Ranking Accounts by Intent: A Practical Scoring Framework

This piece picks up after the score already exists. A sibling article, "Intent Data Scoring: Building a Model That Doesn't Just Reward Website Visits," covers how the underlying 0-100 number gets built: signal weighting, decay, formulas. This one covers what a RevOps or sales leader does with that number once it lands in the CRM: how to cut it into tiers, route each tier to the right owner, re-rank on a schedule, and fit the whole thing inside the hours a rep actually has.

A scored account list with no tiers is a spreadsheet nobody opens twice. Marketing ops builds the model, the CRM populates a score field on 4,000 accounts, and three weeks later the AEs are still working the same 40 logos they always worked, picked by memory and habit, not by the number marketing spent a quarter building. The score didn't fail. The decision layer on top of it never got built.

TL;DR: Convert the continuous score into 3 to 4 named tiers (Hot, Warm, Cool, Cold), set the cutoff points by testing them against last year's actual win rates rather than guessing round numbers like 75 and 50, and attach a specific owner and response-time SLA to each tier (AE direct-dial, SDR cadence, marketing nurture). Re-rank daily for the Hot tier and weekly for everyone else, because Harvard Business Review's 2011 field study of web leads found firms that made contact within an hour were roughly 7 times more likely to qualify the lead than firms that waited even one hour longer, and more than 60 times more likely than firms that waited 24 hours (Oldroyd, McElheran, and Elkington, "The Short Life of Online Sales Leads," Harvard Business Review, March 2011). Finally, cap each rep's Hot-tier queue at a number they can actually call in a week; a ranked list that exceeds rep capacity just produces a longer list of accounts nobody touches.

A Score Is Not a Decision

An intent score answers one question: is this account showing more buying signal than that one. It does not answer who should call them, how fast, or what happens if two accounts both land at 76. Those are routing questions, and they sit downstream of the model entirely.

Most teams skip this layer because it feels like the easy part after the hard part of building the score. It isn't. A model that's 90% accurate but feeds a flat, unranked list into Salesforce produces the same chaos as a model that's 60% accurate, because the rep still has to decide, unaided, which of 200 "scored" accounts to call first. Ranking converts a diagnostic number into an operating rule.

Setting Tier Thresholds

Four tiers work for most B2B SaaS GTM motions: Hot, Warm, Cool, and Cold. Three tiers work for smaller teams with one combined SDR/AE function. Five or more tiers almost always collapses back into four within two quarters, because nobody can hold five distinct playbooks in their head during a live call.

The starting cut points are a hypothesis, not a conclusion. A common first pass looks like this: Hot at 80 and above, Warm at 55 to 79, Cool at 30 to 54, Cold below 30. These numbers come from nowhere in particular, which is exactly the problem: a threshold set before you've looked at a single closed deal is a guess wearing a lab coat. The real cutoffs get set in the next section, against your own pipeline.

Before assigning any tier, an account has to already be on the list in the first place; that starting universe is the target account list purple path walks through in how to build a TAM list, and a tiering exercise applied to the wrong universe just ranks accounts you were never going to sell to.

The Tiering Table

The table below is a working template. Swap the score bands for whatever your own calibration (next section) produces; keep the owner and SLA columns as the part that actually changes rep behavior.

A Hot account that doesn't get a human response inside roughly an hour starts losing the advantage the score gave it. InsideSales.com's 2014 Lead Response Report, built on tracked response data across hundreds of B2B companies, found that most inbound leads still waited hours for a first response despite years of industry data showing the opposite was needed. Ken Krogue, the company's co-founder, wrote about the same gap in Forbes under the headline "Why Companies Waste 71% Of Internet Leads" (July 2012), citing internal tracking showing 71% of leads never got a timely follow-up at all. A tiering system that assigns a Hot label but routes it into the same unprioritized queue as everything else hasn't solved the problem; it's renamed it.

Calibrating Thresholds Against Pipeline and Win-Rate Data

Pull every closed-won and closed-lost opportunity from the last two to four quarters and back-fill what each account's intent score would have been at the point it entered the pipeline, if your tooling retains historical scores (most intent platforms, including Bombora and 6sense, keep a score history per account for exactly this reason). Then plot win rate against score band. The real cutoff for "Hot" is the score above which win rate jumps noticeably, not an arbitrary round number.

In practice this produces uneven bands. A team might find win rate sits flat at 12-15% from score 40 to 70, then jumps to 38% at 71 and above; in that case, "Hot" starts at 71, not 80, and everything below it is relabeled. Recalibrate this quarterly for the first year of a new model and twice a year after that; a threshold set once at launch and never revisited drifts out of sync with the market as deal sizes, buying committees, and even the intent vendor's own signal mix change underneath it.

This is also where firmographic fit earns its place next to behavioral intent, not as a separate score but as a gate. purple path's comparison of firmographic versus behavioral ICP signals makes the case that fit should filter the universe before intent ranks it; applied here, that means a 90-score account outside your ICP never enters the tiering exercise at all, no matter how hot the behavioral signal looks.

Routing Rules: Matching Tier to Owner

The tiering table above assigns an owner to each band, but the handoff mechanics matter more than the label. A Hot account should create a task or alert inside the CRM the moment it crosses the threshold, not sit as a passive field value an AE has to remember to check. purple path's guide to integrating intent data into CRM and automation walks through building that trigger layer in HubSpot or Salesforce so a tier change fires a Slack alert or task, not just a dashboard update nobody opens.

