GTM questions, unpacked: Ask a better
growth question.

GTM Unpacked is a research publication about how companies acquire, convert, retain, and expand customers. It combines detailed company cases, shorter growth explainers, cross-company topics, people profiles, a glossary, and practical questions. Source-backed statements, company-reported claims, analytical inferences, and unresolved evidence gaps are distinguished throughout.

[Inference] Start with a company whose buyer and product resemble your own, read its growth post, and then open the full case. Compare the buying process and economics before borrowing a tactic. The company index and comparison tool help locate relevant cases; the glossary explains unfamiliar terminology.

The company case is the fuller research record, including available sources, people, economics, historical evidence, critique, and gaps. Its companion blog post is an approximately 600-word editorial explanation of the GTM story. Each links to the other so readers can move between the concise narrative and its supporting context.

No. The corpus contains extended cases, source-backed briefs, and partial reconstructions. A longer blog post does not upgrade a company's evidence status. Some records establish substantial history; others support only a narrow product, founder, or distribution account. The case's status, sources, and gaps explain that difference.

No. A first paying customer is named only where the underlying record establishes one. Public launches, waitlists, GitHub stars, design partners, free users, and pilot activity are different evidence. When a first contract or founder closing sequence is missing, the post preserves that uncertainty rather than supplying an anecdote.

An Inference label identifies interpretation, such as a motion classification, an explanation of why a tactic may have worked, or a proposed improvement. It is not presented as a directly verified historical fact. Read the surrounding source-backed observations and consider whether other explanations could also fit the evidence.

An Unverified label identifies a claim, detail, or question that the available research does not establish. It can describe missing customer names, economics, historical periods, or internal processes. The label is a boundary on the account, not evidence that the unknown event did or did not occur.

No. The site organizes companies for comparison and learning, not as an investment or quality ranking. Different buyer groups, market conditions, costs, and capital requirements make a single GTM score misleading. The comparison tool surfaces those differences without assigning a universal winner.

Use the topic and pattern pages, then inspect the company's actual conversion process and segment. A shared label is a starting point for comparison. Product-led entry, enterprise selling, and partner involvement can coexist within one company, and the mix can change over time.

[Inference] Use them as hypotheses to test, rather than instructions to copy unchanged. A tactic can depend on founder access, timing, brand, capital, customer concentration, or a narrow market. Look for the conditions that made the mechanism plausible and decide what evidence would support it in your own setting.

The site uses the book's framework to examine early founder selling, experiments, motion choices, lead generation, conversion, retention, team building, technology, and metrics. It applies that lens retrospectively to company evidence. This does not mean the companies themselves followed the book or used its terminology.

The methodology page explains the framework's educational context and links to official MIT course and Martin Trust Center resources. It also identifies The Revenue Engine and its authors. That context is separate from the evidence supporting individual company stories; institutional references do not verify a company's growth claims.

The framework source used by this edition identifies Emmanuelle Skala, Adam Blake, and Max Faingezicht as the authors of Disciplined Entrepreneurship: The Revenue Engine. The methodology page links to the official book page and explains the supplied manuscript edition used for the site's reading map.

No. A framework organizes questions and comparisons; it does not create missing evidence or establish causality. A founder account, launch sequence, or observed pricing change may support a plausible explanation while leaving alternatives unresolved. The site labels those explanations as inference and records gaps separately.

A channel describes how a prospect reaches an offer, while the motion concerns how the purchase happens. A community can introduce a user who later buys through self-service or an enterprise sales process. Treating the community alone as the motion hides the work required to convert interest into payment.

Founder interviews provide direct accounts of decisions, motivations, and reported results. They can also be retrospective, selective, or incomplete. The site attributes those accounts and avoids treating every recollection as independently confirmed. Contemporaneous product pages, filings, customer accounts, and other records can provide additional context.

No. A company announcement establishes what the company reported, subject to the statement's context and date. It does not automatically establish audited financial performance, representative customer outcomes, or a causal explanation. Company-reported claims should remain attributed even when the announcement is a primary source.

Dated captures can show positioning, packages, calls to action, and commercial routes visible at a particular time. A current page cannot establish what the company offered years earlier. Archive gaps, partial rendering, and missing periods limit the conclusions that can be drawn from those captures.

