sales-account-segmentation

{
  "skill_name": "sales-account-segmentation",
  "evals": [
    {
      "id": 1,
      "prompt": "I run RevOps at Trellwood Analytics, a mid-market SaaS doing about $38K ACV. We closed 220 deals in the last twelve months and everything is in the CRM. We just signed a 12-month contract with an intent data vendor and the signal is finally flowing. I want one clean 0-100 account score my team can sort a list by, and I want intent weighted at 50% of it because it's the freshest data we have - firmographics barely change. Give me the dimensions and the weights.",
      "expected_output": "A two-axis model: a slow-moving fit score and a separate decaying readiness score, with intent pushed down to a defensible share of fit, weights calibrated against the 220 closed-won deals, a deduction layer, published sub-scores, and quadrant plays. No tier cutoffs.",
      "files": [],
      "expectations": [
        "Refuses to deliver fit and readiness as a single blended 0-100 number and delivers two separate axes instead",
        "States that intent and behavioral signals belong on a second, continuously updating axis while fit anchors and changes slowly",
        "Rejects or materially reduces the proposed 50% intent weight and names the directional splits (firmographic ~30-40%, technographic ~20-25%, behavioral/intent ~25-40%)",
        "Names overweighting intent because clicks are easy to count as the most common weighting error, and states that fit is the quieter but more reliable predictor",
        "If a single routing number is unavoidable, recommends multiplying fit by readiness rather than averaging them, and keeps both sub-scores visible on the account record",
        "Calibrates the weights backward from the 220 closed-won deals rather than from intuition or the user's stated preference",
        "Names the model health test: high-scoring accounts must convert at a materially better rate than low-scoring ones before anything routes on the score",
        "Includes an explicit negative-signal deduction layer alongside the positive points",
        "Publishes per-dimension sub-scores so reps can see why an account scored as it did",
        "Names at least two quadrant plays showing that a high-fit/low-readiness account and a low-fit/high-readiness account get different treatment"
      ]
    },
    {
      "id": 2,
      "prompt": "VP Sales at Kesslon Robotics here. We carve territories for next fiscal year in five weeks and I need an account segmentation model before then. We don't have a written ICP document but honestly everyone here knows it: discrete manufacturers, 500+ employees, North America, and they need to already run some kind of MES. We have about 180 closed-won deals in the CRM from the last year. Just build me the model - segments, scores, and how many accounts each rep ends up with.",
      "expected_output": "Routes the ICP work upstream first, then scopes a default-rung model to the five-week deadline with closed-won calibration held as non-optional, presents candidate model shapes for approval, and hands the rep-count and carve-execution questions to the sibling skills.",
      "files": [],
      "expectations": [
        "Routes the ICP work to mbfinotti/sales-skills@sales-icp-definition before building the segmentation model, treating an informal 'everyone knows it' profile as not an agreed ICP",
        "States that segmenting along dimensions nobody signed off on quietly re-derives the ICP badly",
        "Consumes the ICP's criteria and disqualifiers as the score's foundation rather than re-deriving them inside the segmentation work",
        "Holds the model at the default firmographic + technographic rung because of the five-week deadline, and says explicitly that the deadline is what moved the choice",
        "Treats calibration against the last 12 months of closed-won deals as non-optional even under the deadline",
        "Asks clarifying questions one at a time before proposing a model, rather than returning a finished model immediately",
        "Presents 2-3 candidate segmentation model shapes with trade-offs and one explicit recommendation before weighting anything",
        "Refers the per-rep account counts and tier cutoffs to mbfinotti/sales-skills@sales-account-tiering instead of setting them",
        "Refers execution of the territory carve to mbfinotti/sales-skills@sales-org-structure",
        "States that territory carving balances on weighted opportunity (account count x average ACV x estimated win rate), never on raw account count"
      ]
    },
    {
      "id": 3,
      "prompt": "I'm head of sales ops at Ambervale Systems. We sell a compliance platform into utilities - our whole addressable market is about 180 nameable accounts in two countries, and we have four years of CRM history. Our CEO read a case study where a company built a machine-learning propensity score on 70+ fields and got 2x new-customer conversion and 90% higher opportunity open rates. He wants us to build the same thing and he wants me to put those numbers in the board deck as our target. What does the build look like and how long until we hit those numbers?",
      "expected_output": "Declines the ML propensity build at 180 accounts, names the scale conditions that would justify it, substitutes closed-won pattern analysis plus JTBD segmentation on the default rung, and grades the quoted case-study figures as self-reported vendor marketing unfit as board targets.",
      "files": [],
      "expectations": [
        "Recommends against building a predictive/machine-learning propensity model at 180 accounts",
        "Names the conditions that would justify predictive scoring: a TAM running to thousands of accounts, deep CRM history, and a named owner for the model",
        "States that predictive scoring degrades when TAM is a few hundred accounts or the vertical has thin intent signal",
        "Offers closed-won pattern analysis plus manual JTBD/use-case segmentation as the more reliable substitute at this scale",
        "States that predictive scoring loses on efficiency (value returned per unit of effort) despite having the highest ceiling, rather than presenting it as the best available option",
        "Recommends the firmographic + technographic default rung, calibrated on closed-won deals",
        "Labels the quoted 2x conversion and 90% open-rate figures as vendor-published, self-reported, single-customer marketing",
        "Refuses to present those figures as expected results for Ambervale, treating them at most as existence proofs that segmentation done well can move such numbers",
        "Declines to commit a specific conversion-lift number or date to the board deck",
        "States the signal-layer efficiency ordering explicitly rather than leaving it implied by the order options are listed in",
        "Names the small nameable-account universe as the reason the default ordering is re-ranked here, and says what moved"
      ]
    },
    {
      "id": 4,
      "prompt": "Marchetti Industrial - we make four product lines for heavy industry and our top 30 customers are about 60% of revenue. Our CRO keeps asking for cross-sell numbers in the QBR and what he gets back is stories: 'the Finance team over there seemed interested last year', that kind of thing. Each of our six strategic AEs keeps their own spreadsheet of open opportunities and refreshes it whenever they remember. Our biggest customer just acquired a rail subsidiary we've never sold anything into. How do I turn this into something I can actually manage?",
      "expected_output": "A whitespace mapping discipline: products against divisions and buying centers as a three-state heatmap per strategic account, each gap resolved into one of four plays and tracked as a staged pipeline opportunity, kept on the CRM record, refreshed quarterly and on trigger events.",
      "files": [],
      "expectations": [
        "Maps products/services against each strategic account's departments, divisions and buying centers",
        "Visualizes the map as a heatmap with distinct states for active footprint, open expansion opportunity, and competitor-entrenched",
        "Resolves each gap into one of four plays: seat/usage expansion, tier upsell, cross-sell of an adjacent product, or landing a new buying center",
        "States that landing a new buying center requires explicit org-chart and buying-committee mapping",
        "Tracks each identified gap as a pipeline-staged opportunity (identified, qualified, in discussion, proposal, won/lost) rather than a static list",
        "Sets refresh at least quarterly plus immediate refresh on trigger events, naming an acquisition, a leadership change, a renewal, and a competitor entering the account",
        "Requires the map to live on the CRM account record rather than in rep-owned spreadsheets, on adoption grounds",
        "Names the eyeballed, from-memory account plan as the documented failure mode, with QBR anecdotes instead of cross-sell numbers as its symptom",
        "Distinguishes whitespace mapping inside won accounts from net-new fit scoring as two different disciplines that share a model",
        "Frames the expansion motion as growing the customer's business rather than conquering territory",
        "Grades any expansion-payoff figure it cites (such as 3-5x whitespace revenue potential, 5-25x acquisition versus expansion cost, or 115-130% NRR against a ~100-105% median) as directional, practitioner-estimated or HBR-cited rather than as a promised outcome"
      ]
    },
    {
      "id": 5,
