Intent Audience as a Service: Packages, Costs and Measurement

The short answer: sell a governed audience operating cycle

Package the service around a repeatable cycle:

1. define the client’s ideal customer profile, exclusions, topics, territories, and allowed uses;
2. receive first-party or third-party intent inputs on an agreed cadence;
3. resolve each event to the appropriate entity level and enrich only what the workflow requires;
4. apply fit, recency, confidence, duplication, suppression, and validation rules;
5. deliver approved account audiences, contact research queues, branded topic reports, or downstream actions;
6. capture delivery, match, activation, disposition, and outcome evidence; and
7. refresh, remove, tune, and report under a clear service level.

This structure protects margin because the agency can bound volume, destinations, refresh cadence, manual review, content work, media operations, meetings, and change requests. It protects trust because the client sees why an account qualified, what the signal can and cannot establish, and how records are removed.

A manual or non-intent audience is still the baseline. Firmographic targeting, CRM lists, past engagement, form fills, and human account research can be sufficient for a small market or a stable campaign. Intent audience as a service becomes attractive when timing changes, research happens outside the client’s properties, first-party volume is incomplete, audience freshness matters, and the client lacks the operating team to connect data with action.

Define the deliverable before choosing the data source

Choose one primary decision per package. Common deliverables include:

• Weekly priority delivery: eligible companies showing fresh topic or website activity, with reason codes and removals.
• Paid-media audience service: platform-ready account or contact audiences, exclusions, refresh and suppression, match reporting, and creative recommendations.
• Sales research queue: high-fit accounts with validated context and a human-approved next-step brief.
• Branded intent report: a client-facing view of topic activity, qualified companies, key evidence, limitations, actions, and outcomes.
• Client portfolio monitor: separate rules and reporting for multiple agency clients, with tenant-level permissions and no data leakage.

Do not promise a “list of buyers.” Account-level topic research is not proof that a named person researched, has budget, or intends to purchase. The service should describe qualified signal candidates or priority accounts based on disclosed rules. A person-level record requires its own identity, validation, permission, and use assessment.

Write an acceptance definition. An accepted audience member might need to satisfy target-market fit, a relevant signal within the freshness window, a minimum confidence level, no customer or suppression conflict, successful destination matching, and an explainable reason. Records that fail should be counted and categorized, not silently discarded.

Create packages with entitlements and approval limits

A service can have three operational levels without fixing the agency to arbitrary retail prices:

Report package: configured topics, market filters, branded report, agreed refresh, insights call, and recommendations. The client executes actions. This is the simplest delivery model and a useful pilot.

Audience package: report deliverables plus qualified audience files or destination sync, suppression, refresh/removal rules, match QA, and activation reporting. The agency operates the data flow; the client approves channels and creative.

Managed activation package: audience service plus research briefs, ad or CRM workflows, approved content modules, human review, outcome reconciliation, and optimization. This can create more value but adds labor, platform dependencies, media, privacy obligations, and client approval risk.

For each package, specify included topics, accounts or records, client workspaces, destinations, refreshes, retention, reports, meetings, support, manual reviews, and change requests. State overage units and stop rules. Keep paid media, creative production, outreach tooling, and client-specific integration separate unless intentionally bundled.

Approval limits should be explicit. The agency may automatically refresh an approved low-risk audience while requiring client review for new topics, new sensitive segments, person-level outreach, public claims, material budget changes, or production-system actions.

Use a seven-day pilot to evaluate the operating model

A useful pilot is not a compressed promise of revenue. It proves whether the service can produce a credible audience and client-facing artifact.

Pilot input: one defined offer, target market, exclusion list, a small set of topics, destination preference, and agreed success criteria.

Pilot output: a branded topic report; eligible and rejected account counts; sample reason codes; freshness, identity, and validation fields; recommended audience rules; a proposed activation workflow; and an evidence gap list.

Pilot acceptance: the client can understand why records qualified, the agency can operate within expected labor, the destination or report works, and no mandatory rights or governance requirement is unresolved.

Start by reviewing the MarketPulse demo and plans, starting at $500/mo. Choose a relevant topic and ICP, then evaluate audience fit, available contact fields, and the workflow you will use. The first-purchase 3-day satisfaction guarantee requires no data exports and no more than 2 Intent Scans; review the current guarantee terms before testing. A short evaluation can establish fit, but cannot prove future pipeline or revenue.

Price the service using defined delivery costs

Build a unit-economics model before setting retail price:

Contribution margin = client revenue − data and platform cost − delivery labor − variable activation cost − allocated support and compliance cost.

Model cost drivers separately: wholesale platform scope, topics, accounts or records, clients, workspaces, credits, API and destination usage, identity and validation, enrichment, paid media, creative, outreach tools, analyst and strategist time, manual review, client meetings, reporting, privacy and security work, and expected exceptions.

Compare the complete delivery scope rather than a platform price alone. Obtain written terms for client access, permitted use, topics, refreshes, exports, activation, support, renewal and exit. A software subscription and a managed client service are different purchases.