Those measures often are not disclosed with enough detail to verify the relevant cohort, period, or calculation. The site can discuss cost structure and possible expansion mechanisms without inventing a ratio. Missing metrics are important because revenue growth alone cannot establish efficient acquisition or durable retention.

A citation points to material supporting the nearby statement within the scope described by the case. It does not imply that the source verifies every sentence on the page. Check attribution, dates, the distinction between reported claims and analysis, and whether the underlying material directly supports the proposition.

[Inference] Look for a specific group with a consequential problem, reachable buyers, a feasible buying process, and a solution you can deliver. State why this group is more promising than adjacent segments. Test those assumptions through actual behavior and commitments rather than selecting a segment only because its total market looks large.

[Inference] Narrow enough to change whom you contact, which problem you discuss, and what you decline to build or sell. Include observable fit criteria and exclusions. If two supposed ideal customers need different products, buying processes, or economics, separate them for testing instead of hiding the difference in a broad persona.

[Inference] Start with credible access to a tightly defined problem: existing professional relationships, relevant communities, direct outreach, or people already seeking a solution. The objective is not just a meeting. Learn the customer's workflow, decision process, and willingness to commit, then record how interest becomes a paid relationship.

[Inference] Establish a working account of who buys, what triggers evaluation, why customers choose the offer, which objections matter, and what implementation requires. The process need not be fully optimized. It should be clear enough that a new seller can test and improve it without relying entirely on the founder's personal reputation.

Not necessarily. A design partner may provide feedback, access, or collaboration without paying. A paying design partner provides stronger commercial evidence, but an unusually customized engagement may still differ from a repeatable sale. State the payment, scope, and commitment before treating participation as customer acquisition.

[Inference] Evidence becomes more informative when the customer accepts meaningful effort, risk, or payment in pursuit of a specific outcome. Relevant examples include completing an evaluation, using the product in real work, paying, renewing, or expanding. Each answers a different question; none should be substituted for the others.

Identify who performs the workflow, who experiences the problem, who controls the budget, and who must approve the change. These can be different people. Mapping the decision-making unit makes it easier to understand why user enthusiasm may not translate directly into a purchase or production deployment.

[Inference] Ask about specific recent situations, existing alternatives, consequences, and decisions rather than inviting compliments about your idea. Look for observable behavior and conflicting evidence. A useful interview can show that the problem is unimportant or that the proposed buyer lacks authority; learning that early can change your next test.

[Inference] Consider expansion when there is evidence of repeatable value and a credible way to serve the adjacent segment without losing focus. Check whether the new buyer needs different capabilities, support, pricing, or approvals. Shared vocabulary or superficial product similarity does not establish that the original motion will transfer.

Yes, early access and personal trust can make initial conversations and commitments easier than they will be for later sellers. That does not make early customers unhelpful. It means the team should test which parts of the sale survive without the founder's relationship, reputation, or unusual level of involvement.

[Inference] Compare the customer's ability to experience value independently with the complexity, risk, and economics of buying. A self-serve path may suit an individual use case while the same product needs human assistance for enterprise approval. Test the route for a particular segment instead of selecting a company-wide label first.

Yes. Individuals or teams can adopt through a product while larger organizations purchase governance, support, deployment, or broader access through a sales process. The useful question is which route converts each segment and where the handoff occurs. A free product does not make every later contract product-led.

No. A free tier can educate users, create leads, or support a later human-led evaluation. To establish product-led conversion, examine how customers actually move from use to payment and which decisions happen without required seller involvement. Also consider the cost of serving free users and the quality of paid retention.

[Inference] Human involvement is easier to justify when it materially improves qualification, confidence, evaluation, or implementation and the account's contribution can support that effort. Revenue per contract is only one input. Delivery costs, sales duration, renewal behavior, and the amount of customization can change the economic answer.

Inside sales relies primarily on remote engagement; field sales includes substantial in-person work. Both are forms of human-led selling and can coexist. The right format depends on how customers work and buy, the importance of onsite evaluation, and the cost and value of the engagement.