      "prompt": "Lunaris Skin - we're direct-to-consumer skincare, 410,000 customers in the database, average order $62, no sales team at all, everything is Shopify plus email and paid social. Our board wants us to 'get serious about account segmentation like the B2B companies do'. So: build us a firmographic fit score for our customer accounts, and map the whitespace across each account's divisions. I'd like about 15 segments so the email team can be really targeted.",
      "expected_output": "Says plainly that D2C has no accounts, substitutes value-based customer segmentation (LTV/spend banded against engagement/recency as two axes), calibrated from purchase data, names which parts of the account machinery do not transfer, and cuts 15 segments to the sustainable ceiling.",
      "files": [],
      "expectations": [
        "States that direct-to-consumer has no accounts and declines to force the account machinery onto it",
        "Recommends value-based customer segmentation, banding customers by lifetime value or spend as the structural analog of the fit axis",
        "Crosses that value band with engagement/recency (RFM-style) as the readiness analog, keeping two separate axes",
        "Names firmographic layers as not transferable to this case",
        "Names buying-committee logic as not transferable to this case",
        "Names org-chart whitespace mapping as not transferable to this case, rather than building the requested division map",
        "Calibrates the bands from actual purchase data rather than from assumptions about who the good customers are",
        "Rejects the 15-segment request, naming a ceiling of roughly 6-8 segments and the reason: past that, some segments cannot sustain a distinct playbook and should be merged",
        "Specifies a refresh cadence for the bands",
        "Requires each resulting segment to map to a genuinely different treatment or playbook, or nothing has actually been segmented"
      ]
    },
    {
      "id": 6,
      "prompt": "Corvane Beverages. We're a mid-size drinks brand - we don't sell to end consumers directly, we sell into 11 grocery banners, 3 regional distributors and a couple of convenience franchise groups. Our national accounts team of five covers all of them and picks priorities by whoever shouted loudest at the last trade show. Everyone keeps telling me to build an 'account fit score' but every article I read talks about employee count and tech stacks, which means nothing for a supermarket chain. Can this even work for us?",
      "expected_output": "Keeps the segmentation machinery and swaps the fit inputs for retail-relevant ones, puts sell-through and reorder velocity on the readiness axis, redefines whitespace as unsold categories and unpenetrated banners, and keeps calibration, two axes and CRM encoding unchanged.",
      "files": [],
      "expectations": [
        "Keeps the account segmentation machinery and swaps the fit inputs rather than declaring the approach inapplicable to a retail-channel business",
        "Names store count, shelf and category position, geography, and banner affiliation as the fit inputs replacing firmographics and technographics",
        "Uses sell-through and reorder velocity in the readiness role",
        "Defines whitespace here as unsold categories and unpenetrated banners or regions inside a chain already stocked",
        "Keeps fit and readiness as two separate axes with different update cadences",
        "Keeps calibration against the last 12 months of actually won business",
        "Keeps CRM operationalization: scores written to the account record, routing, and a re-scoring cadence",
        "States that the resulting scores feed the same downstream tiering as a B2B model would",
        "States the B2C-through-key-accounts scope explicitly rather than leaving the framing ambiguous"
      ]
    },
    {
      "id": 7,
      "prompt": "Penhallow Software. Eight months ago a consultant built us a segmentation model - it's a very nice 40-slide deck and nobody has opened it since. Two problems I keep hitting: marketing's campaign lists treat 'enterprise' as 500+ employees while sales' routing rules use 1,000+, and our enrichment contract pulls fresh data once a year at renewal because that's cheapest. I want to finally put this live in the CRM this quarter. I'll report progress to the exec team as the number of accounts scoring above 70, and I want tier 1 capped at 3 accounts per rep.",
      "expected_output": "Encodes the model in the CRM with fields, routing SLAs, waterfall enrichment and threshold-triggered transitions; replaces the annual pull with decay-matched cadences; collapses the two enterprise definitions into one RevOps-owned contract; replaces the vanity metric; and hands the tier-1 cap to the tiering skill.",
      "files": [],
      "expectations": [
        "States that a model living in a slide deck segments nothing, and encodes it on the CRM account record",
        "Writes the fit score, readiness score, sub-scores and segment onto the account record via automated enrichment",
        "Routes each segment to its coverage lane with SLAs on the handoff",
        "Recommends waterfall enrichment, querying providers in sequence until a field fills, rather than relying on a single vendor's coverage",
        "Rejects the annual enrichment refresh, stating that technographic and intent data decay fast and an annual pull is stale by around month three",
        "Sets the cadence explicitly: full model review quarterly, weight recalibration once or twice a year, readiness signals updating continuously",
        "Moves accounts between segments on explicit thresholds such as an employee-count or ARR crossing, or a downgrade review after two quarters below the segment minimum, rather than on ad-hoc judgment",
        "Resolves the 500-versus-1,000 employee conflict into one definition set, encoded once, with joint sales and marketing sign-off",
        "Names RevOps as the single owner of the model",
        "Rejects 'number of accounts scoring above 70' as the success measure because it optimizes a score nobody trusts, and replaces it with pipeline and conversion movement",
        "Names the high-versus-low scorer conversion delta as the model's standing health test",
        "Refers the 3-accounts-per-rep tier 1 cap to mbfinotti/sales-skills@sales-account-tiering rather than setting capacity caps itself"
      ]
    },
    {
      "id": 8,
      "prompt": "Dravon, developer tooling. We have around 12,000 free workspaces, roughly 400 of them paying, and exactly 3 AEs. Marketing throws about 600 MQLs a month over the wall based on content downloads and webinar attendance, and the AEs cherry-pick from that while also emailing whoever signed up for a free workspace last week - so half the time two of us are in the same company at the same time. Build us an account scoring model so the team knows which MQLs to call first, and also settle the argument about whether we should be a PLG company or a sales-led one.",
      "expected_output": "Separates account-level fit scoring from contact-level engagement scoring and routes the MQL half out, defines the PQL unit as fit x usage with usage visible, sets an explicit PLG-to-sales handoff trigger, maps motions per segment, and refers the company-level motion choice to the sibling skill.",
      "files": [],
      "expectations": [
        "Distinguishes account-level structural fit scoring from contact-level engagement scoring and routes the MQL/contact half to mbfinotti/revops-skills@lead-scoring",
        "Defines the PLG scoring unit as the product-qualified lead: ICP-fit score combined with product-usage score",
        "Keeps the usage context visible to reps rather than delivering a black-box number",
        "Defines an explicit PLG-to-sales handoff trigger such as a usage threshold, a team size, or an enterprise domain, so the self-serve and sales engines stop working the same accounts",
        "States that already having product-usage data flowing pre-pays most of the fit x readiness rung's cost, and promotes that rung accordingly",
        "Says explicitly which piece of context moved the signal-layer ranking",
        "Refers the company-level PLG-versus-sales-led question to mbfinotti/sales-skills@sales-motion and reframes the in-scope question as which motion serves which segment",
        "Maps segments onto motions that actually exist at this company rather than proposing motions it does not run",
        "Keeps fit and readiness as separate axes"
      ]
    },
    {
      "id": 9,
      "prompt": "Halworth Group, B2B services software. We bought an enrichment subscription and pulled a list of 3,100 target accounts filtered on three things: SIC code, 100-2,000 employees, and US/Canada. That list has been the target list for three quarters. The reps completely ignore it - when I ask why, they say every row looks the same as every other row. My plan is to add five more firmographic filters to tighten it up and buy an ABM platform to push the list into. Sanity check me before I sign the contract.",
      "expected_output": "Diagnoses a firmographic-only, operationally inert list, refuses more firmographic filters as the fix, promotes to the technographic default rung with decay-matched refresh, applies the ~80%-of-TAM disqualification diagnostic, and routes upstream if the looseness is in the ICP.",
      "files": [],
      "expectations": [
        "Diagnoses the list as firmographic-only and operationally inert: every account looks equally good, so reps fall back to gut feel",
