AI Marketing Agency Glossary

AI Marketing Agency Glossary. 22 Key Terms Every B2B Executive Should Know

An AI Marketing Agency Glossary is, in B2B marketing, a reference of terms executives use to evaluate AI-enabled agencies and pilot pipeline performance before committing budget.

Most "AI agencies" pitching your team in 2025 are legacy shops with ChatGPT bolted onto the same workflows they ran in 2019. The vocabulary problem is the buying problem: if you cannot name the capability, you cannot inspect it, and if you cannot inspect it, you will buy theater. Forrester's 2024 State of B2B Marketing report finds most B2B marketing leaders now face board-level pressure to prove AI investments generate qualified pipeline within a quarter. Wasted quarters cost credibility with sales, budget with the CFO, and time you do not have.

This hub defines 22 terms across five clusters: Agency Types, Core AI Capabilities, Pipeline and Performance Metrics, Pilot Evaluation Concepts, and Failure Modes. Generic martech glossaries define "AEO" for SEO practitioners and "pipeline" for RevOps analysts. Neither vocabulary helps a CMO decide whether an agency's AI pitch is real. The Starr Conspiracy scoped every definition to the agency-selection decision and the 30 to 90 day pilot that should precede any long retainer.

How to Use This Glossary

Read the cluster that matches your current decision. Shortlisting agencies? Start with Agency Types and Core AI Capabilities. Structuring a pilot contract? Go to Pilot Evaluation Concepts and Pipeline Metrics. Something feels off in a proposal but you cannot articulate why? Failure Modes will give you the language.

Each entry follows the same pattern: a one-sentence capsule, expanded definition, how it works or why it matters, named examples, related terms, FAQs, and a per-term bottom line tied to selection or pilot measurement.

Cluster 1. Agency Types

This cluster defines the firm you are hiring. Get this wrong and every downstream decision inherits the mistake. Salesforce's 2024 State of Marketing report and IBM's 2024 Global AI Adoption Index anchor the capability claims in this section.

AI Marketing Agency

An AI Marketing Agency is, in B2B marketing, a services firm that embeds generative and predictive AI into strategy, content production, media buying, and analytics workflows rather than treating AI as an add-on tool.

Expanded definition. An AI Marketing Agency is a services firm that embeds generative and predictive AI into strategy, content, media, and analytics workflows. Salesforce reports a majority of marketing organizations now use generative AI in some form, but adoption is not capability. Most agencies claiming AI capability have added a chatbot to a legacy production model. A true AI marketing agency has rebuilt pricing, staffing, and turnaround around AI-native delivery.

Why it matters in agency selection. If AI shows up as a line item rather than an operating model, you are paying legacy rates for legacy output.

How to test it. Ask for the reference architecture behind one recent deliverable. Ask what the agency stopped doing when AI came in. Ask how pricing changed. Vague answers are the tell.

Examples. The Starr Conspiracy operates as an AI-native B2B agency. IBM's consulting practice publishes reference architectures for enterprise AI deployment that illustrate what documented AI capability looks like.

Related terms.

• AI Lead Generation Agency
• AI Advertising Agency
• AI-Native Agency
• AI Capability Theater
• Tool-Not-Strategy Misalignment
• Capability Audit
• Reference Architecture

FAQs.

Is every agency using ChatGPT an AI marketing agency? No. Tool use is not delivery-model transformation.

What is the fastest way to disqualify a claim? Request the reference architecture for one paid deliverable.

Does agency age matter? No. Restructured legacy shops can qualify. Founding date is not the test.

Bottom line. Buy the operating model, not the pitch deck. The Starr Conspiracy structures every engagement around AI-native delivery so the pricing reflects the productivity.

AI Lead Generation Agency

An AI Lead Generation Agency is, in B2B marketing, a firm that uses machine learning, intent data, and generative outreach to identify, qualify, and engage B2B prospects at scale.

Expanded definition. An AI Lead Generation Agency uses ML models, intent signals, and generative outreach to build and engage target account lists. Salesforce's data shows lead quality outranks lead volume as the top pipeline concern for B2B marketers. Vendors in this space include Outreach and B2B Rocket, which publish approach documentation on their sites.

Why it matters in agency selection. The failure mode is volume dressed as pipeline. Ask for the qualification logic in writing.

How to test it. Request the scoring model, the disqualification criteria, and the last three months of MQL-to-SQL conversion rates for a comparable client.