No. A partnership can concern technology, supply, marketing, referrals, delivery, or actual selling. To classify a partner-led conversion route, look for evidence that the partner participates in the commercial decision or transaction. An integration logo or joint press release does not establish partner-sourced customers or revenue.

Community can support discovery, learning, trust, adoption, and retention. It does not by itself explain how a buyer commits payment. A community-supported company can convert through self-service, sellers, marketing, or partners. Describe both the community's role and the purchasing mechanism rather than treating the labels as interchangeable.

In a less established category, customers may need help recognizing the problem, assigning a budget, and understanding a credible evaluation. In an established category, they may compare alternatives using familiar criteria. These conditions can vary by segment, so the relevant question is the buyer's situation rather than the age of the technology alone.

Yes. An easy interface can still trigger security review, procurement, legal terms, integration work, or organizational change. Product complexity concerns using and delivering the solution; deal complexity concerns deciding and committing to buy it. Treating them separately helps explain why a straightforward demo may still lead to a long sales cycle.

[Inference] Revisit the motion when evidence shows a mismatch among buyer behavior, conversion friction, costs, or the segment being pursued. A change should have a specific hypothesis and measurable outcomes. Moving upmarket or adding sellers is not automatically an improvement if the new process increases complexity without enough customer value.

[Inference] Compare channels using consistent definitions of customer quality, acquisition cost, conversion timing, and retained contribution. Include the work and costs required after the first lead. A cheap signup channel can be expensive when its users rarely activate or renew, while a costly channel may reach more suitable accounts.

[Inference] A relevant message connects a credible customer situation to a useful reason to engage. It should reflect the recipient's role and likely problem without pretending to know private facts. The test is whether the intended buyer recognizes the issue and takes a meaningful next step, not merely whether the message receives opens.

[Inference] Start with the job the content is meant to do: introduce a problem, answer an evaluation question, support implementation, or retain users. Then measure relevant behavior and downstream quality. Page views can describe reach, but they do not alone establish incremental pipeline, paid conversion, or customer value.

A marketing-qualified lead meets agreed criteria based on marketing information or engagement. A product-qualified lead meets criteria based on product behavior or experienced value. Neither label is universal. Both need definitions and validation against the outcomes the company wants them to predict.

[Inference] Choose a behavior that gives useful information about customer value or buying readiness, such as repeated completion of an important workflow or meaningful team adoption. Validate the signal against later outcomes. An easily measured event can be operationally convenient without being commercially informative.

[Inference] Define the fit, need, buying process, and next-step evidence required to advance an opportunity. Review both wins and disqualifications to see whether the criteria are useful. Qualification should help stop poor-fit work as well as prioritize promising accounts; otherwise the pipeline can grow without becoming more credible.

[Inference] A demo should help the relevant buyer evaluate whether the offer can address an important workflow or outcome. Connect the demonstration to the customer's situation and identify what remains unproven. A memorable presentation is not the same as agreement on technical fit, business value, or a purchase process.

[Inference] Specify the customer problem, participants, scope, duration, success evidence, cost, and decision that follows the pilot. Also identify who will own production implementation if it succeeds. Without those agreements, a technically successful pilot can remain disconnected from budget approval or sustained operational use.

No. Payment is valuable commercial evidence, but one pilot may reflect a special relationship, custom work, or a narrow experiment. Product-market fit needs broader evidence of repeated value and demand in a defined market. Examine conversion to production, delivery effort, renewal, and the similarity of subsequent customers.

A demonstration may establish interest without resolving urgency, budget, ownership, implementation risk, or approval requirements. Those are different questions. [Inference] Diagnose the missing decision and stakeholder rather than assuming that another demonstration or a larger discount will move the opportunity forward.

[Inference] Help the champion articulate the problem, relevant value, evidence, risks, and next steps to other stakeholders. Learn what the champion needs to navigate internally and verify their influence through actual actions. Do not assume an enthusiastic user controls budget or can obtain approvals alone.

[Inference] A useful experiment tests an important uncertainty with a defined audience, intervention, outcome, and decision rule. For example, test whether a narrower use-case message produces qualified evaluations in one segment. Record the costs and counterevidence, then decide what to repeat, change, or stop.

