{
  "version": "https://jsonfeed.org/version/1.1",
  "title": "The Ownership Brief",
  "home_page_url": "https://buildownsell.vercel.app",
  "feed_url": "https://buildownsell.vercel.app/feed.json",
  "description": "Published editorial snapshot; no automatic refresh is claimed.",
  "items": [
    {
      "id": "urn:owner-thesis:article:revenue-verified-is-not-buyer-ready",
      "url": "https://buildownsell.vercel.app/research/revenue-verified-is-not-buyer-ready/",
      "title": "Revenue verified. Business transferable?",
      "summary": "A payment connection answers one question. A future owner has at least six more.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "The missing bridge\nTrustMRR describes verification of aggregate payment metrics through connected providers. That is useful evidence of recorded money movement. It does not by itself establish the customer outcome, transfer rights, cost of replacing the founder or durability of demand.\n\nOur thesis: the next useful layer is an evidence passport that connects financial claims to the operational facts a buyer would inherit. The passport should be a collection of dated, scoped receipts—not a single reassuring badge.\n\nBuild six folders before building a data room\nStart with financial periods and revenue composition. Then record customer retention and concentration; code and data rights; delivery economics; channel dependence; and an independent operating test. Each claim should name a source, an owner, a time window and what is still unknown.\n\nFor example, “80% gross margin” is incomplete until a buyer can identify whether inference, retries, human review, support, refunds and the owner’s replacement labor are included. A Stripe screenshot can coexist with weak unit economics.\n\nA useful first action\nAsk another person to run one customer job without your credentials, memory or undocumented judgement. Record every interruption. The resulting exception list is an actionable product backlog and a better starting point for diligence than a polished deck.\n\nDo not upload client contracts, customer lists or bank information to a public assessment form. This launch’s calculation and CIM tools run locally. Private deal-room sharing is a separately permissioned product, not an implied use of your free-tool inputs."
    },
    {
      "id": "urn:owner-thesis:article:ai-moat-test",
      "url": "https://buildownsell.vercel.app/research/ai-moat-test/",
      "title": "What remains when the model gets better?",
      "summary": "Stress-test distribution, workflow, data rights and operating economics before calling something a moat.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Cost is not the same as defensibility\nStanford’s 2025 AI Index documents a dramatic historical fall in the inference price for GPT-3.5-level benchmark performance. That supports testing cost and capability assumptions. It does not establish that all software, distribution or customer service is free.\n\nOur analytical view: improvements in models can help a business and weaken its differentiation at the same time. Lower inference costs may improve margins while making an answer-generation feature easier for competitors to reproduce.\n\nRun the substitution test\nImagine a capable customer rebuilding your headline feature over a weekend. Which parts of the customer outcome would remain difficult? Consider access to customers, integration with their operations, permissioned feedback, reliable delivery and the cost of switching.\n\n“We have proprietary data” is not a complete answer. Document the right to use it, whether those rights transfer, whether the data improves outcomes and whether the same useful information is readily available elsewhere.\n\nRun the survival test\nModel three shocks: inference costs triple; the dominant acquisition channel disappears; the founder becomes unavailable for two weeks. Do not assign invented probabilities. Write down the operational consequence and the cheapest experiment that could reduce uncertainty.\n\nThe strongest result may be an unglamorous improvement: diversified lead sources, an export path, a tested fallback model or an exception-handling procedure. Buyers inherit those capabilities, not your enthusiasm about the technology.\n\nTurn assessment into work\nThe free AI Moat Test separates an opinion from documented evidence. Its score is a prioritization aid, not a valuation or a prediction that someone will acquire the company. Complete one missing evidence task before adding more features."