Examples. Outreach.io documents AI-assisted sequencing workflows. B2Brocket.ai publishes its automated prospecting model. Both are useful reference points when evaluating a lead-gen agency's stack claims.

Related terms.

• Qualified Pipeline
• MQL-to-SQL Conversion
• Intent Data
• Predictive Audience Modeling
• Vanity Metric Substitution
• Pipeline ROI

FAQs.

Are AI lead gen agencies just outbound shops with better tools? Sometimes. The differentiator is the scoring model and closed-loop measurement.

What is a reasonable output metric? Qualified pipeline dollars, not MQL count.

Can they replace SDRs? They can compress SDR ratios. Full replacement claims are typically theater.

Bottom line. Buy on qualified pipeline output, not lead volume. Ask The Starr Conspiracy about lead-gen agency evaluation criteria before your next RFP.

AI Advertising Agency

An AI Advertising Agency is, in B2B marketing, a media services firm that applies machine learning to audience targeting, creative optimization, bid management, and attribution across paid channels.

Expanded definition. An AI Advertising Agency applies ML to targeting, creative, bidding, and attribution across paid media. Real capability shows up in creative velocity (variants tested per week), audience modeling depth, and closed-loop measurement, not in the tool logos on a capabilities deck. IBM's index finds marketing and advertising remain among the leading functions for enterprise AI deployment.

Why it matters in agency selection. Media budgets scale fast. An AI advertising agency without closed-loop measurement will burn budget faster than a manual team.

How to test it. Ask how the agency ties ad exposure to qualified pipeline. Ask how many creative variants ship per campaign and how variant selection is decided.

Examples. Salesforce's Einstein-powered advertising integrations demonstrate closed-loop measurement architecture at the platform level.

Related terms.

• Generative Creative
• Creative Velocity
• Predictive Audience Modeling
• Attribution Modeling
• Pipeline ROI
• AI Capability Theater

FAQs.

Is programmatic buying the same as AI advertising? No. Programmatic automates bidding. AI advertising also generates and selects creative and models audiences.

What is the top failure signal? Reporting impressions and CTR without pipeline attribution.

Does this replace paid media teams? No. It changes the ratio of strategists to executors.

Bottom line. Demand closed-loop measurement before demand generation. If an agency cannot tie ad spend to qualified pipeline, the AI label is decorative.

AI-Native Agency

An AI-Native Agency is, in B2B marketing, a services firm founded or restructured so AI capabilities are core to the operating model, not layered onto legacy processes.

Expanded definition. An AI-Native Agency is built so AI drives pricing, staffing ratios, and turnaround times. Founding date is not the test. Delivery model is. IBM's index shows organizations that redesign workflows around AI report materially higher productivity than those bolting AI onto existing processes.

Why it matters in agency selection. AI-native economics should show up on the invoice. If they do not, you are subsidizing legacy overhead.

How to test it. Compare the agency's staff-to-output ratio and hourly economics against a legacy peer. Ask how AI changed pricing.

Examples. The Starr Conspiracy operates as an AI-native B2B agency, which shows up in pilot structure and measurement discipline rather than in tool inventories.

Related terms.

• AI Marketing Agency
• Capability Audit
• Reference Architecture
• Tool-Not-Strategy Misalignment
• Pilot Success Criteria
• 30 to 90 Day ROI

FAQs.

Do AI-native agencies charge less? Often per unit of output, not per retainer. Ask for unit economics.

Can a 20-year-old shop be AI-native? Yes, if it restructured. Ask when and how.

What is the fastest disqualifier? Hourly billing tied to legacy staffing pyramids.

Bottom line. AI-native is an economic claim, not a marketing claim. Inspect the invoice logic.

Cluster 2. Core AI Capabilities

This cluster defines what the firm should actually do. Capability claims without measurement are theater.

Answer Engine Optimization (AEO)

Answer Engine Optimization (AEO) is, in B2B marketing, the practice of structuring content so AI-powered answer engines can extract and cite it in generated responses.

Expanded definition. AEO structures content so answer engines can extract and cite it. Gartner predicts a significant share of traditional search volume will shift to AI-powered answer surfaces within the next several years. AEO differs from traditional SEO in three ways: the ranking unit is a passage or entity, not a page; the success signal is citation, not click; and structured data plus semantic clarity outweigh backlinks. Think of AEO as writing for an extractor, not a reader. The engine wants self-contained answers, named entities, and structured markup. Agencies deliver on AEO by defining a query set, a target answer engine set, and a citation-tracking cadence, then producing content patterned for extraction.

