Territory Planning | Claude Code
Using Octave in Claude
Workflow
Territory
Planning
Surface
Claude Chat (MCP)
Cowork
Claude Code
Difficulty
Easy
Medium
Hard
Audience
Sales leaders, RevOps, and AEs at SKO or start of quarter
What it does
Build an interactive territory planning app that qualifies accounts via Octave, layers in historical win rates by persona and segment, tiers by territory, and flags rebalancing opportunities
Scenario
It's SKO prep or the start of a new quarter. Leadership needs territory plans. Right now someone in RevOps is in a spreadsheet manually tiering 200 accounts based on gut feel and last year's pipeline. There's no systematic way to factor in ICP fit, which personas converted historically, what segments perform in which geos, or how to balance tier distribution against pipeline targets. By the time you've assigned territories, the spreadsheet is stale and the methodology behind the tiers lives in someone's head.
How It Works
You build a Claude Code app. It reads your target account list, runs each account through Octave's qualification and enrichment, pulls historical win rate data by persona and segment via a saved Octave agent, layers in CRM pipeline data, and outputs an interactive HTML dashboard. Accounts are auto-tiered and grouped by territory. The app flags imbalances and recommends territory splits. Re-run it anytime to refresh with updated data.
Since every team's stack is different, this walkthrough focuses on what the app does and what Octave tools it calls — not a line-by-line build guide.
Account List
CSV or CRM query
Octave Qualify
Enrich + score each account
Win Rate Analysis
Historical rates by persona, segment
Tier + Rebalance
Auto-tier, flag territory gaps
Interactive App
Filter, adjust, export
The Old Way
❌ Manual ICP Scoring
You can sort and filter by size and segment in a spreadsheet, but systematically scoring 200 accounts against ICP criteria, personas, and product fit is a separate project nobody has time for
❌ Win Rates Are Separate
Knowing which personas and segments actually closed last year means pulling a report, cross-referencing deals, and doing the math. That analysis usually lives in someone's head or a one-off slide.
❌ Days of RevOps Time
Between the data pull, the scoring, the tiering, the territory balancing, and the formatting, it's a multi-day project every quarter
❌ Static Output
The end result is an Excel that works but is painful to iterate on in real time with leadership and not exactly what you want to walk through at SKO
⏱️ 2-3 days of RevOps work per quarter
The New Way
✓ ICP-Scored Tiers
Every account is qualified via Octave against your product, segment, and persona criteria. Tiers are based on ICP scores and historical close rates, not a manual sort.
✓ Win Rate Overlay
Historical win rates by persona and segment are layered into scoring via a saved Octave agent. Accounts in segments with higher close rates get a scoring boost.
✓ Iterate in Real Time
Build the first version, walk through it with leadership, adjust tier rules, move accounts, split territories. Go back to Claude Code with the feedback and it rescores and regenerates the dashboard. A week of back-and-forth happens in one session.
✓ Purpose-Built UI
A fun, interactive dashboard you can filter by geo, segment, and persona. Way more engaging to walk through at SKO than column Q of a pivot table. Exportable when you need the raw data.
⏱️ Build once, re-run in minutes. Export when ready.
What You Get
✓ Tiered Account Cards
Every account scored and assigned to Tier 1, 2, or 3 with ICP qualification score, persona match, segment fit, and win rate context
✓ Territory Views
Accounts grouped by territory and geo with filterable views by segment, persona, and tier. See pipeline coverage at a glance.
✓ Rebalancing Flags
Territories with too many or too few Tier 1 accounts are flagged. Recommendations on where to split, merge, or reassign.
✓ Approach by Tier
Recommended outreach strategy per tier: Tier 1 gets multi-threaded ABM, Tier 2 gets targeted sequences, Tier 3 gets scaled outbound.
How to Build It
1
Set up your MCP connections
Connect the Octave MCP for qualification and enrichment. Connect your CRM MCP (Salesforce, HubSpot) for pipeline data and deal history. Optionally connect BigQuery or a data warehouse for historical win rate data if it's not in the CRM.
The Octave MCP gives you: enrich_company, qualify_company, run_agent. The CRM MCP gives you deal records, pipeline values, and territory assignments.
2
Define your inputs and rules
Tell Claude Code your account list (CSV or CRM query), target geos, pipeline targets per territory, and tier distribution rules (e.g., "max 20 Tier 1 accounts per rep, 40 Tier 2, the rest Tier 3").
Build me a territory planning app. Inputs: - accounts.csv (200 target accounts with domain, geo, current owner) - Pipeline target: $4.2M per territory per quarter - Tier rules: max 15 Tier 1, 30 Tier 2, rest Tier 3 - Geos: West, Central, East, International
3
Account enrichment and qualification loop
The app iterates through every account in your list. For each one, it calls enrich_company to get firmographics, vertical, and tech stack, then calls qualify_company to score against your product, segment, and persona criteria. Each account gets an ICP fit score with rationale.
4
Historical win rate analysis via Octave skills and agents
The app uses Octave's built-in skills like /octave:wins-losses and /octave:icp-refine to analyze your historical closed-won and closed-lost deals. It pulls win rates by persona, segment, geo, and deal size. Accounts in segments and personas with higher historical close rates get a scoring boost. You can also run saved Octave agents or create your own custom skills for more specific analysis.
This is where the tiering gets smart. An account might score well on ICP fit but sit in a segment where you've historically lost 80% of deals. The win rate layer catches that.
5
Tiering and territory assignment
The app combines ICP qualification score + historical win rate + pipeline value to assign each account to a tier. It groups accounts into territories by geo and checks each territory against your distribution rules and pipeline targets. If a territory is overloaded (too many Tier 1s, or 90% of pipeline concentrated in one geo), it flags the imbalance and recommends splits.
6
Build the interactive dashboard
You work with Claude Code to build an interactive HTML app. Think: account cards showing tier, score, segment, and persona match. Territory views with pipeline coverage bars. Rebalancing recommendations. Approach strategy per tier. Filters for geo, segment, persona, and tier. Expect iteration here. The first version gets the structure right, then you refine the UI, adjust what's shown, and dial in the layout to match exactly what you want to present.
This is a purpose-built app, not a spreadsheet. It's visual, interactive, and ready to present at SKO. The build itself takes some back-and-forth with Claude Code to get the UI and data presentation exactly right for your team.
7
Review with leadership, iterate, regenerate
Walk through the dashboard with leadership. They say "move these three to Tier 1, split California into NorCal and SoCal, we need more mid-market in East." Go back to Claude Code, feed in the feedback, adjust tier rules and territory boundaries. It rescores and regenerates the dashboard with the changes. What used to be a week of back-and-forth between RevOps and leadership happens in one working session.
This isn't a one-and-done output. Re-run mid-quarter with new accounts, updated pipeline data, or revised targets. The app regenerates in minutes.