[Inference] Identify the outcome customers value, how that value varies across segments, and which measurable unit reasonably tracks it. Then test whether buyers understand and accept the package and price. A convenient billing metric can create friction when it rises faster than perceived value or makes costs difficult to predict.

[Inference] It can fit when consumption is measurable, varies meaningfully across customers, and relates to value or delivery cost. Evaluate predictability, budget approval, measurement quality, and margin at different usage levels. Usage pricing is not inherently more customer-friendly or more profitable than a subscription.

[Inference] Unlimited offers can simplify a buyer's decision while shifting heavy-use risk to the provider. The relevant questions are what is actually unlimited, which restrictions remain, and how variable costs behave. Evaluate the distribution of customer usage and the cost to serve before treating a flat price as economically simple.

ARR concerns a defined recurring revenue base expressed annually. A revenue run rate extrapolates a shorter observation period and can include activity that is not contractually recurring. Neither should be substituted automatically for recognized revenue. Comparisons need the metric definition, period, source, and assumptions.

No. Gross merchandise value describes a defined transaction value flowing through a marketplace or commerce system. The platform may earn only a fee or another portion of that value. To understand the business, distinguish GMV, take rate, recognized revenue, incentives, and the costs associated with serving transactions.

A basic calculation divides defined acquisition spending by the corresponding number of acquired customers. The useful work is in defining the numerator, denominator, and period: which salaries, programs, tools, and overhead are included, and how long spending takes to produce customers. State whether the result is blended or channel-specific.

Costs may be omitted, acquisition spending may precede customer conversion, or brand and founder relationships may contribute demand that is difficult to repeat. A blended average can also hide an expensive marginal channel. Read the calculation and cohort definitions before comparing CAC across companies or periods.

CAC payback estimates how long it takes to recover acquisition cost from a defined stream of customer contribution, often using gross-profit-adjusted recurring revenue. It is not simply the time until invoices equal acquisition spending. The calculation should identify margin, timing, churn, and which costs are included.

There is no universal answer independent of the calculation and business. [Inference] Examine retention evidence, margin, timing, concentration, capital needs, and the reliability of the lifetime estimate. A favorable ratio built on untested retention assumptions can be less informative than a modest estimate grounded in mature cohorts.

Growth can bring higher acquisition spending, more expensive customers, larger discounts, heavier usage costs, or greater service requirements. Revenue measures the commercial top line, while unit economics examine the revenue and costs of a defined unit. The two can move in different directions.

[Inference] Separate the price customers pay from the costs of model inference, infrastructure, human review, implementation, and support relevant to the product. Look at different usage and customer cohorts. A compelling demonstration or rapid revenue run rate does not establish healthy contribution margin or repeatable customer retention.

Implementation, training, migration, and customization can be essential to customer value while requiring substantial human effort. [Inference] Measure whether those services are repeatable, paid appropriately, and reducing future support needs. Treating all contract value as equivalent high-margin software revenue can obscure the work needed to deliver it.

[Inference] Onboarding should help the customer reach a meaningful outcome, understand ongoing responsibilities, and establish a credible path to continued value. Completing setup is only one milestone. Define what success means for the user, administrator, and buyer, and make ownership of unresolved implementation work explicit.

Activation is a defined early event or outcome indicating that a user has begun experiencing value. Retention concerns whether the user or customer continues over time. A product can create a dramatic first experience yet fail to support recurring work, so activation needs validation against later behavior.

[Inference] Assign ownership according to the customer journey and the work involved, then make responsibilities explicit across sales, account management, and customer success. Expansion should follow a credible additional customer need. Ambiguous ownership can create duplicated contact or leave commercial opportunities and adoption risks unattended.

Support commonly helps resolve questions and incidents. Customer success commonly coordinates adoption, outcomes, and relationship health proactively. Companies allocate these responsibilities differently, sometimes combining them. The title is less informative than the actual work, account coverage, handoffs, and ownership of renewal or expansion.

No. A role or headcount establishes an organizational choice, not its effectiveness. Retention depends on customer outcomes and commercial behavior that require separate evidence. Look for cohort results, adoption progress, renewal ownership, and delivery capacity rather than treating a CSM title as a verified growth mechanism.