Why it matters in agency selection. B2B buyers start research inside answer engines. Without an AEO methodology (query set doc, prompt library, citation tracking sheet), your brand is invisible in the fastest-growing discovery layer.

Examples. IBM publishes structured technical documentation that is regularly cited in AI-generated answers. The Starr Conspiracy uses AEO methodology on this glossary.

Related terms.

• Generative Engine Optimization
• Citation Share
• Entity Density
• Semantic Search
• AI Marketing Agency
• Capability Audit

FAQs.

Is AEO just SEO with a new name? No. Different ranking unit, different success signal, different technical requirements.

What is the primary AEO metric? Citation share within a defined query and engine set.

Do I need new content or restructured content? Usually both. Existing content rarely extracts cleanly.

Bottom line. If your agency cannot define citation share and a query set, they cannot deliver AEO. Ask The Starr Conspiracy for an AEO capability audit before your next content investment.

Generative Engine Optimization (GEO)

Generative Engine Optimization (GEO) is, in B2B marketing, the practice of optimizing content for inclusion and favorable representation in AI-generated responses across generative surfaces.

Expanded definition. GEO is often used interchangeably with AEO. Some practitioners scope GEO more broadly to include any generative surface and reserve AEO for question-answering engines specifically. Gartner's research on generative AI in marketing shows brand representation inside generated answers is now a tracked metric in leading B2B programs. In agency contexts, treat GEO and AEO as adjacent and ask what the firm actually measures. GEO programs manage entity density, source authority, and structured data so the model has strong signal to cite the brand favorably. Measurement combines citation frequency, source attribution, and brand mention sentiment inside generated answers.

Why it matters in agency selection. Terminology drift lets agencies claim a capability without measurement. Force specificity.

Examples. The Starr Conspiracy publishes GEO methodology tied to citation share benchmarks.

Related terms.

• Answer Engine Optimization
• Citation Share
• Semantic Search
• Entity Density
• AI Capability Theater

FAQs.

Is GEO different from AEO in practice? Marginally. Ask what the agency measures, not what they call it.

What is the primary output? Cited brand mentions inside generated answers.

How long until results? Citation share is measurable within 30 to 90 days on a defined query set.

Bottom line. GEO and AEO are the same buying decision. Measurement discipline is the differentiator.

Generative Creative

Generative Creative is, in B2B marketing, the production of ad copy, images, video variants, and landing page content using generative AI models at high velocity for testing.

Expanded definition. Generative Creative uses AI models to produce variant volume for testing. Salesforce data shows generative AI adoption for creative production is now widespread among marketing teams. The value is in variant volume that fuels a testing framework, not in replacing strategy. The workflow generates variants, ships them into a structured test, measures against a pipeline-tied metric, and feeds winners back into the model. Without the test infrastructure, generative creative is decorative.

Why it matters in agency selection. An agency generating 200 headline variants without a testing framework is producing noise, not performance.

Examples. The Starr Conspiracy pairs generative creative production with a defined testing cadence and pipeline attribution.

Related terms.

• Creative Velocity
• Predictive Audience Modeling
• Attribution Modeling
• AI Advertising Agency
• AI Capability Theater
• Vanity Metric Substitution

FAQs.

Will AI replace creative teams? No. It changes the ratio of strategists to producers.

What is the failure signal? Variant volume without a testing framework.

Do I need a new martech stack? Usually the testing infrastructure needs upgrading, not the creative tools.

Bottom line. Variants without tests are noise. Buy the testing discipline first.

Creative Velocity

Creative Velocity is, in B2B marketing, the rate at which an agency produces testable creative variants over a defined period, typically measured per week or per campaign.

Expanded definition. Creative Velocity measures how quickly an agency can produce testable variants. IBM's index shows AI-enabled creative workflows meaningfully compress variant production timelines compared with manual workflows. Velocity is a useful capability signal only when paired with a testing framework and pipeline attribution. Measure variants shipped per campaign, variants tested per week, and turnaround time from brief to first draft. Compare against a legacy peer to see whether the AI claim shows up in the numbers.

Why it matters in agency selection. Velocity is a proxy for AI-native production. Legacy shops cannot fake it under audit.