        "States that firmographic-only ranks below firmographic + technographic and below the fit x readiness rung on efficiency, despite its near-zero effort",
        "Accepts firmographic-only only as a deliberate short-term stopgap with the technographic layer already scheduled",
        "Promotes the model to the firmographic + technographic default rung",
        "Explains what the technographic layer adds: a paper trail of decisions the account already made, revealing buying maturity, displacement potential and integration fit",
        "Rejects adding five more firmographic filters as the fix",
        "Applies the disqualification diagnostic: criteria that cannot rule out roughly 80% of TAM are too loose to concentrate anything",
        "Recommends tightening the existing criteria rather than stacking more layers on a loose base",
        "Routes back to mbfinotti/sales-skills@sales-icp-definition if the looseness traces to the ICP itself",
        "Requires each resulting segment to map to a genuinely different motion or playbook, or nothing has been segmented"
      ]
    }
  ],
  "trigger_queries": [
    { "query": "design an account segmentation model for our mid-market SaaS", "should_trigger": true },
    { "query": "how should we segment our account universe before annual planning", "should_trigger": true },
    { "query": "build an account fit score from our closed-won data", "should_trigger": true },
    { "query": "we have 6,000 accounts in the CRM and no defensible way to rank them", "should_trigger": true },
    { "query": "every account on our target list looks equally good and reps just pick favourites", "should_trigger": true },
    { "query": "our SDRs work whatever logo they recognize and I need something defensible", "should_trigger": true },
    { "query": "where is the whitespace in our installed base", "should_trigger": true },
    { "query": "find the unsold products and untouched divisions inside our biggest customers", "should_trigger": true },
    { "query": "we want a data-backed ranking of target accounts instead of the founder's list", "should_trigger": true },
    { "query": "what weights should go into our account scoring model", "should_trigger": true },
    { "query": "we just bought intent data, how do we work it into account prioritization", "should_trigger": true },
    { "query": "should fit and intent be one score or two", "should_trigger": true },
    { "query": "calibrate our account score against the last year of closed-won deals", "should_trigger": true },
    { "query": "map our product lines against the buying centers at our strategic accounts", "should_trigger": true },
    { "query": "our account list is pure firmographics from an enrichment vendor and it isn't helping anyone", "should_trigger": true },
    { "query": "carve up the book of business before we plan next year", "should_trigger": true },
    { "query": "is machine-learning account propensity scoring worth it with 200 target accounts", "should_trigger": true },
    { "query": "build a fit and readiness model for our target accounts", "should_trigger": true },
    { "query": "we're a DTC skincare brand, how should we band customers by lifetime value", "should_trigger": true },
    { "query": "we sell through grocery chains, how do we score which chains to push", "should_trigger": true },
    { "query": "how do we organize our qualified account universe so marketing and sales work the same list", "should_trigger": true },
    { "query": "our cross-sell numbers in QBRs are just anecdotes", "should_trigger": true },
    { "query": "expansion revenue is flat and nobody knows where to look inside existing customers", "should_trigger": true },
    { "query": "what signal layers should feed an account score - firmographics, tech stack, intent?", "should_trigger": true },
    { "query": "how often should we re-score accounts", "should_trigger": true },
    { "query": "we need one scoring contract RevOps owns instead of three definitions of enterprise", "should_trigger": true },
    { "query": "our best customers don't look alike on firmographics at all", "should_trigger": true },
    { "query": "help me rank accounts by how well they fit and separately by how ready they are", "should_trigger": true },
    { "query": "the self-serve side and the sales team keep working the same companies, how do we split the universe", "should_trigger": true },
    { "query": "score product-qualified accounts using ICP fit and product usage", "should_trigger": true },
    { "query": "we have technographic data now, does it beat just using company size", "should_trigger": true },
    { "query": "prioritize which of our 900 prospects to work this quarter", "should_trigger": true },
    { "query": "our account scores don't predict anything, high scorers close no better than low ones", "should_trigger": true },
    { "query": "what does a good account scoring rubric look like, including negative signals", "should_trigger": true },
    { "query": "build the expansion map for our top 20 customers", "should_trigger": true },
    { "query": "how do I prove our account targeting model is actually working", "should_trigger": true },
    { "query": "reps say the scoring model is a black box and ignore it", "should_trigger": true },
    { "query": "we want to group customers by value and engagement for differentiated treatment", "should_trigger": true },
    { "query": "how do we rank accounts for an ABM launch next quarter", "should_trigger": true },
    { "query": "our whole account model lives in a slide deck and nobody uses it", "should_trigger": true },
    { "query": "where should the cutoff between tier 1 and tier 2 accounts sit", "should_trigger": false },
    { "query": "how many named accounts can one enterprise AE realistically cover", "should_trigger": false },
    { "query": "what coverage model should tier 3 get - touch cadence, channel mix, QBR frequency", "should_trigger": false },
    { "query": "name our account tiers and set the capacity cap per rep", "should_trigger": false },
    { "query": "define the ideal customer profile for our new security product", "should_trigger": false },
    { "query": "what disqualifiers belong in our ICP rubric", "should_trigger": false },
    { "query": "our ICP hasn't been refreshed in two years, redo the criteria", "should_trigger": false },
    { "query": "score inbound leads by email engagement so SDRs work the hot ones first", "should_trigger": false },
    { "query": "our MQL threshold is letting junk through, retune it", "should_trigger": false },
    { "query": "design the sales org topology - pods or assembly line", "should_trigger": false },
    { "query": "execute the territory carve and assign reps to regions", "should_trigger": false },
    { "query": "what SDR to AE ratio should we run at 40 reps", "should_trigger": false },
    { "query": "set next year's quotas for the mid-market team", "should_trigger": false },
    { "query": "estimate the TAM for warehouse automation software in Europe", "should_trigger": false },
    { "query": "how much pipeline do we need to hit a $14M number", "should_trigger": false },
    { "query": "should we be product-led or sales-led as a company", "should_trigger": false },
    { "query": "design the commission plan and pay mix for our new AE role", "should_trigger": false },
    { "query": "write a cold email sequence for CFOs at logistics companies", "should_trigger": false },
    { "query": "plan a 14-touch outbound cadence across email, phone and LinkedIn", "should_trigger": false },
    { "query": "score this deal against MEDDPICC", "should_trigger": false },
    { "query": "who is the champion in this deal based on the call notes", "should_trigger": false },
    { "query": "review this call transcript and tell me what the rep did wrong", "should_trigger": false },
    { "query": "build discovery questions for a first call with a VP of Engineering", "should_trigger": false },
    { "query": "handle the 'we already use a competitor' objection", "should_trigger": false },
    { "query": "plan my concessions before the renewal negotiation", "should_trigger": false },
    { "query": "write the recap email after today's meeting with the buyer", "should_trigger": false },
    { "query": "build the ROI business case for this one deal", "should_trigger": false },
    { "query": "check this cold email setup for SPF and DMARC alignment", "should_trigger": false },
    { "query": "write subject line variants and a split test plan", "should_trigger": false },
    { "query": "what red flags are hiding in this deal's notes", "should_trigger": false },
    { "query": "design the interview loop for hiring SDRs", "should_trigger": false },
    { "query": "how do I get promoted from SDR to AE", "should_trigger": false },
    { "query": "which sales podcasts and newsletters should I follow", "should_trigger": false },
    { "query": "segment our VLANs so the lab network is isolated from the rest of the house", "should_trigger": false },
    { "query": "set up a Segment tracking plan with a clean event taxonomy", "should_trigger": false },
    { "query": "write a cold call opener for a head of ops at a 300-person manufacturer", "should_trigger": false },
    { "query": "split our customer data platform audiences for a paid media test", "should_trigger": false },
    { "query": "build a partner tiering model for our reseller program", "should_trigger": false },
    { "query": "what health score should trigger a churn save play on existing customers", "should_trigger": false },
    { "query": "which accounts get an executive sponsor assigned this quarter", "should_trigger": false }
  ]
}

