|8 min read

Reviewed by LeadPilot Editorial

Lead Scoring Matrix: How to Build a Two-Axis Grid That Tells Your Team What to Do Next

Learn how to build a lead scoring matrix: choose two axes, write level rules, map each cell to an owner and action, and test the grid on your own CRM data.


A lead scoring matrix is a grid that places each lead in a cell by crossing two dimensions, usually how well the lead matches your ideal customer and how ready the lead is to talk to sales. Each cell has a named action and an owner. A points model squeezes everything into one number. A matrix keeps the two dimensions separate, so your team can see why a lead is a priority and what to do about it.

This guide shows how to build a matrix, how to write the rules for each level, and how to check whether your grid is sorting leads usefully. The grid and the cell actions below are recommendations for you to adapt.

Why use a matrix instead of a single score

A single total can hide two very different situations that produce the same number:

  • A lead that is a poor match for your offer but asks pointed questions about pricing and timing.
  • A lead that is an excellent match but is only browsing.

Both could land in the middle of a points scale. Yet the right responses differ. The first may deserve a polite, fast disqualification or a referral. The second may deserve a low-effort nurture sequence and a check-in later.

A matrix makes that difference visible. The trade-off is that you manage more states than a single number gives you. If your team is small and your lead volume is low, a simple list with a few labels may be enough. A matrix pays off when your team routinely asks, "Why is this lead flagged?" and nobody can answer from the score alone.

Step 1: Choose your two axes

The most common pairing, and the one this guide uses, is:

  • Fit (rows): Does this person or company look like someone you can serve profitably?
  • Readiness (columns): Is there a real need with a plausible timeline, and is the lead willing to take a next step?

You can substitute other axes if they better describe your sales motion. For example, a services firm might replace readiness with "project urgency," and a SaaS company with a free tier might use "product usage" on one axis. Two rules apply whichever you choose:

  1. The axes should be as independent as possible. If one axis is mostly a proxy for the other, the grid collapses into a diagonal line and you have built a single score with extra steps.
  2. Every criterion should be something you can actually capture. If you cannot collect company size or timeline before a sales call, it cannot sit in a rule.

Step 2: Define three levels per axis in observable terms

Three levels per axis give you nine cells, which is usually enough to separate actions without creating a maintenance burden. Write each level as a rule someone else could apply and get the same answer.

Fit levels (examples for a B2B services or SaaS business):

  • Strong fit: Meets all of your must-have criteria, such as target industry, company type, and a use case you serve well.
  • Possible fit: Meets some criteria, or a key criterion is unknown.
  • Weak fit: Clearly outside your target, such as a student project, an unsupported use case, or a geography you do not serve.

Readiness levels:

  • Ready: States a specific need, gives a timeframe, and has accepted or asked for a next step such as a call.
  • Exploring: Asks substantive questions but has no stated timeline or next step.
  • Early: General browsing or a single informational question, with no stated need.

Notice the "possible fit" level treats unknown data as its own state. Missing information is not the same as a bad answer, and treating it that way can quietly route good leads into the wrong cells. A useful habit is to record "unknown" explicitly rather than defaulting to the lowest level.

Step 3: Map every cell to an action and an owner

The grid below is a starting recommendation. Edit the actions to match your team's capacity.

ReadyExploringEarly
Strong fitPersonal follow-up by sales as soon as possible; book the callSales or marketing sends relevant material and a soft invitation to talkAdd to nurture; revisit if behavior changes
Possible fitSales asks the missing fit questions, then decidesCollect missing fit information through a short exchangeLight nurture only
Weak fitReply politely, redirect or refer; no sales timeSelf-serve resources onlyNo action

Two design notes:

  • Every cell needs one owner. A cell labeled "marketing and sales" usually means nobody acts. Name a role or a queue.
  • Check which before you celebrate.

Step 4: Walk one lead through the grid (hypothetical example)

The following scenario is invented for illustration. It is not a customer result.

A visitor to a web design agency's site asks whether a five-page site can launch before a conference. The agency's published fit criteria are small to mid-sized B2B companies that already have copy and branding. The visitor says the conference is six weeks away and the copy is done.