Examples. The Starr Conspiracy tracks creative velocity as a standard pilot metric.

Related terms.

• Generative Creative
• AI Advertising Agency
• Pilot Success Criteria
• Attribution Modeling
• Capability Audit

FAQs.

Is higher velocity always better? Only when paired with a testing framework.

How do I benchmark velocity? Compare variants per campaign against a legacy peer.

Is velocity a contract term? It should be, alongside quality gates.

Bottom line. Velocity without tests is waste. Contract for both.

Predictive Audience Modeling

Predictive Audience Modeling is, in B2B marketing, the use of machine learning to identify look-alike, in-market, or high-propensity audiences from first-party and third-party signals.

Expanded definition. Predictive Audience Modeling scores accounts and individuals for likelihood to buy. Salesforce reports predictive modeling is among the most-adopted AI use cases in B2B marketing. Model quality depends on training data volume, signal freshness, and retraining cadence. The model ingests firmographic, behavioral, and intent signal, scores accounts, and outputs a ranked target list. Freshness matters: models trained on stale data misfire on current market conditions.

Why it matters in agency selection. Bad model, wasted spend. Ask for the data sources and retraining schedule in writing.

Examples. Salesforce and M1 Project publish approach documentation for predictive audience modeling that is useful reference material.

Related terms.

• Intent Data
• Account-Based Marketing
• AI Lead Generation Agency
• Lead Scoring
• Qualified Pipeline

FAQs.

What data does the model need? First-party CRM data plus at least one intent signal.

How often should it retrain? Cadence depends on market volatility. Quarterly at minimum.

Can I audit the model? Yes. Ask for feature importance and validation results.

Bottom line. Audit the model before you buy the audience.

Intent Data

Intent Data is, in B2B marketing, third-party or first-party behavioral signal indicating a buying committee is researching a category, competitor, or solution.

Expanded definition. Intent Data is behavioral signal indicating active research. Forrester's research on B2B buying behavior finds buying committees consume most of their research content before contacting a vendor, which makes intent signal the primary early indicator. Intent is a probability input, not a ready-to-call list. Providers aggregate third-party publisher signal or first-party site behavior into topic-level scores. The agency filters by fit criteria, sequences outreach, and measures conversion to qualified pipeline.

Why it matters in agency selection. Agencies that treat every intent spike as a qualified lead are optimizing for activity, not pipeline.

Examples. B2Brocket.ai and Outreach.io publish intent-data-driven outreach methodologies.

Related terms.

• Predictive Audience Modeling
• Qualified Pipeline
• Account-Based Marketing
• Lead Scoring
• Vanity Metric Substitution

FAQs.

Is intent data worth the cost? Only if paired with disciplined qualification.

What is the top misuse? Treating spikes as sales-ready leads.

How is intent scored? By topic engagement volume and recency, weighted by fit.

Bottom line. Intent is an input to qualification, not a substitute for it.

Semantic Search

Semantic Search is, in B2B marketing, a retrieval approach that matches queries to content based on meaning and entity relationships rather than exact keyword match.

Expanded definition. Semantic Search uses embeddings and entity graphs to match queries to meaning. IBM's research on enterprise search finds semantic approaches materially improve retrieval accuracy over keyword-only systems. AEO and GEO programs depend on semantic infrastructure at the engine level. Content is embedded into vector space, entities are tagged and linked, and queries are matched by proximity in meaning rather than lexical overlap.

Why it matters in agency selection. If your agency talks about keywords without talking about entities and embeddings, they are optimizing for a discovery layer that is losing share.

Examples. IBM Watson Discovery documents semantic search implementation for enterprise content.

Related terms.

• Answer Engine Optimization
• Generative Engine Optimization
• Entity Density
• Citation Share

FAQs.

Is semantic search replacing keyword search? In discovery, yes. In some technical use cases, no.

What does it mean for content strategy? Write for entities and meaning, not keyword density.

Do I need new tooling? Usually a content restructuring, not a tool swap.

Bottom line. Keyword-only strategies age out. Buy for semantic and entity discipline.

Entity Density

Entity Density is, in B2B marketing, the concentration of named entities per unit of content that helps AI engines identify, disambiguate, and cite a source.

