Go-to-market failure.

A few customers can love a product while the company still has no way to reach the next hundred.

Go-to-market failure is not “sales are low.” It is the inability to reproduce the path by which a qualified customer discovers, evaluates, buys and adopts a useful product at economics the company can survive.

The product may work. Retention may be good in a defined group. Founders may close impressive accounts. Press, referrals or a launch event may create a wave. None of those proves a motion.

The sequence is:

real customer value → isolated wins → path remains implicit → scaling expense amplifies variance → acquisition cost and cycle expand → cash expires.

This manual excludes absent demand. If customers do not care enough to use, retain or sacrifice for the outcome, begin with no-market-need or product-market-fit work. It also excludes a principally wrong price architecture. Pricing can damage conversion, but go-to-market diagnosis asks whether the same buyer, trigger, message, channel and process can be repeated.

The causal thesis is:

Go-to-market failure occurs when a useful product generates isolated wins but leadership cannot specify and reproduce the same buyer, trigger, message, channel, sales steps and economics before cash expires.

The word “repeatable” is often abused. A CRM workflow is repeatable administratively. A commercial motion is repeatable causally: another qualified operator can take a comparable customer situation through substantially the same evidence and decisions, inside a bounded time and cost.

Founders resist constraining the motion because every exclusion appears to shrink the market. So the ideal customer profile remains broad, each seller improvises, partners receive a generic pitch and aggregate pipeline becomes proof of progress. The company looks active while learning is diluted across incomparable attempts.

The discipline is to make one commercial unit explicit before multiplying it.

## Founder charisma hides missing structure

Founders carry context, authority and conviction that a new seller does not. They can change roadmap, pricing and contract terms in a single call. They know the product’s history and can translate vague objections into an improvised story.

A founder-led win is evidence that a deal can close. It does not yet reveal why.

Replication requires decomposing the win:

• the customer’s situation before contact;
• the event that made inaction costly now;
• the user, champion, economic buyer and veto holder;
• the message that earned the next step;
• the proof that retired the largest perceived risk;
• the sequence and time between decisions;
• the product or service concessions; and
• the acquisition and delivery cost.

Without this record, the company hires sellers to imitate a personality.

## The customer definition is aspirational

“Mid-market companies,” “developers” and “people with photos” are populations, not operating segments. A usable segment shares a problem context, trigger, buying process and reachable gathering place.

Broad definitions keep reported total addressable market large. They also make every loss explainable. When one cohort stalls, leadership points to another. The team never sees enough comparable attempts to learn.

## Interest is mistaken for urgency

A buyer can agree that the product is useful and still not act. Go-to-market needs a trigger: a deadline, failure, cost increase, policy, hiring constraint, migration, renewal, new executive or visible risk.

Absent a trigger, sellers create education and follow-up work. Pipeline grows while decision dates slide. “No decision” is not neutral. It is evidence that the motion has not attached value to a current cost.

## Pipeline stages describe the seller

Weak CRMs record meetings, demos and proposals. Strong stages record buyer evidence:

• acknowledged problem and current cost;
• named decision process;
• access to payer and veto;
• agreed proof plan;
• completed proof with success criteria;
• commercial approval; and
• live use.

Seller activity can be manufactured. Buyer progression cannot.

## Channels arrive before direct knowledge

Partnerships, agencies, app stores and reseller networks promise reach. A channel can multiply a working sale. It rarely invents one.

The partner needs a sale-in-a-box: exact target, recognizable trigger, simple qualification, proof, price, objection handling, delivery responsibility and attractive economics. If founders cannot state these, a partner has less information and weaker incentives. The channel produces logos and introductions rather than revenue.

## Scale hiring freezes ambiguity

A senior sales leader is often hired to “build the motion.” If the company has not found a repeatable unit, that person must run discovery, design process, recruit and hit a forecast simultaneously. Compensation and board expectations push toward apparent pipeline before causal learning.

More sellers increase the number of incomparable conversations. They also make changing segment or message politically expensive. Scale spend is therefore a commitment, not just a cost.

Create a GTM replication sheet for each candidate motion. Do not merge enterprise direct sales, self-serve, founder networks and partnerships into one funnel.

## Define the situation

State:

• customer segment and operating context;
• costly before-state;
• trigger that makes action timely;
• current alternative, including doing nothing;
• user, champion, payer and veto;
• reachable source of qualified opportunities; and
• minimum product conditions.

The profile should let an operator reject a tempting but incomparable deal.

## Reconstruct wins and losses

For every material opportunity, record the first source, initial trigger, messages, evidence exchanged, participants, elapsed time, concessions, implementation work and outcome.

