By Wednesday, the SDR floor was quiet—ten “Tier 1” accounts had been called to death, and the queue was empty.

We’d spent Monday and Tuesday burning through our best guesses. Calls, emails, LinkedIn touches, ad impressions—bandwidth gone. Not because the market was dry, but because our ICP was carved into a slide once and left there to fossilize.

And yet, timing moves. Hiring moves. Budgets move. Even your own brand’s reach moves. So why was our ICP frozen?

This is how we unfroze it—and how you can, too.

The Slide-Deck ICP That Ate Your Bandwidth

Most teams treat ICP like a buyer persona with better punctuation. Job titles, industries, company size. Neat for a board meeting; useless for a Tuesday morning queue.

That slide does three things badly:

First, it overweights firmographics and underweights timing. Fundraises, new leaders, reshuffles—none of that shows up on a static slide.

Second, it ignores how prospects react to you. Opens, clicks, replies, visits, forms—engagement decays by the hour, and your “ideal” can go cold by lunchtime.

Third, it never links to channel bandwidth. You have finite dials, sends, and impressions each day. If your ICP can’t shape those choices, it’s not an ICP. It’s trivia.

Here’s the rub: an ICP isn’t a portrait. An ICP is a scoring system. Scores roll up into ICP tiers. And the whole thing must be dynamic—because the market is.

Signals, Not Stereotypes

At Side Kick, we rebuilt ICP around decisions machines can take in seconds and humans can trust at a glance. The goal: fill human revenue bandwidth daily with the right accounts, not just the “right-looking” accounts.

What changed:

We tied ICP to live, external signals and fast internal reactions. We pushed it into the tools that actually move work—HubSpot, Smartlead, Apollo, ad platforms—so the model shapes queues, not slides.

We also forced weekly improvement. Scores are living. Tiers are living. The rules are simple, the outcomes visible, and the adjustments small but frequent.

The mental shift is simple: ICP is not “who we sell to.” ICP is “how we prioritize now.”

The Five That Move the Needle

Our dynamic ICP scoring model uses five signals. Four you can observe. One you must earn.

Fitment. Who they are and how they sell. B2B, ~50–1000 employees, actively doing outbound. No internal ops team is a strong fit signal. We source Fitment from Apollo, LinkedIn, and the website. We separate it from Infrastructure on purpose.

Intent. Timing clues that whisper “now.” Funding rounds. A fresh junior SDR class. A new senior GTM hire. An executive customer changing companies. ABM readiness shows up in plans and posts. Job boards and LinkedIn are gold here.

Engagement. How they react to you. Opens, clicks, replies, site visits, forms, ad interaction across LinkedIn and display. Engagement decays faster than Intent—it’s a fruit with a short shelf life. Why that matters and how to score it in Score Decay,

Infrastructure. Tech readiness and acquisition investment. ZoomInfo, 6sense, HubSpot Marketing, Marketo, LinkedIn Ads, plus outbound tools like Apollo. We read this via BuiltWith, public stacks, and AI over job posts.

Pain (from conversations only). Public pain is mostly theater. Real pain is in calls and transcripts. We use AI to score it. When someone says, “Our reps spend two hours a day chasing bad leads,” that’s a 10/10 signal. Pain is the strongest driver of action.

Our weighting reflects the world as it is, not as decks prefer it:

• Fitment 20%

• Intent 20%

• Engagement 20%

• Infrastructure 10%

• Pain 30%

These weights force trade-offs early. If Pain is high, it can carry a merely “okay” stack. If Engagement is spiking, it can tip a borderline Fitment into today’s queue.

Buckets for Humans, Scores for Machines

Scores decide actions. Tiers help humans talk about those actions.

We keep it simple:

• 80–100 → Tier 1

• 50–79 → Tier 2

• Below 50 → Tier 3

Tiering exists for humans and reporting. The work stays score-based. Two Tier 2s with different reasons (high Intent vs. high Engagement) won’t get the same play. The machine routes them differently. Your team sees a clean label; your system executes a precise plan.