Expanded definition. Entity Density measures how many named entities per unit of content give retrieval systems clear signal. Gartner's research on generative AI shows entity clarity is a leading factor in whether generative engines cite a given source. Higher entity density, when accurate and relevant, correlates with citation likelihood. Editors identify canonical entities (brands, people, products, concepts), use consistent term strings, and link entities to authoritative references so engines can disambiguate.

Why it matters in agency selection. Content without entities is content without signal. Ask how the agency structures for entity extraction.

Examples. The Starr Conspiracy structures glossary entries for high entity density and consistent term strings.

Related terms.

• Answer Engine Optimization
• Generative Engine Optimization
• Citation Share
• Semantic Search

FAQs.

Can entity density be too high? Yes, when it degrades readability.

Does it replace keywords? It supplements them for generative surfaces.

How is it measured? Entities per 100 words, tracked against citation outcomes.

Bottom line. Write for extractors as well as readers. Density with accuracy earns citations.

Cluster 3. Pipeline and Performance Metrics

This cluster defines how you measure whether it worked. Metrics without shared definitions are how partnerships die in month four. Forrester's 2024 State of B2B Marketing and Salesforce's 2024 State of Marketing anchor the benchmarks below.

Qualified Pipeline

Qualified Pipeline is, in B2B marketing, the dollar value of open opportunities that meet a jointly agreed fit and intent threshold, tracked as the primary output metric of B2B demand programs.

Expanded definition. Qualified Pipeline is the dollar value of opportunities meeting a shared fit and intent threshold. Forrester's research shows alignment on qualification criteria is one of the strongest predictors of marketing-sourced revenue. The definition must be shared between marketing, sales, and the agency before the pilot starts. Marketing and sales define fit criteria (firmographics, ICP fit) and intent criteria (behavior thresholds). Agency-sourced opportunities meeting both are counted at deal value.

Why it matters in agency selection. Ambiguity here is where most agency partnerships fail. Agree in writing or plan to argue in month four.

Examples. The Starr Conspiracy defines qualified pipeline criteria as part of every pilot SOW.

Related terms.

• MQL-to-SQL Conversion
• Pipeline ROI
• 30 to 90 Day ROI
• Lead Scoring
• Attribution Modeling
• Pilot Success Criteria

FAQs.

Who owns the definition? Marketing, sales, and the agency jointly, in writing.

Is deal stage part of the definition? Usually. Common threshold is Stage 2 or "sales-accepted."

What if sales won't agree? Do not start the pilot until they do.

Bottom line. No shared definition, no pilot. Get it in the SOW.

MQL-to-SQL Conversion

MQL-to-SQL Conversion is, in B2B marketing, the percentage of marketing-qualified leads that sales accepts as sales-qualified opportunities within a defined window.

Expanded definition. MQL-to-SQL Conversion measures how many marketing-qualified leads convert to sales-qualified opportunities. Forrester's B2B benchmarks show conversion rates vary widely by segment, and disciplined qualification is the primary driver of higher conversion.

Formula. (SQLs Accepted / MQLs Delivered) x 100. Variables: SQLs Accepted is the count of leads sales formally accepts within the window; MQLs Delivered is the count of leads marketing hands off in the same window.

Worked example. An agency delivers 400 MQLs in a quarter. Sales accepts 92 as SQLs. Conversion rate is (92 / 400) x 100 = 23%.

Why it matters in agency selection. Low conversion usually signals lead quality, not sales execution. This is the metric that reveals whether an agency is sourcing pipeline or manufacturing MQLs.

Examples. The Starr Conspiracy tracks MQL-to-SQL conversion as a standard pilot metric.

Related terms.

• Qualified Pipeline
• Lead Scoring
• Vanity Metric Substitution
• Pipeline ROI
• Attribution Modeling

FAQs.

What is a healthy rate? Segment-dependent. Track trend against your baseline, not against a public benchmark.

Who defines an SQL? Sales, formally, before the pilot starts.

What does a falling rate mean? Usually lead quality or ICP drift.

Bottom line. Track conversion, not volume. Volume without conversion is a vanity metric.

Pipeline ROI

Pipeline ROI is, in B2B marketing, the ratio of qualified pipeline generated to agency and program cost, typically expressed as a multiple.

Expanded definition. Pipeline ROI expresses qualified pipeline as a multiple of program cost. Forrester's benchmarks show pipeline-to-cost ratios vary by industry and program maturity, so use your baseline as the reference, not a published number.