Interview buyers after the decision. Ask what changed internally, which alternative was considered, what nearly stopped the deal and which evidence mattered. Do not ask only why they liked the product.

Losses need a controlled vocabulary: no current problem, wrong actor, no trigger, missing capability, risk not retired, price/economics, implementation cost, internal priority, competitor or no decision. “Timing” is too vague unless the next trigger is named.

## Test replication

Choose one narrow motion and run comparable cohorts. A unit is approaching repeatability when:

• qualified opportunity rate is bounded by source;
• conversion by buyer-evidence stage stabilizes;
• cycle time has a useful distribution;
• the same proof closes the same risk;
• concessions and implementation remain bounded;
• a non-founder can execute the process; and
• contribution margin can repay acquisition inside the company’s cash horizon.

The standard is not identical deals. It is shared causality.

## Separate discovery from scale

Discovery metrics:

• comparable conversations;
• trigger frequency;
• stage-specific losses;
• proof completion;
• time to the next buyer decision; and
• changes to the motion thesis.

Scale metrics:

• qualified pipeline per operator;
• cohort conversion;
• cycle distribution;
• acquisition cost and payback;
• implementation burden; and
• expansion and retention.

Do not pay scale incentives during discovery and then wonder why the CRM is full of weak opportunities.

## Use motion states

• Green: multiple comparable cohorts and a non-founder reproduce the route within bounded economics.
• Amber: value and buyer are clear, but one stage, source or proof remains unstable; hiring is capped.
• Red: wins depend on founder relationships, customization or incompatible segments.
• Black: no motion can reach payback before runway expires.

Everpix built an elegant service for organizing and viewing a person’s photos across devices and services. The Verge’s detailed account described strong product ratings, engaged users and a meaningful paid conversion rate, but also very limited marketing until late in the company’s life ^[1].

This matters because Everpix is not a clean no-market-need story. A real group valued and paid for the product.

The commercial system was weaker. Product development consumed the founders’ attention. Growth relied heavily on organic discovery, coverage and word of mouth. The company explored partnerships and financing, but it did not establish an acquisition engine that could predictably add enough paying users relative to storage, development and company cost. It closed in 2013 ^{[1] [2]}.

The lesson is not “spend more on advertising.” Paid acquisition can accelerate loss when targeting and payback are unknown. The lesson is that product admiration and retention do not specify a route.

A replication sheet would have forced choices:

• Which photo owner had the sharpest pain—family archivist, serious hobbyist, multi-device professional or another group?
• What event created urgency—new child, device migration, storage failure, duplicate chaos or sharing need?
• Which existing product or behavior did Everpix replace?
• Which channel held that cohort at a reachable cost?
• What proof overcame privacy and longevity concerns?
• What storage and support cost followed each acquired subscriber?
• Did paid retention repay the source before cash expired?

Everpix’s broad consumer appeal made narrowing emotionally difficult. “Everyone has photos” sounds like a huge market and provides little acquisition guidance. The company needed not merely people who liked a beautiful photo product, but a segment with an acute trigger, willingness to pay and a repeatable place to reach it.

The counterfactual is not certain. A narrower professional or family segment might have reduced the product’s elegance or market size. A different partnership might have worked. The diagnostic conclusion is firmer: the company produced value without converting the path to enough customers into an owned, measurable operating system.

## RethinkDB: technical regard without a sustainable company

RethinkDB developed an open-source database and attracted a developer community. In its shutdown announcement, the company said it had built an elegant system but could not build a sustainable business ^[3]. The project later moved to the Linux Foundation ^[4].

Co-founder Slava Akhmechet’s retrospective argued that the company chose a difficult market and optimized for product attributes that did not translate cleanly into commercial adoption ^[5]. That account includes product, market and management choices, not only sales execution.

The go-to-market lens isolates a difficult open-source transition:

• the user who tries software may not control budget;
• developer enthusiasm may not reveal an urgent enterprise trigger;
• adoption can occur without a monetizable event;
• the buyer may require proof around operations, support and migration rather than technical elegance; and
• community scale can flatter the funnel when few users enter a buying process.

The replication artifact would connect usage evidence to a commercial situation: which production workload, at which organizational threshold, created a support or operational need that a named buyer would fund? It would track time from developer adoption to production, organizational actors, competing alternatives and the proof required for paid conversion.

RethinkDB is not evidence that open source cannot support a business. It shows that product use, technical affection and a business motion are distinct systems.

