.

What is the 3-3-3 rule in sales?

The 3-3-3 rule is not a standard or widely defined lead scoring framework, so there is no single agreed version of it worth repeating as fact. If you have encountered this term from a specific source, it is worth checking that source’s own definition rather than assuming it maps onto standard lead scoring criteria.

How do you score a lead?

Scoring a lead means combining a small set of explicit criteria, such as role, industry, and company size, with implicit behaviours, such as content downloads and demo requests, into a weighted total. Pipedrive’s lead scoring guidance recommends starting with a short list of criteria on each side, piloting the model against historical data, and refining the weights once real outcomes are available.

What counts as a reasonable SLA for following up with a hot lead?

There is no single universal standard for lead follow-up timing, since it depends on your sales cycle and industry, but many teams set an internal SLA of 24 hours or less for their highest-scoring leads to prevent them from going cold. The exact window should be agreed between sales and marketing during the alignment stage, before the scoring model goes live, rather than following a strict predefined deadline.

Should a beginner use a 0 to 10 scale or a 0 to 100 scale?

A compact scale like 0 to 10 is generally easier for a beginner team to build, explain, and act on, since it forces clearer differentiation between leads and speeds up day-to-day decisions. A 0 to 100 scale can offer more granularity, but it tends to suit teams with a mature, data-rich model rather than a first attempt.

Sources

• Getting consent to send email
• Monday
• Effective lead scoring tips and techniques for SMBs (Pipedrive)
• Lead scoring guide (Adobe)

Recommended

• Google Business Profile optimization: your 30-day Canadian guide
• Website maintenance plan: your complete Canadian guide
• LinkedIn lead generation: your 30-day B2B campaign plan