# Case studies and figure grading

What the segmentation payoff literature actually contains, and how much weight each figure deserves. Attribute every number below when citing it - none of them are independent research.

## Named case studies (vendor/consultant-published, self-reported)

- **Snowflake + ZoomInfo** - Snowflake's sales data-science team built an Account Propensity Score on 70+ firmographic/technographic fields, at least a third of the critical features sourced from ZoomInfo. Reported (ZoomInfo's own case study, quoting Snowflake's sales data-science manager):
  - 25% higher customer engagement.
  - 2x new-customer conversion.
  - 90% higher opportunity open rates on top-scored accounts.

  The scores feed territory planning and account distribution directly.

- **Ivanti + 6sense** - reported in 6sense's customer story:
  - +71% opportunities created ($263.2M influenced pipeline).
  - +94% opportunities won ($18.4M revenue).
  - +154% year-over-year win rate.
- **SAP Concur + Demandbase** - after three to four months of advertising to a prioritized target-account group:
  - Closed-won revenue +52%.
  - Deal size +57%.
  - Pipeline +59%.
- **SalesGlobe (consulting)**:
  - A 1,000+ rep distributor: tiered re-segmentation plus territory optimization, 39% market growth vs. 24% in the comparison cohort, +20% selling time.
  - A post-merger healthcare SaaS company: a new four-band segmentation model plus a dedicated inside-sales function grew SMB penetration from under 3% to 28%.

## How to grade the figures

- **Vendor bias is pervasive.** The firms behind most of this literature (6sense, Demandbase, ZoomInfo, Terminus, enrichment and ABM platforms) sell the scoring tooling the case studies showcase, and every metric above is self-reported single-customer marketing. Use them as existence proofs - segmentation done well can move these numbers. Never use as expected results.
- **Survey-based benchmarks** (ITSMA/ABM Leadership Alliance, Bridge Group) have a real methodology but skew toward organizations with already-mature GTM functions, and widely circulated derivatives of their figures often drift from the primary publication. Verify against the primary source before quoting a specific percentage.
- **Granular numbers come from practitioner blogs, not peer review.** Rubric point allocations, weighting splits (firmographic ~30-40% / technographic ~20-25% / intent ~25-40%) and tier account counts represent common practice, not proven optima - present them as directional conventions.

## How much weight each figure in SKILL.md deserves

| Figure                                                                                                           | Grade                                                                                          |
| ---------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------- |
| 3-5x revenue potential in won-account whitespace                                                                 | Practitioner estimate, directional                                                             |
| 5-25x cost of new acquisition vs. expansion                                                                      | HBR-cited range, widely reproduced                                                             |
| 115-130% NRR with operationalized whitespace vs. ~100-105% SaaS median                                           | Self-reported by practitioners, directional                                                    |
| Layer weighting splits (30-40 / 20-25 / 25-40)                                                                   | Practitioner convention, calibrate locally                                                     |
| 10%+ revenue drained by poor data quality                                                                        | Industry estimate, directional                                                                 |
| ~80%-of-TAM disqualification test                                                                                | Practitioner heuristic, useful as a diagnostic                                                 |
| ACV thresholds in the segment-of-one debate (1:many uneconomic at low ACV; 1:1 needs six/seven-figure potential) | Practitioner consensus, economics-dependent                                                    |
| Rep selling hours (~1,500-2,000/year)                                                                            | Published figures vary by source; use as order of magnitude                                    |
| 15-25% higher quota attainment from data-driven territory sizing                                                 | Attributed to Forrester territory-design research, reproduced rather than traced to the report |

## Consensus vs. contested

Stated by the research literature itself, worth preserving when advising:

- **Consensus**:
  - Combine fit with intent.
  - Calibrate from closed-won data.
  - RevOps owns the model.
  - Map segments to differentiated motions.
  - Hybrid territory carving.
  - Whitespace drives NRR.
- **Contested or business-specific** (genuinely local decisions, not universal constants):
  - Segment-of-one at scale.
  - Annual vs. quarterly territory re-carving.
  - Whether any blended fit×intent number is ever acceptable.
  - The precise weighting splits.

# Worked example: two-axis fit×readiness scoring grid

A worked pass for a fictional mid-market data-integration vendor ("Nortida", ~$40K ACV, sales-led with a self-serve entry tier). Every weight below is a directional convention adapted from published practitioner rubrics - calibrate against the user's own closed-won base before routing anything on it.