## Standout Jobs: a channel before the sale-in-a-box

Standout Jobs offered recruitment-oriented career sites and employer tools. In his postmortem, founder Ben Yoskovitz described pursuing a channel strategy before the company had mastered the direct sale: who would buy, at what price and through which repeatable process. Partners did not receive a finished sale-in-a-box, and the recession further reduced their willingness to take risk ^[6].

This is the correct use of a channel case. External conditions mattered. Product and market choices mattered. But the founder’s own account identifies a controllable sequence error:

channel recruitment came before commercial unit definition.

A channel partner has its own quota, customer trust and opportunity cost. It will not spend scarce attention translating an evolving startup story unless the economics are unusually attractive. The startup must supply qualification, message, proof, price, delivery boundary and clear revenue.

The better order would have been:

1. founders close a constrained set directly;
2. reconstruct the common buyer and trigger;
3. remove founder-only concessions;
4. have a new seller reproduce the route;
5. package qualification and proof; and
6. test one partner against the direct baseline.

The partner is then a multiplier with measurable lift, not an escape from unfinished discovery.

Maintain one GTM replication sheet per motion.

## Motion header

• segment and excluded segments;
• trigger and costly before-state;
• user, champion, payer and veto;
• product and service boundary;
• opportunity source;
• owner and learning period; and
• current state: green, amber, red or black.

## Stage contract

Each stage needs:

• buyer evidence required to enter;
• seller action;
• proof asset;
• expected elapsed time;
• exit or loss rule; and
• fields that cannot be self-certified by optimism.

For example, “proposal sent” is seller activity. “Economic buyer confirms the decision process, success criteria and commercial review date” is buyer evidence.

## Replication table

List comparable opportunities by cohort. Record source, trigger, actors, stage dates, proof used, objections, concessions, implementation effort, outcome and loss reason. Show medians and distributions only after confirming the rows are comparable.

## Economics

Include fully loaded acquisition and delivery cost:

• marketing and partner fees;
• seller time;
• founder intervention;
• proof or pilot cost;
• onboarding and implementation;
• discount and custom work;
• support burden; and
• time to collected gross margin.

Founder labor is not free. Custom features are acquisition cost when required to close.

## Learning decision

At each review choose one:

• repeat unchanged;
• narrow segment or trigger;
• change one message or proof variable;
• stop a source;
• redesign a broken product or price condition;
• graduate to a bounded scale test; or
• kill the motion.

Multiple simultaneous changes destroy learning. A large quarter of mixed deals is weaker evidence than ten comparable attempts.

## First week: freeze false scale

Pause hiring, broad campaigns and new channel commitments. Do not stop serving customers or qualified opportunities. Stop multiplying an undefined process.

Segment the current pipeline by actual situation, trigger and buyer—not CRM industry labels. Mark founder-sourced, custom and relationship-dependent wins.

## Weeks two to four: reconstruct causality

Interview recent wins, losses and no-decisions. Examine recorded calls, email sequences, proof plans and implementation. Build the first replication sheet from evidence.

Select one motion with:

• demonstrated customer value;
• a recognizable trigger;
• reachable customers;
• tolerable product gaps; and
• plausible payback.

Reject incompatible opportunities during the learning period, even when their logos are attractive.

## Following six weeks: run the constrained replay

Use the same qualification, message, proof and stage definitions across a bounded cohort. Change one major variable at a time. Have a non-founder run part of the route while a founder observes.

Review weekly:

• where buyer evidence stopped;
• whether the same trigger predicts progress;
• which proof changed the decision;
• concessions and implementation work;
• cycle distribution; and
• economics at observed conversion.

## Then choose

Scale only if the unit repeats. Otherwise narrow again, change the offer or acknowledge that the problem is demand, product or price rather than distribution.

Repair is not a new sales deck. It is the ability to explain a win before it happens and then produce another within a bounded range.

1. Which customers demonstrably receive value and continue using the product?
2. What event makes them act now rather than admire the idea?
3. Who uses, champions, pays and can veto?
4. Which wins depended on founder relationships or custom work?
5. Can a non-founder reproduce the same route?
6. Are pipeline stages buyer evidence or seller activity?
7. What is the largest stage-specific loss reason?
8. Which source produces comparable qualified opportunities?
9. What proof retires the buyer’s largest risk?
10. What are acquisition cost and payback after founder time and implementation?
11. Which apparently attractive deals are outside the current motion?
12. What evidence permits the next hiring or channel tranche?

The board should reject “we need more leads” until it sees the replication sheet. It should reject an aggregate pipeline that combines different buyers, triggers and sales systems.

The final test:

If every win must be explained as special, the company has customers but no go-to-market motion.