Formula. Qualified Pipeline Generated / Total Program Cost. Variables: Qualified Pipeline Generated is the dollar value of opportunities meeting the shared qualification threshold; Total Program Cost includes agency fees, media spend, and tooling.

Worked example. Agency generates $2,000,000 in qualified pipeline against $400,000 in total program cost. Pipeline ROI is 5.0x.

Why it matters in agency selection. Pipeline ROI measures efficiency of the demand program. It does not measure revenue ROI, which requires close rates and average contract value.

Examples. The Starr Conspiracy reports Pipeline ROI as a standard pilot output.

Related terms.

• Qualified Pipeline
• 30 to 90 Day ROI
• CAC Efficiency
• Attribution Modeling
• Pilot Success Criteria

FAQs.

What ratio should I target? Your improved baseline, not a public benchmark.

Does it include media spend? Include all program cost the agency influences.

Is Pipeline ROI the same as revenue ROI? No. Complete with close rate and ACV.

Bottom line. Pipeline ROI is a pilot metric. Revenue ROI is the year-one goal.

30 to 90 Day ROI

30 to 90 Day ROI is, in B2B marketing, the measurable pipeline or revenue return produced within the first 30 to 90 days of an agency partnership, used as the primary pilot success criterion.

Expanded definition. 30 to 90 Day ROI is the pipeline or revenue return within the pilot window. Forrester's research shows executive tolerance for extended agency ramp periods has fallen significantly under current budget scrutiny. The window is deliberately short. It filters agencies that need two quarters of strategy before producing measurable output. Define pipeline and citation targets in the SOW, measure weekly, review at day 30, 60, and 90 against contracted thresholds.

Why it matters in agency selection. Before you sign a 6-month retainer, run a 30 to 90 day pilot. Executives cannot wait two quarters to know if a partnership is working.

Examples. The Starr Conspiracy structures every engagement around a 30 to 90 day pilot with contracted success criteria.

Related terms.

• Pilot Success Criteria
• Pipeline ROI
• Qualified Pipeline
• Capability Audit
• Reference Architecture

FAQs.

Is 30 days too short? For pipeline sourced from cold starts, sometimes. For citation share and lead quality signals, no.

What if the pilot fails? You have lost 90 days and learned the delivery model. That is the point.

Can pilots convert to retainers? Yes, on evidence, not on hope.

Bottom line. Pilot first, retainer second. Ask The Starr Conspiracy for a 30 to 90 day pilot scope and success criteria.

Citation Share

Citation Share is, in B2B marketing, the percentage of AI-generated answers within a defined topic set that cite your brand as a source, used as the primary output metric for AEO and GEO programs.

Expanded definition. Citation Share measures how often your brand appears in generated answers across a defined query set. Gartner's research shows brand representation inside AI-generated answers is now a tracked KPI in leading B2B programs.

Formula. (Answers Citing Brand / Total Answers in Query Set) x 100. Variables: Answers Citing Brand is the count of generated answers naming or linking to the brand; Total Answers in Query Set is the count of answers produced across the defined query and engine set.

Worked example. Query set of 100 prompts across three engines yields 300 answers. Brand is cited in 45. Citation Share is (45 / 300) x 100 = 15%.

Required inputs.

• A defined query set (typically 50 to 200 prompts)
• A defined answer engine set
• A defined tracking cadence

Why it matters in agency selection. Without those three inputs, citation share is unmeasurable and any claim is theater.

Examples. The Starr Conspiracy uses Citation Share as a primary AEO pilot metric.

Related terms.

• Answer Engine Optimization
• Generative Engine Optimization
• Entity Density
• Semantic Search

FAQs.

What is a good citation share? Depends on category competition. Track your trend.

How often should it be measured? Weekly during pilot, monthly at steady state.

Which engines count? The ones your buyers use. Define in the SOW.

Bottom line. Define the query set or the metric is fiction.

CAC Efficiency

CAC Efficiency is, in B2B marketing, the ratio of customer lifetime value or annual contract value to customer acquisition cost, used to measure the sustainability of demand programs.

Expanded definition. CAC Efficiency compares the value of a customer against the cost to acquire them. Salesforce reports CAC discipline is now a board-level metric in most B2B tech organizations.

Formula. LTV / CAC, or ACV / CAC for shorter-horizon views. Variables: LTV is customer lifetime value; ACV is annual contract value; CAC is fully loaded acquisition cost including agency fees, media, and internal cost.