The company begins with a limited customer commitment and seeks broader use or spending after demonstrating value. Expansion may involve more users, locations, workloads, or products. [Inference] The critical questions are what triggers that second purchase, who approves it, and whether the account can support the additional service effort.

Gross revenue retention excludes expansion and shows how much starting recurring revenue remains after churn and contraction. Net revenue retention includes expansion within that same starting customer cohort. New-customer revenue is excluded from both; consistent cohort and period definitions are essential.

Yes. Expansion from a subset of accounts can offset losses elsewhere, leaving the aggregate attractive while customer-count retention or particular segments deteriorate. [Inference] Examine gross retention, account concentration, cohort maturity, and the distribution of expansion before concluding that the entire customer base is healthy.

[Inference] Combine relevant product behavior, customer goals, relationship changes, support history, and commercial context, then investigate the account directly. A health score is a prompt for action, not a customer decision. Low activity may signal risk for one use case and be normal for another.

A useful account identifies the initial purchase, the value achieved, the stakeholders involved, the additional scope, and the commercial or operational evidence of expansion. A vendor-published example should remain attributed. It does not establish that the same expansion pattern applies to all customers.

No. Appropriate usage frequency depends on the job the product performs. A monthly financial task, occasional security event, or periodic purchase can create substantial value without daily activity. [Inference] Define healthy usage around the intended outcome and evaluate retention at a cadence suited to the customer workflow.

[Inference] Consider services when customers need help reaching value and the work has a clear scope, owner, and economic model. Learn whether repeated engagements can become documentation, tooling, or a standardized service. Bespoke work that wins a contract can also create obligations the core product cannot yet support.

[Inference] Hire for the most consequential demonstrated bottleneck, not a fashionable title. The need might be qualification, closing, technical evaluation, onboarding, or customer education. Define the work, expected learning, and support available before deciding whether a generalist, specialist, or experienced manager is appropriate.

[Inference] A sales engineer becomes useful when technical evaluation repeatedly constrains commercial progress and the work requires focused expertise. Clarify the boundary with product engineering and implementation. The role should help validate fit and reduce uncertainty, rather than compensate indefinitely for unclear requirements or unsupported capabilities.

[Inference] Document actual customer situations, buying roles, objections, qualification criteria, evidence, and next steps. Use observed calls and deal reviews to test whether the process is understandable to someone else. Preserve the reasoning behind choices so the team can adapt when the next customer differs.

[Inference] Begin with a consequential operational ambiguity, such as conflicting stage definitions, lost handoffs, unreliable account ownership, or inconsistent revenue reporting. Establish a workable process and source of truth before adding tools. A larger stack can make inconsistent definitions more expensive without resolving them.

[Inference] Identify the job, data ownership, workflow, and integration requirements first. Choose tools that support those requirements at the team's actual operating scale. A competitor's stack is not a complete buying rationale, and current website technology does not establish which tools powered a company's early growth.

[Inference] A useful partnership creates a specific benefit for the customer and a viable incentive for each participant. Define whether the partner supplies access, selling, implementation, technology, or service. Track the resulting behavior and economics rather than treating signed agreements or marketplace listings as revenue outcomes.

Partner-sourced revenue is attributed to opportunities originated by a partner under defined rules. Partner-influenced revenue includes opportunities the partner helped advance, which may have originated elsewhere. The categories can overlap, so attribution rules should make double counting and the actual partner role visible.

Not by itself. A listing can enable discovery, procurement, or transactions without creating active demand or resolving evaluation complexity. [Inference] Examine what customers actually do through the marketplace, what the platform contributes, and which selling, onboarding, and support responsibilities remain with the vendor.

[Inference] Define milestones appropriate to the role and ramp stage, such as qualified discovery, technical validation, successful onboarding, or closed business. Separate learning and execution quality from outcomes the hire cannot yet influence. Evaluate whether the territory, product, support, and process make the target credible.

Incentives affect which activities and customers receive attention. A plan that rewards a signature may create different behavior from one that also recognizes durable adoption or account quality. [Inference] Examine handoffs and unintended consequences, particularly when one team creates commitments another team must deliver.