Warm accounts go to a structured SDR cadence, typically 6 to 10 touches across email, phone, and LinkedIn over two to three weeks, with an AE pulled in only once a meeting is booked. Cool accounts either get a lighter SDR touch (2 to 3 attempts, then release back to nurture) or get folded into paid and content retargeting rather than human outreach at all. Cold accounts get no individual rep attention; they sit inside whatever ongoing demand generation program is running, which is where purple path's framework for designing high-impact demand generation programs applies directly, since that's the mechanism meant to eventually move a Cold account into Cool on its own.

Re-Ranking Cadence

Score decay means yesterday's Hot account can be today's Warm account, and a tiering system that only recalculates monthly will have AEs calling accounts whose buying window already closed. Daily recalculation for the Hot tier is the right default for any team with real-time or daily-batch intent data; weekly is acceptable for Warm and Cool; monthly is enough for Cold, since nothing routes to a human from that tier anyway.

Re-ranking is not the same exercise as recalibrating thresholds. Re-ranking runs the existing formula against fresh signal data every day or week; recalibration (the previous section) revisits where the threshold lines sit, and happens quarterly at most. Conflating the two is a common failure mode: a team that only touches its tiering logic once a quarter will also only move accounts between tiers once a quarter, which defeats the purpose of having near-real-time intent data in the first place.

Handling Ties and Borderline Accounts

Two accounts at score 79 and 81 sitting on opposite sides of the Hot/Warm line will behave almost identically in reality; the model has more precision than the underlying signal deserves. Build a buffer zone, typically 3 to 5 points on either side of a cutoff, where a secondary factor decides placement rather than the raw score alone.

The most useful secondary factor is deal size times ICP fit, not a second intent signal, since stacking more intent data on an already-borderline score just adds noise. An account at score 78 with a $120,000 ACV potential and a strong firmographic match should round up into Hot; an account at score 82 that's a poor fit or a small deal should round down. This is a judgment call made once per recalibration cycle, documented as a standing rule (for example, "borderline accounts within 5 points of a cutoff round up if ACV potential exceeds $50K and ICP fit score exceeds 80"), not re-litigated account by account.

Ties within the same tier, where ranking order inside the Hot list itself matters for queue sequencing, get broken by recency of the triggering signal: the account whose score crossed the Hot threshold most recently goes first, on the logic that fresher intent correlates with a shorter window before the buying committee moves on without you.

Capacity Is the Real Constraint

A ranking system that ignores rep capacity just produces a longer list of accounts nobody works. If a model tags 340 accounts as Hot and one AE owns the territory, 300 of those accounts will sit untouched no matter how accurate the score was.

Work backward from capacity instead of forward from the score. An AE with a 40-hour week, after internal meetings, admin, and existing pipeline management, typically has room for meaningful first-touch outreach on 15 to 25 net-new or re-engaged accounts per week, depending on deal complexity and average sales cycle length. That number sets the real size of the Hot queue each AE can carry, regardless of how many accounts the model scores above 80.

The Bridge Group's 2025 SDR Models, Motions & Metrics Report, one of the longest-running benchmarking studies of sales development organizations in SaaS, tracks exactly this kind of capacity data across hundreds of companies each year and is worth pulling before setting your own caps rather than guessing them from one team's experience. When a model produces more Hot accounts than available capacity, the answer isn't to work faster; it's to build a visible, ranked backlog so the 301st account isn't simply lost, and to feed the overflow into the Warm cadence instead of leaving it stranded in an oversized Hot list.

Frequently Asked Questions

How many tiers should we actually use?

Four (Hot, Warm, Cool, Cold) works for most B2B SaaS teams with separate SDR and AE functions. Three works if one role handles both prospecting and closing. More than four tiers tends to collapse back down within two quarters because reps can't hold that many distinct playbooks in working memory during a live call.

Should every account have a tier, or only accounts above a minimum score?

Every account in your target account list should carry a tier, including Cold, because the tier is what tells the system an account gets automated nurture rather than silence. An account with no tier at all is an account nobody has made a decision about, which is a worse state than correctly assigning it to Cold.

What happens when an account moves from Warm to Hot mid-cadence?

The SDR cadence should stop and hand off to the AE inside the same response window as any other new Hot account, not finish out its remaining steps first. Build this as an automated trigger in the CRM rather than relying on the SDR to notice the score change manually.

How often should we revisit our tier thresholds?

Recalibrate the cutoff scores against closed-won and closed-lost win rate quarterly in year one, then twice a year afterward. This is separate from re-ranking, which moves individual accounts between existing tiers daily or weekly; recalibration only touches where the tier lines themselves sit.

What if our AE team doesn't have capacity for all the Hot accounts the model produces?

Cap the Hot queue at what an AE can realistically work in a week (15 to 25 accounts is a reasonable starting range) and route the overflow into a visible, ranked backlog rather than an oversized list that gets ignored. The backlog should promote accounts into the active queue as capacity frees up, in score order.

Putting It to Work

A tiering system is only as good as the handoffs it automates. If your CRM still relies on a rep remembering to check a score field, the ranking exists on paper and nowhere else. purple path works with B2B SaaS revenue teams on exactly this layer, building out the routing, cadence, and capacity rules that turn a scored account list into something reps actually run on. If your intent data is sitting in a dashboard instead of a queue, get in touch with purple path and we'll help you build the operating system around it.