Worked example. ACV of $60,000 against fully loaded CAC of $20,000 yields ACV / CAC of 3.0x.

Why it matters in agency selection. CAC Efficiency reveals whether an agency is sourcing profitable customers or just closable ones.

Examples. The Starr Conspiracy tracks CAC Efficiency alongside Pipeline ROI in pilot reporting.

Related terms.

• Pipeline ROI
• Qualified Pipeline
• Attribution Modeling
• Lead Scoring

FAQs.

Should agencies own CAC? They should influence it and report on it.

What is a healthy ratio? Category-dependent. Track your baseline.

What is the fastest way to improve CAC? Improve qualification, not volume.

Bottom line. Sustainable growth is a CAC question, not a volume question.

Cluster 4. Pilot Evaluation Concepts

This cluster defines how you structure the first 90 days so the answer is unambiguous. Forrester's 2024 research and IBM's 2024 Global AI Adoption Index anchor the claims below.

Pilot Success Criteria

Pilot Success Criteria are, in B2B marketing, the specific, measurable outcomes agreed before an agency pilot begins that determine whether the partnership expands, restructures, or ends.

Expanded definition. Pilot Success Criteria are contracted, measurable outcomes agreed before kickoff. Forrester finds pilots with pre-agreed numeric success criteria are materially more likely to convert to productive long-term engagements. Marketing, sales, and the agency co-author criteria in the SOW, define measurement cadence, and set expand/restructure/end triggers at day 30, 60, and 90.

Why it matters in agency selection. Good criteria are numeric, time-bound, and tied to pipeline or revenue. Bad criteria are qualitative ("improve brand awareness") or activity-based ("launch 12 campaigns").

Examples. The Starr Conspiracy co-authors Pilot Success Criteria as part of every pilot SOW.

Related terms.

• 30 to 90 Day ROI
• Qualified Pipeline
• Vanity Metric Substitution
• Capability Audit
• Attribution Modeling
• Pipeline ROI

FAQs.

Who writes the criteria? Jointly, before kickoff.

What if the agency resists numeric criteria? Do not sign unless they will commit to numeric criteria within a defined revision window.

How many criteria are enough? Three to five, tied to pipeline and citation.

Bottom line. No numeric criteria, no pilot.

Capability Audit

A Capability Audit is, in B2B marketing, a structured evaluation of an agency's actual AI infrastructure, workflows, and delivery model conducted before contract signing.

Expanded definition. A Capability Audit inspects tools in use, prompt libraries, quality-control processes, and staff-to-output ratios. IBM's index shows the gap between claimed and operationalized AI capability remains one of the largest sources of vendor selection error.

Required inputs.

• Tool inventory with usage evidence
• Prompt libraries and QA processes
• Staff-to-output ratios
• Reference architecture for one representative deliverable

Why it matters in agency selection. Agencies that resist a capability audit are usually hiding AI capability theater.

How it works. Send the audit questions ahead of the finalist round. Score answers against evidence, not claims. Require live walk-through of one recent deliverable.

Examples. The Starr Conspiracy welcomes capability audits as part of the finalist round.

Related terms.

• AI Capability Theater
• Tool-Not-Strategy Misalignment
• Pilot Success Criteria
• Reference Architecture
• AI-Native Agency

FAQs.

Is a capability audit standard? It should be. Most RFPs skip it.

How long does it take? One to two weeks with an engaged agency.

What is the top red flag? Refusal or vague answers on staffing ratios.

Bottom line. Audit before you sign. Ask The Starr Conspiracy for a capability audit question set.

Reference Architecture

A Reference Architecture is, in B2B marketing, a documented map of how an agency's tools, data sources, and workflows connect to produce a specific outcome.

Expanded definition. A Reference Architecture documents the tools, data flows, and workflow steps that produce a defined outcome. IBM's research on enterprise AI shows documented reference architectures correlate strongly with successful deployment outcomes. Think of it as the wiring diagram behind the demo. Anyone can film a demo. Only teams that actually built the workflow can draw the wiring.

Why it matters in agency selection. Executives should request the reference architecture for the exact outcome the pilot will measure. Vague answers indicate the workflow does not exist yet.

What to ask for. A diagram or documentation showing inputs, tools, human decision points, and outputs for one pilot-relevant outcome.

Examples. IBM publishes reference architectures for enterprise AI patterns. The Starr Conspiracy publishes reference architectures for pilot outcomes on request.

