Every company has an ICP document. Almost none of them have an ICP that Sales accepts, Marketing scores against, and the CRM enforces. We build the segmentation model from your won, lost, and churned data, then ship it as scoring rules, lifecycle logic, playbooks, and page architecture.
Trusted by revenue teams at
years building revenue systems
B2B GTM implementations
Strategy + execution under one roof
Marketing scores leads against a workshop output from 2023. Sales disqualifies half of them on the first call using criteria that live in a rep's head. Both are right. Neither has been written into a system. So every pipeline review turns into a lead-quality argument nobody can settle with data.
MQLs are up and acceptance is down while cycle length is climbing. Nobody has looked at whether the accounts you're pouring budget into resemble the ones you actually win, because the analysis would require joining three systems and nobody owns that.
There's no Tier field. No scoring model. No disqualification rule. No lifecycle logic. So the ICP has zero operational consequence, it doesn't route a lead differently, doesn't change a sequence, doesn't stop a rep working a bad account for six weeks.
It's the segment your strategy has always named, so nobody re-tests it. It has the longest cycle, the lowest ACV, and the highest churn. The segment that actually pays back has never made it into a campaign or a CRM property.
Most $30M–$500M ARR companies live at Stage 2 and believe they're at Stage 3. They have the document. The CRM doesn't know about it.
Tier 1, Tier 2, expansion, and experimental, defined by what your won, lost, and churned cohorts actually have in common. Firmographics, technographics, behavioral signals, and trigger events. Not “companies with 200–1,000 employees” because that's what the template said.
Fit score and intent score as properties in HubSpot or Salesforce. Applied to leads and accounts. Visible to every rep before they pick up the phone. Routing, SLAs, and sequences that actually differ by tier.
The accounts you should refuse. Explicit, agreed by Sales and Marketing, and enforced in the system, so nobody spends six weeks on an account that was never going to buy.
Who initiates, who evaluates, who blocks, who signs, and how it differs between a Tier 1 fintech and a Tier 2 professional services firm. Because “the persona” is not one person and doesn't behave the same way across segments.
Each tier mapped to a motion, inbound, outbound, ABM, partner, PLG, with the sequences, offers, and sales plays that fit it. Tier 1 gets a different treatment than Tier 3, because it should.
Segment-aware site architecture, landing pages, and conversion paths, so a Tier 1 fintech buyer doesn't land on a page written for everyone. Built, not just recommended.
Governed, not frozen: we build the model, ship it into your systems, and set the cadence for re-validating it as new data arrives.
Before any data, we settle a question most teams have never explicitly answered: ideal by what measure? Highest ACV? Fastest cycle? Best retention? Highest expansion? Best gross margin? These often point at different segments, and if leadership hasn't chosen, no model can. Half of all ICP arguments are actually disagreements about the objective, not the data. We surface that in week one.
Six work streams. Every deliverable ships as a field, a workflow, a page, or a play, not a recommendation.
The objective the model optimizes for, and who owns it after we leave.
The evidence layer. Won, lost, churned, expanded, profiled, not asserted.
The model itself. Weighted, tiered, tested against holdout cohorts.
Who's in the room, what triggers them, what kills the deal, per tier.
The model becomes infrastructure, built, tested, documented in your CRM.
The model becomes go-to-market. Plays, messaging, and segment-aware pages.
Download the ICP readiness checklist. Twelve questions. If you can answer all twelve with a CRM field rather than an opinion, you don't need this engagement.
Tiered, weighted, validated against your won/lost and churn data, with disqualification criteria and the assumptions written down.
Properties, scoring workflows, routing rules, and lifecycle logic built and tested in HubSpot or Salesforce. Not a spec. Shipped.
Who decides, what triggers them, what kills the deal, for each tier, grounded in real won/lost interviews.
Plays for SDRs, AEs, and CSMs. Messaging guidance by tier. Enablement sessions so the model survives contact with the team.
Landing pages and site architecture that speak to the tier that landed on them, recommended, spec'd, and (optionally) built.