## The fit axis (0-100, slow-moving)

Built from the ICP's criteria plus the chosen signal layers. The firmographic breakdown follows a published 100-point rubric's proportions (15/10/10/5 inside a 40-point firmographic block); the rest is reweighted to Nortida's layers.

| Dimension             | Points    | Scoring notes                                                                                                                                                    |
| --------------------- | --------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Industry vertical     | 15        | Full points for the two verticals holding 70% of trailing-12-month ARR                                                                                           |
| Employee count        | 10        | Full points in the 200-1,000 band the closed-won analysis surfaced                                                                                               |
| Revenue band          | 10        | Proxy for budget; sourced from enrichment, refreshed quarterly                                                                                                   |
| Geography             | 5         | Serviceable regions only                                                                                                                                         |
| Technographic         | 25        | Runs a cloud warehouse (+15); runs a displaceable legacy ETL tool (+10)                                                                                          |
| JTBD / pain-point fit | 20        | Evidence the account has the job the product is hired for: relevant job postings, a stated integration project, a trigger event (funding, regulation, migration) |
| Growth signals        | 15        | Headcount growth, new market entry - expansion pressure on data infrastructure                                                                                   |
| **Deductions**        | up to -25 | Competitor multi-year contract just signed (-15) · regulated data residency the product can't meet (-10, from the ICP's disqualifiers)                           |

## The readiness axis (0-100, decaying)

Updates continuously; every signal carries an expiry so the axis decays back toward zero without fresh evidence.

| Signal                                         | Points | Expiry  |
| ---------------------------------------------- | ------ | ------- |
| Third-party intent surge on category keywords  | 30     | 30 days |
| Pricing/docs page visits (first-party)         | 25     | 30 days |
| Self-serve workspace activated (product usage) | 25     | 90 days |
| Relevant hiring (data engineering roles)       | 20     | 90 days |

## The grid and the quadrant plays

| Account           | Fit | Readiness | Quadrant  | Play                                                                 |
| ----------------- | --- | --------- | --------- | -------------------------------------------------------------------- |
| Meridian Foods    | 91  | 82        | high/high | Route to AE now, full sequence, exec touch                           |
| Cardell Logistics | 88  | 31        | high/low  | Nurture: 1:few ABM, quarterly check for trigger events               |
| Ostrow Media      | 62  | 89        | low/high  | Qualify hard before spending AE time; steer to self-serve entry tier |
| Byrne & Klee LLP  | 34  | 12        | low/low   | Disqualify; automated touch only                                     |

The play differs per quadrant - that is the entire argument for two axes. High-fit/low-readiness accounts are the long-term pipeline; low-fit/high-readiness accounts are where reps waste quarters chasing enthusiasm the economics never reward. If low-fit/high-readiness accounts keep _closing well_, that is not a scoring bug - it is an ICP drift signal to route upstream.

## Calibration pass (never optional)

Score the last 12 months of closed-won and closed-lost deals on the draft fit axis. Nortida's pass: 74% of closed-won ARR landed in the top fit quartile, and top-quartile accounts converted at roughly 3x the bottom half - the model health test passes.

If high scorers do not convert materially better than low scorers, reweight and re-run before anything routes on the score. Re-run this calibration once or twice a year, and check the cohort-bias trap each time: if the closed-won base is still dominated by the founding customer type, adjacent growth segments are being systematically under-scored.

## Negative example: the blended score

The tempting shortcut is one composite: fit and readiness averaged into a single 0-100. Then Meridian-style and Ostrow-style accounts converge on the same number: a 90-fit/35-readiness account and a 65-fit/88-readiness account both read as "≈78/100" and get worked identically, even though one needs patient nurture and the other needs hard qualification.

The blend erases exactly the information that changes the play, and reps learn to ignore the number within a quarter.

- Keep the axes separate.
- If a routing rule genuinely needs one number, multiply fit × readiness so only accounts strong on both dominate.
- Keep the two sub-scores visible on the record.

# Worked example: whitespace map for one strategic account

A worked expansion map for a fictional strategic customer ("Helvane Group", industrial conglomerate, current ARR $220K on a four-product line). Whitespace here means the products, divisions and buying centers inside an already-won account where the vendor has no presence yet - distinct from prospect whitespace (net-new accounts), which the fit score ranks instead.

## The heatmap

Products × buying centers, three states:

- **active** - footprint exists.
- **open** - relevant, unsold.
- **entrenched** - a competitor holds it.

|                  | NA Operations | EMEA Operations | Finance | HR  | Subsidiary: Helvane Rail |
| ---------------- | ------------- | --------------- | ------- | --- | ------------------------ |
| Core platform    | active        | active          | open    |     | open                     |
| Analytics add-on | active        | open            | open    |     | open                     |
| Security module  | open          | open            |         |     | entrenched               |
| Premium support  | active        | open            |         |     | open                     |

Reading the heatmap:

- EMEA Operations is a penetrated division with three unsold products (cross-sell).
- Finance is an untouched buying center on two relevant products (new-buying-center land).
- Helvane Rail is a whole unpenetrated subsidiary.
- The Rail security cell is competitor-entrenched, a displacement bet rather than routine expansion, so it carries its own timeline and risk note.

## Resolve each gap into one of the four plays

| Gap                            | Play                     | Next step                                                              |
| ------------------------------ | ------------------------ | ---------------------------------------------------------------------- |
| EMEA analytics + support       | Cross-sell / tier upsell | Champion intro from NA sponsor; EMEA ops review                        |
| Finance (platform + analytics) | New buying center        | Org-chart mapping - no contact exists today. Identify the budget owner |
| Helvane Rail (all lines)       | New subsidiary land      | Exec-to-exec introduction via the group-level sponsor                  |
| NA seat growth (platform)      | Seat/usage expansion     | Usage sits at 92% of licensed seats - expansion trigger already fired  |

New-buying-center plays are the ones that require explicit org-chart and buying-committee mapping: surface the dormant stakeholders and untapped business units the current relationship doesn't touch, and track multi-product attach across them.

## Account-plan fields (kept on the CRM account record)

- Services currently provided, per buying center.
- Services relevant but not provided (the map above, as data).
- Org coverage map: known relationships vs. dark buying centers.
- Prioritized gap list with the play and owner per gap.
- Each gap tracked as a pipeline-staged opportunity: identified → qualified → in discussion → proposal → won/lost.

The pipeline-stage rule is what turns the map into management-grade data: leaders asking for cross-sell numbers in a QBR get stage-by-stage movement, not recollections.

## Refresh triggers

Refresh strategic-account maps at least quarterly, and immediately on any of the following:

- An acquisition (Helvane buying a new subsidiary redraws the columns).
- A leadership change in a mapped buying center.
- A renewal window opening.
- A competitor entering any cell.

## Negative example: the eyeballed account plan

The failure documented across sources: the seller eyeballs the account, remembers a few open conversations, and calls it a plan.

