ICP scoring is the practice of rating each lead or account against your ideal customer profile, the set of traits that describe the companies and buyers who get the most value from your product and are most likely to close. Instead of treating every lead the same way, a scoring model ranks them, so a rep can work the strongest fits first instead of working a list in whatever order it arrived.

Why it matters is straightforward. Sales time is limited, and not every lead deserves the same amount of it. A team with fifty leads and no scoring model tends to work them roughly in order of arrival or however loudly a prospect asks for attention. A team with a working score can see, at a glance, which ten accounts look the most like their best customers and start there.

There are two broad approaches to scoring, and most mature programs end up combining them.

• Demographic scoring looks at fixed attributes: company size, industry, location, and the title of the person you are talking to. These traits answer the question “does this account look like our customer, in general.” They are stable and easy to collect, but they do not tell you anything about timing. A company can be a perfect demographic fit and have zero interest in buying anything this quarter.
• Signal based scoring looks at recent events: funding, hiring, executive changes, tech stack moves, and similar activity. These answer a different question, “is something happening at this account right now that makes it a good time to reach out.” Signal based scoring adds the timing dimension that demographic scoring is missing.

A useful way to think about it: demographic fit tells you whether an account belongs on your list at all. Signals tell you when to actually call.

A workable ICP scoring model usually draws on five categories of input. None of these need to be complicated to start; the goal is consistency, not precision on day one.

• Company size. Employee count or revenue range, matched against the size of company where your product has historically closed and retained well.
• Industry. The verticals where your product solves a problem people already recognize, versus industries where you would need to do more education before a conversation goes anywhere.
• Tech stack. Tools already in place that suggest compatibility or a gap your product fills, and tools in place that suggest a poor fit or an existing competitor relationship.
• Geography. Regions where you can actually support and sell effectively, including time zone overlap for sales conversations and any regulatory or localization constraints.
• Persona fit. Whether the specific person you have a contact for holds a title and role with the authority or influence to move a purchase forward.

Most teams start by weighting these based on their existing closed won customers: look at the accounts that became your best customers, find the traits they share, and weight the model toward those traits. From there, the model gets refined over time as more deals close or fall through.

A demographic only model has a real limitation: it does not change. A five hundred person software company in your target industry scores the same in January as it does in December, regardless of anything happening inside that business. That is a static snapshot, not a reflection of current buying likelihood.

Adding signals makes the score move with reality. A company that just raised a funding round, just hired ten new salespeople, or just brought on a new VP of operations should score noticeably higher than an otherwise identical company that has not changed in a year, even if their demographic profile is the same. The first company has active reasons to be evaluating new tools right now. The second one might be a fine fit eventually, but there is no evidence of urgency today.

Many scoring models, including Shaku AI’s, express fit as a single number from 0 to 100 so a rep can scan a list and know where to start without reading every field. In plain language, the ranges tend to break down like this:

• 80 to 100. A strong demographic match with active signals in play. These are the accounts to work first.
• 50 to 79. A reasonable fit, but either the demographic match is partial or there is no strong recent signal. Worth working, but not before the top tier.
• Below 50. Either a weak demographic fit, no current signal activity, or both. Usually not worth manual outreach time unless volume allows for it.

The exact thresholds should be tuned to your own close rates over time. The point of the number is not precision to the decimal, it is giving a rep a fast, consistent way to triage a list instead of reading every account from scratch.

Shaku AI scores every lead it surfaces against your ideal customer profile on a 0 to 100 scale, combining demographic fit with live buying signals so the score reflects both who the account is and what is happening there right now. Each lead arrives with plain language notes explaining why now is a good time to reach out and how to pitch, so a rep does not have to reconstruct the context behind the number before making a call.

If you want to see how your own ideal customer profile translates into a working score, Shaku AI offers a free trial with full access and no credit card required.