Territory Design Vendor Checklist: Questions Before You Buy Mapping, Routing or Protection Logic

Buying territory-design software, consulting or channel-routing services is easy. Buying a decision system that still works after the first dispute is harder.

A strong vendor should be able to explain where the market data came from, how account hierarchies are built, how routing rules are prioritized, what happens when two rules conflict, and how a manager can reconstruct a past decision. If the product only produces attractive maps, it is not enough.

Before you sign, ask the questions below. They apply whether the vendor sells mapping software, partner relationship management, sales-territory optimization, channel consulting or a managed routing service.

Start with the hard facts: what does the system actually ingest?

Ask for the exact input inventory.

Typical inputs include:

CRM accounts and opportunities;
billing/shipping locations;
partner locations and capabilities;
industry codes;
employee or revenue bands;
customer parent-child relationships;
service radii;
historic wins/losses;
web/inbound lead location;
protected account lists;
opportunity-registration events.

Then ask which fields are mandatory and which are merely “nice to have.”

A model that requires perfect employee counts and clean global hierarchies may look sophisticated in a demo and fail in a CRM where 30% of accounts lack those fields.

How current is the market data?

If the vendor supplies external opportunity or establishment data, ask the vintage and refresh cycle.

U.S. Census County Business Patterns is an example of a transparent public benchmark: the 2023 CBP dataset provides annual establishment and employment information by geography and detailed industry. Census Business Builder provides an accessible way to explore selected demographic and economic data. Those public sources are useful for market sizing, but annual public statistics are not a substitute for current account-level CRM data.

The question is not “Do you have data?” It is:

reference year;
source;
refresh cadence;
geography available;
industry classification used;
suppression/imputation method;
how the vendor handles businesses that opened or closed after the reference period.

Ask the vendor to label modeled data separately from observed data.

Can the system model account hierarchy without hiding uncertainty?

A territory can look clean until one enterprise has 200 locations.

Ask how the vendor represents:

parent companies;
subsidiaries;
franchises;
dealer groups;
branches;
joint ventures;
acquisitions in progress;
brands owned by the same parent.

More importantly, ask how uncertain relationships are flagged. A guessed parent should not silently override a local opportunity.

Request an “explain this route” view: given one lead, can a user see which hierarchy and rule caused the assignment?

What rule takes precedence?

Ask the vendor to write the precedence stack in plain English.

For example:

named global account;
active opportunity registration;
existing customer expansion;
service/installation requirement;
geographic default.

Then create a lead that matches three rules at once.

If the system cannot show which rule won and why, disputes will move from sales meetings into technical support tickets.

Can rules be time-bounded?

Protection should rarely be a timeless Boolean field.

Ask whether the system can support:

effective dates;
expiration dates;
renewal criteria;
ramp periods;
grandfathered active opportunities;
post-termination service obligations;
future boundary changes.

A territory design without time is a snapshot. Commercial programs are moving systems.

Can you separate selling ownership from service ownership?

This matters in national and digital accounts.

The commercial owner may negotiate price and forecast demand. A local partner may install, deliver or support the customer. If the system can store only one “owner,” incentives become distorted.

Ask whether it supports:

account owner;
opportunity owner;
referral source;
fulfillment partner;
service partner;
commission/revenue-credit fields.

Even if the CRM ultimately stores fewer fields, the design process should recognize the distinction.

How are exceptions handled?

Every territory system has exceptions. The problem is invisible exceptions.

Ask:

who can override;
whether a reason is mandatory;
whether overrides expire;
whether original and changed assignments remain visible;
whether managers can report on frequent exception classes;
whether partners can see status without seeing confidential details.

An exception log is a diagnostic dataset. If 20% of national accounts require the same manual workaround, the core rule needs redesign.

What does the vendor mean by “optimization”?

Make them define the objective function.

Possible objectives include:

equal account count;
equal revenue opportunity;
equal workload;
minimum travel;
maximum service coverage;
balanced partner capacity;
growth potential;
preservation of existing relationships.

Two optimization tools can produce completely different maps because they optimize different outcomes.

Ask for the weights, constraints and trade-offs. Do not buy a black box whose definition of “balanced” cannot be explained to the sales team.

Can we run scenarios without publishing them?

A buyer should be able to ask:

What if we add two dealers?
What if a national account moves to direct sales?
What if one partner loses service certification?
What if we protect e-commerce differently?
What if we combine two territories?

Scenario modeling should show who gains and loses accounts, opportunity and workload before any live assignment changes.