Symptoms:

- QBR answers are anecdotes (e.g. "Finance felt lukewarm last year").
- No one can say which products a division has never been offered.
- The map exists only in the rep's head, so it leaves when they do.

The documented consequence: products that should have been sold into the account two years ago still sit on the shelf, and a competitor walks in through a door the team never noticed was open. The countermeasure is the structure above, in the CRM, on a refresh cadence, framed as growing the customer's business, not conquering territory.

---
name: sales-account-segmentation
description: Designs the account segmentation model between ICP and tier structure - which firmographic, technographic, intent and jobs-to-be-done signal layers define account fit, the weighted fit score calibrated from 12 months of closed-won deals, fit and readiness as separate axes, whitespace and expansion mapping, and how the model is encoded, routed and re-scored in the CRM. Covers B2B account segmentation and B2C value-based customer segmentation. Use whenever the user mentions account scoring, fit score, whitespace, book-of-business carve-up, or reps picking accounts on gut feel, even without the word segmentation. Do NOT use for tier cutoffs and coverage (mbfinotti/sales-skills@sales-account-tiering) or ICP criteria (mbfinotti/sales-skills@sales-icp-definition).
license: MIT
metadata:
  author: Maya-Beth Finotti
  version: "1.2.2"
---

# Sales Account Segmentation

You are an advisor to sales leadership designing the account segmentation model - the layer that turns an agreed ICP into a scored, organized account universe that tiering, territory design, and coverage decisions consume. Produce the signal-layer choice, the fit-scoring model, the whitespace map, and the CRM operationalization plan - never the tier cutoffs built on top of the score.

Hold the boundary in both directions:

- The ICP criteria beneath the model belong to mbfinotti/sales-skills@sales-icp-definition. If no agreed ICP exists, route there first, because segmenting along dimensions nobody signed off on quietly re-derives the ICP, badly.
- Where the tier cutoffs sit, what the tiers are named, and the per-rep capacity caps belong to mbfinotti/sales-skills@sales-account-tiering.
- Executing the territory carve belongs to mbfinotti/sales-skills@sales-org-structure (see References).

## Invocation examples

Each ask enters at a different point. Run the interview first regardless.

- _"Segment our accounts"_ - full build: signal layers, fit score, readiness axis, whitespace map, CRM plan.
- _"Reps chase accounts on gut feel"_ / _"every account looks equally good"_ - diagnostic entry: the list is probably firmographic-only and operationally inert; run the failure-mode checks, then rebuild from the signal-layer menu.
- _"Build an account fit score"_ - scoring entry; confirm the ICP and 12 months of closed-won data exist first, because a score calibrated on intuition is a ranking of opinions.
- _"Where's the whitespace in our customer base?"_ - expansion-mapping entry. Whitespace mapping and net-new fit scoring are different disciplines that share a model; say so and scope which one the user needs before building either.

## Interview

How to run it:

- Ask before proposing, one question per message; offer the multiple-choice options where given.
- Skip anything already answered by prior context.
- If you can read CRM exports or the company's website, offer to pre-fill answers and draft candidate models for the user to correct instead of interviewing from a blank page.

1. Does an agreed ICP exist: (a) a validated rubric with disqualifiers, (b) an informal profile everyone "knows", (c) none? Ask first - (b) and (c) route upstream to mbfinotti/sales-skills@sales-icp-definition before any segmentation starts.
2. Is this B2B, or B2C? For B2C: do you sell through key accounts (retail chains, distributors, franchise groups), or direct to consumers? See B2B vs B2C below - the exercise changes shape.
3. How many closed-won deals from the last 12 months have usable CRM data: (a) under ~20, (b) ~20-100, (c) 100+? This is the calibration source - the weights get built backward from it, not from intuition.
4. What must the model drive: (a) net-new prospect prioritization, (b) expansion inside existing customers, (c) both? Net-new scoring and whitespace mapping are different disciplines - a model built for one silently underserves the other.
5. Which data sources are live today: CRM firmographics, an enrichment vendor, a technographic source, an intent vendor, product-usage data in a warehouse, first-party web analytics? The answer sets which rung of the signal-layer menu is reachable this quarter.
6. Which motions run today (self-serve/PLG, sales-led, enterprise named-account), at roughly what deal size per segment? Segmentation's output must land accounts into motions that actually exist.
7. Who consumes the model and who is its named owner - RevOps, sales ops, marketing ops? Split ownership with separately maintained definitions is a documented failure source.
8. By what date must the model drive a real decision - a territory carve, annual planning, an ABM launch?
9. Do you want a one-off win or a compounding asset: (a) a prioritized account list for this quarter, (b) a standing scoring model with refresh governance the next several planning cycles run on?
10. What is your effort ceiling: analyst hours, enrichment budget, RevOps capacity to encode fields and routing in the CRM, and the political capital to tell reps their favorite accounts score low?

Re-rank the signal-layer menu against answers 8-10 before proposing anything, and say which answer moved what:

- A hard date inside weeks holds the model at the default rung and spends the remaining hours on closed-won calibration, which is never optional.
- A compounding mandate (9b) promotes the fit×readiness rung plus CRM encoding despite the effort.
- A low political-capital ceiling means publishing per-dimension sub-scores first, since reps accept a model they can see the reasons inside.

## Segmentation is the middle layer

Three layers answer three different questions, in order:

- **ICP** asks "should we pursue this account at all" and disqualifies.
- **Segmentation** asks "how is the qualified universe organized and ranked" - it scores fit, reads readiness, and maps whitespace.
- **Tiering** asks "given a ranked account, how much effort does it get".

This skill owns the middle: it consumes the ICP, and it stops at a score plus an organized universe. The moment a cutoff or a coverage ratio appears, treat it as the sibling skill's ground.

The reason the middle layer exists at all is rep capacity: a rep has roughly 1,500-2,000 selling hours a year (published figures vary), and spreading them evenly across a flat list starves strategic accounts while over-serving low-fit ones. Firmographics alone give the universe; the further layers decide where those hours concentrate.

Segmentation is also not lead scoring: leads are individual contacts scored on engagement readiness; segmentation scores _accounts_ on structural fit. The contact half lives in mbfinotti/revops-skills@lead-scoring.

## Brainstorm before committing

A segmentation model hardens fast - routing rules, campaign lists and territory carves get keyed to it within a quarter.

1. After the interview, present 2-3 candidate segmentation models with trade-offs and one explicit recommendation:
   - Size-band model: fast, maps cleanly onto motions, risks being operationally inert.
   - Multi-signal fit×readiness model: precision, but calibration and data cost.
   - JTBD/problem-clustered model: reach for it specifically when the winners don't cluster on firmographics, since buyers span industries but share the same trigger and job.
2. Ask remaining clarifying questions one at a time, multiple-choice where possible, and get explicit approval on a model shape before weighting anything.
3. Build the charter section by section, validating each with the user before the next: signal layers → fit score and calibration → readiness axis → whitespace map → CRM operationalization → governance. A wrong layer choice invalidates every weight downstream.
4. Gate finalization on user approval of the assembled charter.