Why dynamic tiers? Because you aren’t optimizing a spreadsheet. You’re filling daily bandwidth. We wanted the calling queue to refresh with the best available right now. We wanted automations to decide who gets dials, who gets emails, and who gets warmed by ads—without manual tier babysitting.

So tiers are just score buckets, recalculated as signals move.

Bandwidth Is a Budget (A Week in the Queue)

On Monday, a client started with ten Tier 1 accounts. Great logos, clean Fitment, steady Intent. Their three SDRs could eat that list by lunch—and they did.

By Tuesday afternoon, silence. They “ran out of ICP.”

The old way would wait for next week’s research day. The dynamic ICP way lowered the gate.

We nudged the threshold so that 75–79 (previously Tier 2) with spiking Engagement slid into the calling queue. A webinar click. A second reply thread. A repeat visit to the pricing page. “Not perfect,” you might say. But perfect isn’t the job. Filled, prioritized bandwidth is the job.

By Wednesday morning, the queue was full again. The SDRs weren’t guessing. They were calling. The ad team had a matched list for warm retargeting. Marketing saw which content actually pushed scores up. Sales leadership saw capacity stay at 100%, not 62%.

This is why dynamic ICP matters. It converts “we know our market” into “we know what to do today.”

The Plumbing That Makes It Real

Data comes from places your team already uses:

Fitment from Apollo, LinkedIn, and your site.
Intent from LinkedIn, job boards, and leadership changes.
Engagement from Google Analytics, LinkedIn Ads and email sequencer (opens, clicks, replies).
Infrastructure from BuiltWith and job posts.
Pain from call transcripts—AI-scored snippets that flag real friction.

You don’t need a PhD to wire this. You need clean fields, clear weights, and the discipline to review weekly.

The Gotchas (And How We Learned Them)

We tried scoring “public pain” from posts and PR. It looked scientific and performed terribly. We dropped it. Conversation-only, or don’t score pain.

We once kept tiers static for a quarter to “stabilize reporting.” Calling productivity fell. Marketing flighted the wrong segments. Lesson learned: reporting should adapt to reality, not the other way around.

We also bundled Fitment and Infrastructure at first. Big mistake. Great stacks with no outbound motion clogged Tier 1. Separating the two fixed it.

Small, honest wins beat broad, theoretical ones. Every time.

What To Do Next (Takes One Afternoon)

List your five signals—Fitment, Intent, Engagement, Infrastructure, Pain—and give them the weights above as a starting point. Map score ranges to three tiers (80–100, 50–79, <50). Tie each tier to real actions: calls first, emails second, ads for warm-up. Plug your sources (Apollo, LinkedIn, job boards, HubSpot, Smartlead, BuiltWith, transcripts). Then, every Friday, review deltas and nudge thresholds.

Within a week, your “ICP” will stop being a slide and start being a system.

What Not To Do (Because You Will Be Tempted)

• Don’t stop at industry and size and call it ICP.

• Don’t freeze tiers for months to make a neat chart.

• Don’t ignore how prospects engage with your content.

• Don’t pretend public pain equals real pain.

• Don’t do ICP work in a vacuum—automation is the point.

The Quiet Outcome

When your ICP becomes a dynamic scoring model, Tier 1 stops being a trophy case and starts being a pipeline engine. Marketing, sales, and revenue ops finally share the same reality. ABM readiness isn’t theoretical; it shows up in scores. SDRs call more, guess less. Leaders see capacity stay full without adding heads.

We didn’t invent this in a lab. We built it for client work, broke it a few times, and kept what held under pressure.

This piece is the first in a series. We’ll go deeper on each signal (especially Pain), on bandwidth math, and on the mechanics of ICP tiers inside your tools. If you want a calm, ten-minute audit of your current ICP doc, we can do that—no slides, just truth.