  • Fit: Meets the stated criteria, so strong fit.
  • Readiness: Specific need, stated deadline, and a request for a next step, so ready.
  • Cell and action: Strong fit and ready. The owner books a scoping call quickly and receives the deadline and the visitor's questions as context.

Now change one detail. The visitor says they are a student comparing agencies for a class assignment. Fit drops to weak, readiness may still look like exploring, and the action becomes self-serve resources. A single points total might have scored that conversation similarly to the first, because both asked about timelines and pricing. The matrix separates them.

Step 5: Test whether the grid actually sorts leads

A matrix is a hypothesis about which leads deserve attention. You can check it against your own records without any outside benchmark.

A reader-run check:

  1. Export a sample of recent leads from your CRM, including those that were never contacted if you can.
  2. Apply your fit and readiness rules to each lead by hand, or have two people do it independently on the same leads.
  3. For each cell, record how many leads fell in it and what happened next: no response, meeting held, opportunity created, closed-won, closed-lost, disqualified.
  4. Compare the outcomes across cells.

What to look for:

  • Do the cells you rank highest produce better outcomes than the ones you rank lowest? If strong-fit, ready leads do not progress further than weak-fit, early leads, your definitions need rework. Either the criteria do not predict what you thought, or your follow-up process is the bottleneck.
  • Do two people place the same lead in the same cell? If not, your wording is too vague. Rewrite the level in terms of specific, checkable facts.
  • Is a cell mostly "unknown"? That points to a data capture gap, not a scoring problem. Fix the question you are not asking.
  • Are any cells empty? Your axes may be correlated, or one level may be unreachable with the data you collect.

Keep in mind that a small sample can mislead. Treat the result as a prompt to investigate, not as proof, and repeat it periodically as your offer and audience change.

Common mistakes when building a lead scoring matrix

  • Too many cells. A 5x5 grid has twenty-five states, and few teams can define a distinct action for each. Start at 3x3, and only split a cell when you can name a different action for each half.
  • Using activity as a stand-in for need. Page views and email opens are easy to collect, but they may not reflect a stated need. Weigh what a lead says more heavily than what they click when you can capture it, and verify with your own outcomes that this holds for your business.
  • Letting the rules go stale. Revisit definitions when your pricing, target market, or product changes. Put a recurring review on the calendar and assign an owner.
  • Hiding the reason. Whatever tool places a lead in a cell, the sales rep should see the facts behind the placement: what the lead asked, what they said about timing, and which fit criteria were met.
  • No path for exceptions. Give reps a way to move a lead to a different cell and record why. Those overrides are useful feedback on where your rules are wrong.

Capturing the inputs a matrix needs

A matrix only works if you can fill in both axes. Forms are a poor fit for some of these inputs. A visitor may not want to complete several fields before asking a question, and a timeline or a stated need is often easiest to learn through conversation.

If you use a chat-based tool for this, the useful question is whether it can collect the specific inputs your rules need and pass them on in a form your team can read.

You choose the goal, such as capturing a lead, starting a signup, or booking a call, and you add the qualification questions that change that decision. It asks one relevant question at a time and sends your team contact details along with intent, fit, timing, a score, and the conversation itself.

Those outputs map naturally onto a matrix: fit and timing details can inform your row and column placement, and the saved conversation lets a rep see the reasons. You would still define the levels and cell actions yourself, and you should validate them against your own outcomes as described above.

If you want to test this approach, you can Build your LeadPilot agent and compare its handoffs with the grid you have drawn.

A short checklist before you roll out the grid

  • Two axes, each with three levels, written as observable rules.
  • "Unknown" is recorded as its own state.
  • Every cell has one owner and one action.
  • Reps can see the facts behind a lead's placement and can override it with a reason.
  • You have run the sample check against your own CRM records and adjusted the definitions.
  • A review date is on the calendar.

A matrix that three people apply consistently is more useful than a detailed one nobody follows.

Put your website conversations to work

Build a LeadPilot agent from your website and review it before you publish.

Start Free