If your harness has persistent memory, store the approved charter (layers, weights, calibration result, owner, and review dates) so later runs and the sibling tiering skill start from the recorded decision.

## Signal-layer menu

Which layers build the fit model, ranked by efficiency (value returned per unit of effort):

- value: `predictive scoring > fit×readiness two-axis > firmographic+technographic > firmographic-only`
- effort: `predictive scoring > fit×readiness two-axis > firmographic+technographic > firmographic-only`
- efficiency: `firmographic+technographic > fit×readiness two-axis > firmographic-only > predictive scoring`

**Default rung: firmographic + technographic, calibrated on closed-won.**

- Firmographics answer "who" and size the universe.
- The tech stack answers "how": a paper trail of decisions the account already made, revealing buying maturity, displacement potential and integration fit.

Together they produce a fit score most teams can build in weeks from an enrichment vendor plus their own CRM.

The remaining rungs follow, in that same efficiency order:

- **Fit×readiness two-axis** - add intent and first-party behavioral signals (the "when/why" layer), and keep them on a _second axis_ rather than blending them into fit. Promote to this rung as soon as a real readiness source exists - an intent vendor, product-usage data, first-party web signals. Most of the promotion's value is in keeping the axes separate (see the next section).
- **Firmographic-only** - near-zero effort, and it still ranks below both rungs above on efficiency: a firmographic-only list is directionally right but operationally inert - every account looks equally good, reps fall back to gut feel, and the hours spent building it buy almost nothing. Acceptable only as a deliberate two-week stopgap with the technographic layer already scheduled.
- **Predictive scoring** - the starved option: ML models trained on CRM/MAP history predicting fit and timing, the highest ceiling, and it loses every efficiency round on effort. Promote it anyway when TAM runs to thousands of accounts, the CRM history is deep, and someone owns the model; skip it below that scale. Predictive scoring degrades when TAM is a few hundred accounts or the vertical has thin intent signal, where closed-won pattern analysis plus manual JTBD segmentation is the more reliable substitute.

Layer JTBD/use-case framing onto the chosen rung when firmographics fail to cluster the winners: segment by the job the account hires the product for and the trigger event that opens the buying window (regulation, incident, funding round, leadership change, contract renewal).

This ordering is a default, not a law. Re-rank it against what you know about the user:

- A PLG company already has product-usage data flowing, which pre-pays most of the fit×readiness rung's cost.
- An enterprise team with 40 nameable target accounts gets nothing from predictive scoring at any price.

Say what moved on each re-rank.

## Build the fit score

1. Consume the ICP's criteria and disqualifiers as the score's foundation - never re-derive them.
2. Assemble dimensions per chosen layer, with an explicit negative-signal deduction layer alongside the positive points.
3. Weight the layers. Published splits are directional, not universal:
   - Firmographic: ~30-40% of the composite.
   - Technographic: ~20-25%.
   - Behavioral/intent: ~25-40%.
   - Deductions on top.

   The most common weighting error is overweighting intent because clicks are easy to count: fit is the quieter but more reliable predictor.

4. Calibrate backward from the last 12 months of closed-won deals: if 70% of ARR traces to one segment shape, that shape weights highest. Re-run the calibration once or twice a year. Watch the cohort-bias trap: a training set dominated by the company's first big customer type systematically under-scores adjacent growth segments.
5. Validate with the model health test: high-scoring accounts must convert at a materially better rate than low-scoring ones, or the weights need updating before anything routes on them.
6. Publish per-dimension sub-scores so reps can see _why_ an account scored as it did - explainability is what buys adoption.

**Keep fit and readiness as two separate axes - never one blended number.**

- A 90-fit/35-readiness account is a long-term target worth nurturing.
- A 65-fit/88-readiness account is near-term but lower-quality.
- A single blended "78/100" erases exactly the distinction that changes the play.

Fit anchors and changes slowly; readiness (intent, engagement, product usage) decays on a timer and updates continuously. The quadrants of the resulting matrix each carry a different play. See [fit-readiness-scoring-example.md](./references/fit-readiness-scoring-example.md) for the worked grid, the quadrant plays and the blended-score negative example.

## Whitespace and expansion mapping

Net-new scoring ranks accounts not yet won; whitespace mapping finds revenue inside accounts already won (unsold products, unpenetrated divisions and buying centers). Run it as its own discipline whenever the interview says expansion matters: whitespace inside a won account is estimated at 3-5x the original deal's revenue potential, and acquiring a new customer runs 5-25x the cost of expanding an existing one (HBR-cited range). Teams that operationalize it report 115-130% net revenue retention against a SaaS median around 100-105% (directional, self-reported figures, but the direction is consistent across sources).

1. Map products/services against each strategic account's departments, divisions and buying centers.
2. Visualize as a heatmap (active footprint, expansion opportunity, competitor-entrenched) per department.
3. Resolve each gap into one of these plays:
   - Seat/usage expansion.
   - Tier upsell.
   - Cross-sell of adjacent products.
   - Landing a new buying center (needs explicit org-chart and buying-committee mapping).
4. Track each identified gap as a pipeline-staged opportunity (identified → qualified → in discussion → proposal → won/lost), never as a static list.
5. Refresh strategic-account maps at least quarterly, and immediately on a trigger event:
   - An acquisition.
   - A leadership change.
   - A renewal.
   - A competitor entering the account.

The documented failure is doing this by memory: sellers eyeball an account, recall a few open opportunities, and call it a plan. QBRs then produce anecdotes instead of cross-sell numbers while a competitor walks in through a door nobody noticed was open.

- Keep the map inside the CRM account record: account planning that lives outside the CRM is doomed to low adoption.
- Frame the motion as growing the customer's business, not conquering territory: the conquest metaphor misreads the relationship being expanded.

Worked map and account-plan fields: [whitespace-map-example.md](./references/whitespace-map-example.md).

## Operationalize in the CRM

A model that lives in a slide deck segments nothing. Encode it in the CRM:

1. **Fields and routing** - write the fit score, readiness score, sub-scores and segment onto the account record via automated enrichment; route each segment to its coverage lane with SLAs on the handoff. Use waterfall enrichment (query providers in sequence until a field fills) rather than betting on one vendor's coverage.
2. **Re-scoring cadence** - review the full model quarterly; recalibrate weights once or twice a year; let readiness signals update continuously. Technographic and intent data decay fast - an annual pull is stale by month three.
3. **Threshold-triggered transitions** - move accounts between segments on explicit thresholds (an employee-count or ARR crossing, a downgrade review after two quarters below the segment minimum), never on ad-hoc judgment.
4. **One owner, one definition set** - RevOps owns the model as a shared scoring contract with joint sales/marketing sign-off, encoded once in the CRM. When marketing's "enterprise" starts at 500 employees and sales' at 1,000, scoring, routing and campaigns all silently diverge.

## From segments to motion

Segmentation's output must be motion-actionable. The practical question is never "which motion for the company" (mbfinotti/sales-skills@sales-motion's ground) but "which motion for which segment", since most B2B companies run several at once:

- Self-serve/PLG for the low end.
- Sales-led for mid-market.
- Named-account for enterprise.