[Inference] Track a small set tied to current decisions: qualified customer conversations, progression to a real evaluation, paid conversion, time to value, retention, and delivery effort. Definitions matter more than dashboard size. Add metrics when they answer an important question about the current bottleneck.

[Inference] Examine where suitable customers stop progressing, how long stages take, and why. Combine quantitative stage data with customer and team evidence. The constraint may be demand, qualification, evaluation, purchase approval, implementation, or retention; optimizing a different stage may create activity without improving the commercial system.

Cohorts can show how customers acquired at different times or through different routes behave as they mature. An aggregate chart mixes new acquisition, retention, expansion, and customer mix. Cohort analysis is useful when the decision concerns whether newer customers retain or produce value like earlier customers.

No. Attribution assigns credit using observed interactions and a model. Causality asks what would have happened without the activity. The two questions can produce different answers, especially when brand demand, untracked conversations, or multiple teams influence a purchase. A credible incrementality test needs an appropriate comparison.

[Inference] State the original hypothesis, audience, execution, result, cost, and decision made afterward. Distinguish evidence against the hypothesis from a poorly executed or inconclusive test. A tactic not mentioned in public materials should not be described as a failed experiment without supporting evidence.

[Inference] An actionable metric connects to a specific decision, has a reliable definition, and can be interpreted in context. Identify who can respond, what alternative explanations exist, and which guardrails matter. A number that rises without changing any decision may describe activity rather than useful progress.

Small samples can be useful for learning but provide limited evidence of repeatability. A few wins may reflect founder relationships, unusual buyers, or chance. [Inference] Record the exact conditions and uncertainty, look for counterexamples, and avoid projecting a precise conversion rate or lifetime value far beyond the observed population.

Two products can look similar while requiring different budgets, approvals, deployment work, or risk tolerance. Those differences affect the GTM motion and economics. [Inference] Compare the customer problem and buying process alongside product capabilities to decide whether a lesson is likely to transfer.

Each post uses a clear company-specific search question about growth, customers, or go-to-market strategy, followed where useful by its central angle. The page has a distinct title, description, canonical URL, and visible headings. The wording is intended to describe the content rather than promise a ranking outcome.

Every post includes a descriptive link to its company case, and every company case links back to its companion post. Related-reading links connect posts with shared research themes. Glossary anchors explain terminology, while sources and methodology links let readers inspect the basis and limits of the narrative.

Editorial links between the site's posts, cases, glossary, and FAQ use ordinary crawlable HTML anchors with descriptive text. They are not marked nofollow. External source links opened in a new tab use noopener. Link attributes describe their actual purpose rather than being added indiscriminately for SEO.

No. Structured data describes content in a machine-readable form; it does not establish rankings, indexing, accuracy, or eligibility for every search feature. The blog uses article markup, the glossary uses defined terms, and the FAQ describes its visible questions and answers. The content remains readable without relying on that markup.

Google's June 15, 2026 documentation update states that its FAQ rich result feature is no longer shown in Search. The FAQ is built for useful reading and semantic description, not a promise of that retired display. Its questions, answers, and internal links remain part of the site's visible content.

No. These files provide a readable directory for systems that choose to use them. Google's June 15, 2026 update says llms.txt is not needed for Google Search and does not positively or negatively affect its visibility or rankings. The site keeps the files as discovery aids, separate from any claimed search outcome.

A private access gate restricts unauthenticated access to the site's content. Titles, sitemaps, and structured data do not remove that restriction. Public discoverability requires an appropriate public deployment and remains subject to search-engine behavior. This content expansion preserves the site's existing audience rather than changing its access settings.

No. The glossary contains editorial explanations of GTM, product, sales, customer-success, pricing, economics, research, and publishing terms. The book's framework attribution is handled separately on the methodology and framework pages. Links to company material are related reading, not claims that the companies used the same terminology.

No. An approximately 600-word explanation can add context and careful analysis without establishing new historical facts. The evidence status belongs to the underlying sources, not the word count. Posts based on partial records preserve the missing details and label interpretations rather than treating length as verification.

Follow the nearby source number, read the associated company case, and check the source's date and attribution. Look for whether the statement is a direct observation, a company-reported result, or an inference. If the source does not resolve the question, the appropriate conclusion remains uncertain rather than implied by the surrounding narrative.