MarketPulse plans are starting at $500/mo. Starter includes 1 topic, weekly refreshes, and 25,000 exportable contacts per month. Growth includes 3 topics, daily refreshes, and 100,000 exportable contacts. Scale includes 10 topics, daily refreshes, and unlimited contact exports. AIMEE for the initial user is included with Growth and Scale. Compare MarketPulse plans and current offers.

Protect margin with minimum terms where justified by setup effort, prepaid or committed capacity where appropriate, explicit overages, change control, approval deadlines, and a separate pass-through policy for media or third-party tools. Do not sell “unlimited” manual research, destinations, revisions, or client calls unless delivery capacity truly supports it.

Use calculators and operating templates that preserve denominators

The service does not require a complex software model at first. A controlled workbook or database can include:

• a client entitlement and change-request register;
• topic, target-market, exclusion, and approval definitions;
• signal ledger with source, entity, timestamp, confidence, freshness, and reason codes;
• acceptance/rejection log with duplicate, no-fit, stale, identity, and permission reasons;
• destination ledger with delivered, matched, activated, removed, failed, and retried counts;
• labor and exception tracker by client and workflow;
• outcome ledger for responses, accepted meetings, opportunities, stages, pipeline, and revenue;
• invoice, usage, contribution-margin, renewal, and capacity forecast; and
• privacy, security, data-use, retention, deletion, and incident checklist.

Keep client data segregated. Version the rules. Reconcile vendor delivery, agency acceptance, destination match, client action, and outcome as different stages. A delivered record is not automatically an accepted audience member, and a matched account is not automatically activated.

Measure revenue quality and profitability together

Operational measures include eligible-market coverage, signal freshness, accepted-signal rate, rejection reasons, identity confidence, destination match rate, activation latency, removal SLA, delivery failures, manual-review minutes, and client adoption.

Commercial measures include revenue per client, gross and contribution margin, delivery cost, support burden, overages, expansion, churn risk, capacity utilization, and time to recover setup effort.

Revenue-quality measures include positive response, accepted meeting, opportunity creation, stage progression, pipeline, won revenue, sales rejection, unsubscribe, complaint, and client trust feedback. Segment results by client, topic, source, channel, fit, confidence, and audience vintage.

Use a properly designed randomized holdout where feasible to estimate incrementality, with uncertainty and the full relevant costs. A prior-period, phased or matched-cohort comparison is observational unless its design supports stronger inference; report selection, timing and other limitations. Influenced pipeline describes association and does not by itself establish causal ROI.

Renew the service when it continues to improve a recurring decision and maintains target margin. Redesign or reduce scope when one topic, channel, destination, or high-touch process consumes disproportionate labor. Stop when the client cannot act, rights are unresolved, audience quality remains below acceptance, or evidence does not justify cost.

Manage margin, data-use, billing, and client-trust risks

The most common margin failures are underpricing setup, hiding manual cleanup, promising excessive refresh speed, absorbing client-specific integrations, bundling media or creative without limits, and failing to price exception handling. Track minutes and variable costs from the pilot onward.

The most serious trust failures are calling accounts “buyers,” implying person-level behavior from account signals, mixing client data, using stale or disallowed records, ignoring suppression, and making performance claims without evidence. Define source categories, entity level, purpose, client rights, retention, deletion, access, regional rules, sensitive-data exclusions, security, subprocessors, incident handling, and exit.

Review actual contracts, sources, permissions, jurisdictions, client access, retention, suppression and termination with qualified advisers. A service checklist is not a legal determination or independent validation of the records.

Contracts should state who owns raw input, qualified outputs, reports, derived data, configurations, and outcome history; who pays the provider and activation platforms; what happens on nonpayment; what the agency may retain; and how export and deletion work at termination.

Agent-ready audience workflow for Claude, ChatGPT, or AIMEE

Provide the agent with the approved client entitlement, eligible market, exclusions, topics, signal schema, identity thresholds, validation rules, destinations, service levels, message boundaries, metrics, and approval matrix. Instruct it to:

1. segregate every client and reject records that fail fit, freshness, identity, permission, duplication, or suppression rules;
2. create a qualified record with source category, entity level, reason codes, confidence, and expiration;
3. prepare only the approved report, audience, research brief, or action recommendation;
4. route new topics, person-level outreach, production changes, spend, and external claims to human approval;
5. reconcile delivered, accepted, matched, activated, removed, failed, and outcome counts;
6. calculate delivery labor, variable cost, and contribution margin without inventing missing values; and
7. create client reports that separate observed association from incremental evidence and state limitations.

AIMEE can prepare configured work through connected tools in BrandWell. Document the tools, permissions, allowed actions and review points, then check results at the destination. An instruction does not establish an integration or grant client delivery rights.

The service-design rule

Define the recurring decision the client is paying to improve. Document who qualifies, the permitted next action, expiration, observed outcomes and proposed rule changes. Evaluate whether the ongoing service is worth its full cost rather than assuming recurring revenue will follow.

Start by reviewing the MarketPulse demo and plans, starting at $500/mo. Choose a relevant topic and ICP, then evaluate audience fit, available contact fields, and the workflow you will use. The first-purchase 3-day satisfaction guarantee requires no data exports and no more than 2 Intent Scans; review the current guarantee terms before testing. A short evaluation can establish fit, but cannot prove future pipeline or revenue.