ICP lead scoring for your whole prospect list

What is ICP scoring?

Your ICP (ideal customer profile) describes the company and the person your offer helps most: an industry, a size, a region, a function and a level of authority. ICP scoring rates every row of a prospect list on how closely it matches that profile. It is also called a fit score, and it is the first building block of lead scoring.

In practice, each criterion is worth points, weighted by how much it matters to you. A VP Sales at a 120-person software company in a country you serve ticks every box and scores close to 100; a bookkeeper at a five-person construction firm does not. The A, B, C or D tier that follows is how you slice the list: A first, D out of the campaign or parked for later.

Fit, intent, reachability: three scores you should not mix

Complete lead scoring looks at three different questions. Adding them into one number blurs the decision: a perfect-fit prospect you cannot reach and an average one who just raised money are not handled the same way. This tool only computes the first one.

Once the list is sorted by fit, reachability sets the order inside a call session, and qualification happens on the call itself, with a BANT or MEDDIC scorecard.

How to build your ICP in five steps

A score is only as good as the profile behind it. Here is the method the starting profiles in this tool follow.

1. 1

   Start from your best customers, not your wish list

   List the customers who signed fast, stayed long and pay well. What they have in common (industry, size, country, who signed) is your ICP. A dream profile that no customer confirms produces a score that predicts nothing.

2. 2

   Separate the account from the person

   Size, industry and country describe the company; function and seniority describe the contact. A good account with the wrong contact can be fixed by asking who owns the topic. The reverse rarely works.

3. 3

   Write your knock-outs separately

   Competitors, current customers, countries you don’t sell in, interns: these are not fewer points, they are rows to pull out of the campaign. Folding them into the score lets a well-rated competitor slip through.

4. 4

   Weight what actually makes deals close

   If company size decides everything (software that only pays off above 200 seats), give it the biggest weight. If the buyer is always the same role, function weighs more. The starting profiles use reasonable assumptions: adjust them.

5. 5

   Check against your results

   After a few weeks of calls, compare the meeting rate of tiers A, B, C and D. If C converts as well as A, a criterion matters less than you thought: lower its weight or tighten the target.

How the tool computes the score

Six steps, run in your browser, on every row.

1. 1

   Reading the columns

   Headers are recognized in seven languages: Title, Poste, Funktion, Cargo, Ruolo, Functie, Função for the job title; Employees, Effectif, Mitarbeiter, Empleados, Dipendenti, Medewerkers, Funcionários for size; Industry, Secteur, Branche, Settore; Country, Pays, Land, País. A Seniority or Department column (Apollo, Sales Navigator exports) is used as a fallback. You can reassign any column.

2. 2

   Job titles

   A dictionary of more than a thousand expressions maps each title to a function (sales, marketing, leadership, HR, finance, tech, procurement, operations, customer success, product, legal) and a seniority level. It reads abbreviations (CEO, VP Sales, CRO, SDR, AE, PDG, DAF, GF), gender markers (m/w/d, H/F) and avoids the classic traps: an Account Manager is a sales rep, not a manager; a Business Developer is not a software developer; an assistant to the CEO is not the CEO.

3. 3

   Company size

   Ranges are read however they are written: 51-200, 201–500 Mitarbeiter, 1 001 à 5 000, 10k+, 10,001+, More than 500, or a plain number. Inside your target range: full points. A neighboring range (down to half your minimum or up to double your maximum): half.

4. 4

   Industry and country

   The industry is mapped to about twenty families from the LinkedIn or Apollo taxonomy, or free text in seven languages. The country is read as a name, an ISO code or inside a location (“Austin, Texas, United States”); if missing, it is inferred from an international dialing code or a country web domain, and flagged as inferred.

5. 5

   Seniority and weights

   Each criterion counts for its weight, scaled to 100. Seniority gets full points inside your target, half if it is one level above (a director when you target managers), zero below.

6. 6

   Score, completeness and tiers

   The score is computed on the criteria your file fills in: a missing company size costs nothing, it lowers completeness, shown next to it. Tier A from 80, B from 60, C from 40, D below. Under 60% completeness, the tier is provisional: fill in the gaps before you trust it.

The same job title in seven languages

International lists mix languages. These are some of the titles the tool reads as the same function and seniority level.

What to do with each tier

The tier is a decision, not a grade: it tells you where each row goes next.

• A Tier A: top priority, personalized approach

  They look like your best customers. Prepare an opener about their industry and role, check the number (mobile or direct line) and put them at the start of the call session.

• B Tier B: your standard cadence

  One criterion is missing or slightly off: a good target for a regular sequence, call then email. Read the reason: a B “one level above target” is often the right account with a contact who is too senior; ask who owns the topic.

• C Tier C: later, or another channel

  Not worth a call today. Keep them for an email campaign, content, or the day a buying signal moves them up.

• D Tier D: off-target

  You don’t have to delete them, but take them out of the campaign. If most of your list is D, the list is the problem, not the reps.

• ? Provisional tiers: complete them first

  A provisional A is a strong profile with the size, industry or country missing. Fill in those fields (CRM, company website, enrichment) before calling: the tier may drop.

Common mistakes

• Targeting everyone

  Ticking every function and every industry gives you A-tiers everywhere and sorts nothing. A good ICP excludes more than it includes.

• Treating missing data as a bad fit

  A row without a headcount is not a small company. That is why the tool scores on the data present and shows completeness separately.

• Forgetting the knock-outs

  A well-scored competitor or current customer ends up in the sequence. Paste their domains once and save the profile for next time.

• Taking job titles at face value

  “Head of” means director in some companies and team lead in others; “Partner” is an owner at a law firm and a channel role at a software vendor. Classify the ambiguous titles of your market in the panel: the rule then applies to every file.

• Scoring once and for all

  People change jobs and companies grow: a six-month-old list deserves a cleanup and a fresh score before the next campaign.

• Never checking the model

  If your meetings come from C-tier rows as often as from A-tier ones, the weights do not reflect reality. Adjust them, then rescore.

Keep going

Tools and guides for the steps before and after the score.

• Direct Dial vs Switchboard Checker Mobile, direct line or switchboard: how reachable each contact is, and the call order.
• Contact CSV Cleaner Delimiter, accents, columns, duplicates: repair the export before scoring it.
• Waterfall Enrichment Merger Several providers, one file, the best number per contact.
• B2B Data Decay Calculator How many rows of your list have probably changed since the export.
• BANT & MEDDIC Qualification Scorecard On the call: budget, authority, need, timing, scored live.
• From buying signals to today’s call list The signals that justify a call now, and how to combine them with fit.
• B2B lead lists for cold calling Where calling data comes from, what to check before buying, how to score a list.
• The complete lead qualification guide Frameworks, process and scorecards for B2B qualification.