The quarterly process for re-validating tiers against new data, so the model doesn't rot the way the last one did.
No 40-slide ICP deck. Everything lands as a field, a workflow, a page, or a play.
Patchstack is the leader in open-source vulnerability intelligence, protecting thousands of sites and applications worldwide. Their HubSpot instance had never been configured: no scoring, no pipeline visibility, and a product catalog living entirely outside the CRM, so no deal could be attributed to what was actually sold. We audited and cleaned the properties, rebuilt the pipeline around their real sales process, synced the full product library so every deal carries its plan and renewal terms, migrated lead scoring onto weighted conversion and renewal-risk signals, and shipped forecasting and board-level dashboards. Three months in, the team runs it without us.
Growth-stage and mid-enterprise B2B companies, $30M–$500M ARR, with at least 12–18 months of deal and retention data. SaaS, technology services, professional services, fintech, healthtech, energy/sustainability, legal, and PE portfolio companies. Especially those whose win rates vary wildly by segment and who can't explain why.
We'd defined our ICP twice before. The difference this time was that it ended up as a field in HubSpot with a routing rule attached. Suddenly Sales and Marketing stopped arguing about lead quality, because we were both looking at the same score.
A workshop produces consensus, not evidence. Everyone leaves agreeing on an ICP that reflects what the room believes, usually the founding segment, because that's who everyone remembers. Then nobody checks it against the churn data. We start with the data and use the workshop to interpret it, not invent it.
If you have someone who can join CRM, product, and billing data, run cohort analysis, build a weighted scoring model, validate it on holdout, and ship it as HubSpot workflows, do it. Most $30–500M ARR teams have one of those five capabilities, and that person is drowning in reporting requests.
They'll produce a beautiful segmentation model. Then activation is your problem, and activation is the entire point. An ICP that isn't enforced in the CRM has no operational consequence, which means it has no revenue consequence.
Then this is a validation engagement. We'll test your current scoring against actual won/lost outcomes. If it predicts, you'll have proof. In our experience most scoring models were built on plausible attributes rather than predictive ones, and the difference shows up in acceptance rate.
A 45-minute working session. Bring your current ICP and one number: your MQL→SQL acceptance rate. We'll tell you whether the definition is the problem or the enforcement is.
The questions RevOps, Marketing, and Sales leaders ask us on the first call, answered directly.
Having an ICP means you've written down who you sell to. Segmentation means you've tiered them by fit, scored them in the CRM, routed them differently, and can show the win rate gap between tiers. The first is an opinion. The second changes what happens on Monday.
Three signals: your teams argue about lead quality, your win rates vary wildly by segment without explanation, or you're entering a new market with no basis for focus. You'll also want 12–18 months of deal and retention data, below that, the cohorts are too thin.
Yes, and we consider it the deliverable, not an add-on. Properties, fit and intent scoring workflows, routing and SLA rules by tier, lifecycle logic, list segmentation, and dashboards. Built, tested, documented, handed over.
8–12 weeks, depending on data complexity, number of segments, and how many teams are involved. Segment payback findings typically land in week 3–4 and are actionable immediately.
Market research looks outward across the whole market, which segments exist, what they're worth, what competitors are doing. ICP segmentation looks inward at your own data to decide which of them you should actually pursue, and then enforces that decision in your systems. Many clients run research first. Neither strictly requires the other.
It includes messaging guidance per tier and buying committee mapping. Full persona-pain-message architecture is a separate engagement in our GTM pillar, this one produces the segment definitions it depends on.
Segmentation surfaces where pricing and packaging are misaligned with segment value. Acting on that, offer segmentation and price optimization, is our Pricing & Packaging engagement.
CRM access (ideally read/write in a sandbox), marketing automation, product usage and billing data, closed-won/lost and churn exports, and introductions for 6–10 interviews across reps, customers, and lost deals.
Then you've found the most expensive assumption in the business, and you found it from a model instead of a bad year.