Where PLG and sales coexist, define an explicit PLG-to-sales handoff trigger (a usage threshold, a team size, an enterprise domain) so the two engines stop cannibalizing the same accounts. In a PLG segment the scoring unit becomes the product-qualified lead: ICP-fit score × product-usage score, with the usage context visible to reps, not a black-box number.

The engagement bands the model feeds are the settled 1:1 / 1:few / 1:many vocabulary (ITSMA's tiering, popularized by Jon Miller).

"Segment of one", AI-personalized treatment of every account, is the aspiration vendors are pushing down-market for the top band, not a rival model. The genuine debate is economic: at what per-account value bespoke treatment pays for itself.

- 1:many programs generally aren't worth running at low ACVs.
- True 1:1 needs six/seven-figure potential.

Define the _entitlements_ (what treatment each band gets) before assigning anyone to a band, or political pressure inflates the top one.

Where the cutoffs sit is mbfinotti/sales-skills@sales-account-tiering's job. Stop at a score every account carries, including executive gut-feel picks, which get scored on the same model as everyone else and re-scored quarterly.

The same output feeds territory design. The carving dimensions all balance on weighted opportunity (account count × average ACV × estimated win rate), never on raw account count:

- Geography.
- Vertical.
- Size-band.
- Hybrid.

Executing the carve belongs to mbfinotti/sales-skills@sales-org-structure.

## The model charter

Deliver the decisions as one artifact the next planning cycle can execute without re-litigating:

```
CONTEXT: ICP source · closed-won volume used for calibration · net-new/expansion/both · named owner
SIGNALS: layers chosen and why · source per layer · decay and refresh per layer
FIT SCORE: dimensions and weights · negative deductions · calibration result against closed-won
READINESS: second-axis signals · decay timer · update frequency · quadrant plays
WHITESPACE: per-account expansion maps · gap pipeline tracking · refresh triggers
CRM: fields · routing rules and SLAs · re-scoring cadence · transition thresholds
MOTION MAP: which motion serves which segment · PLG-to-sales handoff trigger
GOVERNANCE: RevOps owner · quarterly review · 1-2x/year recalibration · event triggers
CONFIDENCE: which figures are published research vs. directional convention vs. vendor claim
```

## B2B vs B2C

**B2B** is the default framing above: accounts, firmographic/technographic layers, buying committees.

**B2C through key accounts** - a brand selling through retail chains, distributors or franchise groups - keeps the machinery and swaps the fit inputs:

- Fit inputs: store count, shelf and category position, geography and banner affiliation replace firmographics and technographics.
- Readiness role: sell-through and reorder velocity.
- Whitespace: unsold categories and unpenetrated banners or regions inside a chain already stocked.

The two-axis rule, closed-won calibration and CRM operationalization transfer unchanged, and the resulting scores feed the same downstream tiering.

**Direct-to-consumer** has no accounts - value-based customer segmentation plays the structural role: customers banded by lifetime value or spend (the fit analog) crossed with engagement/recency (the readiness analog, RFM-style), driving differentiated treatment.

What transfers:

- Calibrating bands from actual purchase data rather than assumptions.
- Keeping value and engagement as two axes.
- The over-segmentation warning.
- The refresh cadence.

What doesn't transfer:

- Firmographic layers.
- Buying-committee logic.
- Org-chart whitespace mapping.

Say so rather than forcing the account machinery onto it.

## Failure modes

Run the finished charter against each of these before it ships:

- **Scores that aren't actionable.** Check: does each segment map to a genuinely different motion or playbook? A model that produces a score but no differentiated treatment has segmented nothing.
- **Stale, firmographic-only data.** Check: does each layer have a refresh mechanism matched to its decay speed? Poor data quality is estimated to drain 10%+ of revenue through misfired targeting.
- **Cohort bias.** Check: was the calibration base broadened beyond the first big customer type, and is recalibration scheduled?
- **Fit and readiness blended into one number.** Check: two axes, two update cadences, quadrant plays.
- **Over-segmentation.** Check: more than 6-8 segments means some can't sustain a distinct playbook - merge them.
- **Misaligned definitions.** Check: one definition set, encoded once, jointly signed off.
- **No whitespace visibility.** Check: do strategic accounts have a data-backed map, or anecdotes?
- **Set-and-forget.** Check: cadence and event triggers scheduled; success measured on pipeline and conversion movement, never on "number of accounts above threshold" - that optimizes a score nobody trusts.

Diagnostic worth running on the whole model: if its criteria can't disqualify roughly 80% of TAM, the rubric is too loose to concentrate anything - tighten the existing criteria rather than stacking more layers on a loose base. If the looseness traces to the ICP itself, route back to mbfinotti/sales-skills@sales-icp-definition instead of fixing it here.

## Measurement

The charter is not done until all of these pass; iterate until 100%:

- Every layer names its source, refresh cadence and decay assumption; the fit score reports its closed-won calibration result.
- Fit and readiness are separate axes with distinct update cadences, and each quadrant names its play.
- If expansion is in scope, every strategic account has a whitespace map with gaps tracked as pipeline stages.
- The CRM plan names fields, routing SLAs, transition thresholds, the RevOps owner and the review calendar.
- Every figure carries a confidence grade - the case-study numbers this domain circulates are vendor-published and self-reported (see the case-study reference below); the B2B or B2C scope is stated explicitly.

Live KPIs after shipping:

- The high-vs-low scorer conversion delta (the model's standing health test).
- Share of new pipeline inside top segments.
- Whitespace-gap pipeline movement.
- NRR trend, where expansion is in scope.

If high scorers stop outperforming, recalibrate. Don't defend the model.

## References

- See mbfinotti/sales-skills@sales-icp-definition for the ICP criteria and disqualifiers this model consumes - the upstream boundary. Segmentation never re-derives them.
- See mbfinotti/sales-skills@sales-account-tiering for the tier cutoffs, names, capacity caps and coverage levels built on the fit score this skill produces.
- See mbfinotti/sales-skills@sales-org-structure for executing the territory carve the weighted-opportunity output feeds.
- See mbfinotti/sales-skills@sales-quota-setting for the territory-potential weighting this model supplies - quotas weighted on unsegmented accounts split the target by headcount rather than by opportunity.
- See mbfinotti/sales-skills@sales-motion for choosing the company-level motion mix the per-segment motion map plugs into.
- See mbfinotti/revops-skills@lead-scoring for contact-level engagement scoring - a different object from account fit.
- See [./references/fit-readiness-scoring-example.md](./references/fit-readiness-scoring-example.md) for the worked two-axis scoring grid, quadrant plays, calibration pass and the blended-score negative example.
- See [./references/whitespace-map-example.md](./references/whitespace-map-example.md) for the worked whitespace heatmap, account-plan fields and gap-pipeline tracking.
- See [./references/case-studies-and-figure-grading.md](./references/case-studies-and-figure-grading.md) for the named case studies and how much weight each figure this skill cites deserves.