Related terms.

• Capability Audit
• AI-Native Agency
• Tool-Not-Strategy Misalignment
• Pilot Success Criteria

FAQs.

Is a slide deck enough? No. Ask for a diagram plus documentation.

Who at the agency should present it? Delivery lead, not sales.

What if they cannot produce one within 48 hours of the request? Treat as a red flag requiring escalation, and disqualify if they cannot produce one before the finalist round closes.

Bottom line. No reference architecture, no pilot.

Attribution Modeling

Attribution Modeling is, in B2B marketing, the methodology used to assign pipeline or revenue credit across marketing touchpoints, ranging from single-touch to multi-touch and algorithmic models.

Expanded definition. Attribution Modeling assigns credit for pipeline and revenue across touchpoints. Forrester's benchmarks show attribution disagreements are among the most common sources of marketing-sales friction. Agreement on the attribution model must exist before the pilot. Think of attribution as the box score. Argue about the scoring rules before the game, not after. Choose a model (first-touch, last-touch, multi-touch, algorithmic), define the touchpoints in scope, and document the reporting cadence.

Why it matters in agency selection. Post-hoc attribution debates are how agencies win arguments and lose contracts.

Examples. Salesforce publishes attribution modeling documentation across its analytics products. The Starr Conspiracy contracts the attribution model in the pilot SOW.

Related terms.

• Pipeline ROI
• Qualified Pipeline
• MQL-to-SQL Conversion
• CAC Efficiency
• Pilot Success Criteria

FAQs.

Which model is best? The one both sides commit to before the pilot.

Can we change models mid-pilot? Only with mutual agreement in writing.

Do we need algorithmic attribution? Not for a pilot. Multi-touch usually suffices.

Bottom line. Contract the model. Do not argue it later.

Lead Scoring

Lead Scoring is, in B2B marketing, the rule-based or model-based system that assigns a numeric value to leads based on fit and behavior to prioritize sales action.

Expanded definition. Lead Scoring prioritizes leads by fit (firmographics, ICP) and behavior (engagement signals). Salesforce reports AI-assisted lead scoring is one of the fastest-growing applied use cases in B2B marketing. Fit and behavior signals are weighted, summed, and thresholded. Leads above threshold become MQLs. Model quality is validated against SQL conversion.

Why it matters in agency selection. Scoring is where an agency's qualification discipline is visible. Ask for the model in writing.

Examples. The Starr Conspiracy documents scoring logic as part of pilot handoff.

Related terms.

• MQL-to-SQL Conversion
• Qualified Pipeline
• Intent Data
• Predictive Audience Modeling
• Account-Based Marketing

FAQs.

Should scoring be rule-based or model-based? Both, layered.

Who owns the scoring model? Marketing operations, with sales input.

How often should it be retuned? Quarterly, validated against SQL conversion.

Bottom line. No documented scoring, no defensible qualification.

Account-Based Marketing

Account-Based Marketing (ABM) is, in B2B marketing, a strategy that treats individual target accounts as markets of one, coordinating marketing and sales action against a defined account list.

Expanded definition. Account-Based Marketing focuses program investment on a defined target account list. Forrester's research shows ABM programs with disciplined account selection and shared sales-marketing execution consistently outperform broad demand programs on pipeline efficiency. Define target accounts, personalize outreach and content at account level, coordinate sales and marketing plays, and measure account engagement and pipeline.

Why it matters in agency selection. Ask how the agency selects accounts, coordinates with sales, and measures account-level engagement, not just lead volume.

Examples. Salesforce documents ABM workflow patterns across its marketing cloud products.

Related terms.

• Predictive Audience Modeling
• Intent Data
• Lead Scoring
• Qualified Pipeline
• Pipeline ROI

FAQs.

Is ABM different from targeted demand gen? Yes, in coordination and personalization depth.

How many accounts is reasonable? Program-dependent, from dozens to low thousands.

Should ABM live in marketing or sales? Jointly.

Bottom line. ABM is a coordination discipline, not a channel.

Cluster 5. Failure Modes

This cluster gives you the vocabulary to reject bad-fit proposals before they become expensive mistakes.

AI Capability Theater

AI Capability Theater is, in B2B marketing, the practice of marketing AI capabilities that do not exist in the agency's actual delivery model.

Expanded definition. AI Capability Theater is claimed capability without operational substance.