This is especially important because boundary changes create human and commercial consequences even when the math improves.

How does it measure whitespace?

A vendor may label an area “under-covered” because it has few partners. That can be wrong if demand is low.

Ask whether whitespace uses:

target establishment count;
historic demand;
current pipeline;
demographic/economic indicators;
travel/service constraints;
digital lead volume;
industry concentration.

Public Census establishment data can provide one external benchmark for U.S. markets. It should be combined with the company’s actual conversion and economics.

What is the integration failure mode?

Ask what happens if CRM sync stops.

Does routing fail closed, fail open, or continue with stale data? How does a user know?

Document:

API limits;
sync frequency;
retry behavior;
duplicate logic;
account-ID strategy;
field ownership;
rollback procedure.

A territory tool should not create a second invisible CRM.

How portable are our rules and data?

Ask what you can export:

territory definitions;
shapes/geographies;
account assignments;
rule precedence;
exception history;
scoring outputs;
user-entered notes;
partner capability data.

If you leave the vendor, can another system reconstruct the decisions?

Data portability is not only an exit concern. It prevents routing logic from becoming institutional knowledge trapped inside one dashboard.

How does the vendor handle security and role access?

Territory data can contain sensitive customer, pipeline and partner information.

Ask about:

authentication;
role-based access;
audit logging;
data retention;
subprocessors;
incident response;
customer data used for product training or benchmarking;
regional hosting needs if relevant.

Route visibility should follow business need. A partner usually does not need to see every competing partner’s pipeline to understand its own assignment.

What legal assumptions are built into the template?

This question separates operational tooling from legal judgment.

FTC public guidance explains that manufacturer-imposed territory/customer restrictions and exclusive dealing can be lawful in many contexts under U.S. federal antitrust principles, while market facts, coordination, state rules and other jurisdictions can matter. Franchise rules and contract law may also become relevant depending on the structure.

A software vendor should not quietly encode a legal conclusion into “best practice.”

Ask:

Does the template assume exclusivity?
Does it set resale-price behavior?
Does it restrict customers or channels?
Is any rule described as “legally required”?
Can those assumptions be disabled?
Which terms should counsel review?

Use legal professionals for material legal interpretation in the applicable jurisdiction.

How will we know the project worked?

Demand a before-and-after scorecard.

Useful measures can include:

time to assign an inbound lead;
duplicate ownership disputes;
manager hours spent on exceptions;
percentage of target accounts with active coverage;
response time;
conversion by territory;
gross profit by territory;
partner activation;
service travel/time;
opportunity reassignment rate.

Do not let the success metric be “maps delivered.”

What happens after go-live?

Territories decay.

Accounts open and close, employees move, partners gain and lose capability, products change, national accounts restructure and strategic priorities shift.

Ask who owns quarterly or annual maintenance, what data is refreshed, how changes are proposed, and whether there is a governance calendar.

A good vendor should make the operating cadence explicit.

Run a dispute test before signing

Take five historic disputes and recreate them.

Give the vendor the data that existed at the time — not the perfect data you know today. Ask the system to route each case. Then ask the vendor to explain the decision in business language.

Finally, change one fact and see whether the route changes predictably.

That test reveals more than a polished demonstration. A territory-design vendor should help the organization make decisions that are repeatable, explainable, reversible when needed and measurable over time. If the tool cannot do those four things, the map is decoration.

Ask for a live data-lineage demonstration

A vendor should be able to take one territory recommendation and trace it from source data to final assignment.

Choose a difficult account and ask the team to show:

the raw company identity used;
geography and industry fields;
any external market-sizing inputs;
parent-child relationship;
partner-capacity fields;
scoring or rule transformations;
final route;
user overrides after the route.

Then ask what happens if one upstream value is corrected. A dependable system should recompute or flag the affected decision without forcing analysts to rebuild the model manually.

This exercise is more valuable than a long feature list because territory errors are often lineage errors. The map is wrong because a parent was wrong, a ZIP was stale or a partner capability field was copied forward.

Ask who owns model calibration

Optimization tools become stale even when integrations keep running.

Someone must compare predicted opportunity with actual results and decide whether weights still make sense. Ask the vendor whether calibration is:

automatic;
vendor-managed;
customer-managed;
or simply not part of the product.

Then define the review cadence. A quarterly review might compare predicted account potential, actual pipeline, conversion, service burden and exception frequency. If the tool repeatedly labels a territory “balanced” while managers constantly override it, the model is not calibrated to the business.

Make calibration an operating responsibility with an owner, not an optional analytics project.