    },
    {
      "id": "urn:owner-thesis:article:q2-2026-buyers-selective",
      "url": "https://buildownsell.vercel.app/research/q2-2026-buyers-selective/",
      "title": "A slower market is not an empty market",
      "summary": "The Q2 business-for-sale report makes a case for preparation—not for applying a universal multiple to AI companies.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "What changed\nBizBuySell reports 2,117 closed transactions in Q2 2026, down 10% both sequentially and year over year. Its reported average cash-flow multiple was 2.7 and median price was $349,250. These are broker-reported Main Street observations, not a benchmark for a new AI product.\n\nIn its owner survey, 14% reported a professional valuation, while many others had only a rough estimate or no estimate. The report stresses earnings durability and preparation. It does not prove that every prepared company sells quickly.\n\nWhat an owner should do with this\nUse the report as a reason to improve the evidence behind your business, not as a calculator input without comparable context. A local service business, a subscription application and a pre-revenue software asset have different economics and buyer populations.\n\nOur recommended operating move is a monthly owner-readiness review: reconcile the current financial period, record concentration, check assignment clauses and identify the one activity only the founder can perform.\n\nWhat deserves watching\nTrack completed transactions separately from inventory, and distinguish time on market from time since business creation. Changes in transaction count can reflect financing, supply quality or reporting, not simply demand.\n\nFor an AI-enabled business, add inference costs and human-review minutes to this monthly review. An apparent margin improvement caused by unrecorded owner work is not a durable operational gain."
    },
    {
      "id": "urn:owner-thesis:article:price-is-not-proceeds",
      "url": "https://buildownsell.vercel.app/research/price-is-not-proceeds/",
      "title": "Your exit price is not your take-home",
      "summary": "A simple bridge from enterprise value to the cash you can actually use.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Start with the basis\nA valuation scenario is a set of assumptions. State whether the multiple applies to normalized profit, seller discretionary earnings, EBITDA or annual recurring revenue. Do not substitute one measure for another because its resulting number looks better.\n\nNext distinguish enterprise value from equity value. In the simplified cash-free/debt-free scenario used by our tool, equity value equals enterprise value plus excess cash less debt. A real agreement can include working-capital targets, preferences and other adjustments.\n\nFollow the closing waterfall\nAn illustrative company valued at $200,000, with $20,000 of debt and $5,000 excess cash, has $185,000 of modeled equity value. An owner of 80% has a $148,000 gross share before fees, tax and other terms. If 70% of that share is paid at closing, the modeled closing cash is $103,600—not $200,000.\n\nSeller financing, holdbacks, earnouts and retention compensation are not interchangeable. Some amounts arrive later, some are contingent, and some compensate future work rather than ownership.\n\nKeep assumptions editable\nThe free calculator separates those inputs and exposes the fee model. It does not estimate your tax or certify a market price. Bring its output to the appropriate transaction, legal and tax professionals rather than treating a range as a promise."
    },
    {
      "id": "urn:owner-thesis:article:cim-buyers-can-underwrite",
      "url": "https://buildownsell.vercel.app/research/cim-buyers-can-underwrite/",
      "title": "Write a CIM a buyer can question",
      "summary": "An honest information memorandum makes the unknowns visible and the important claims traceable.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Replace persuasion with inspection\nA confidential information memorandum is a structured description of a business offered for a potential transaction. It should help a buyer decide which questions to pursue. It is not a substitute for source records or a way to make an untested business look mature.\n\nBegin with the customer outcome, the operating model and the transaction scope. Separate what is being sold from what stays with the owner. Identify proposed transition work rather than quietly assuming the seller will remain available indefinitely.\n\nUse a claim–receipt structure\nFor each material metric, include the period, currency, accounting basis, source and important adjustments. Make the distinction between recurring contracts, usage, services and one-time sales explicit. Give retention and customer concentration denominators.\n\nAI businesses need additional specificity: model and platform dependencies, training or inference data rights, review labor, failure handling, provider cost sensitivity and ownership of generated or contractor-created assets.\n\nRoast the document, not the founder\nA sharp review can say “This ARR claim has no period or source,” without insulting the seller or inventing fraud. Our initial local CIM Lens finds topic coverage and risky certainty language. It cannot verify those claims or discover every omission.\n\nThe CIM drafting tool only rearranges the facts you provide and leaves missing sections plainly marked. A future generative reviewer must cite the uploaded evidence, abstain when it is missing and remain private to the authorized owner."
    },
    {
      "id": "urn:owner-thesis:article:quickbooks-valuation-not-one-number",
      "url": "https://buildownsell.vercel.app/research/quickbooks-valuation-not-one-number/",
      "title": "What QuickBooks can—and cannot—tell a buyer",
      "summary": "Importing a profit-and-loss statement is the beginning of a valuation workflow, not its conclusion.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "An accounting report is a scoped observation\nIntuit’s accounting API supports reports, but the standard accounting OAuth scope is broad. A product that promises read-only behavior must enforce that restriction in its own backend; it should not imply that this standard scope is a narrow, read-only token.\n\nThis launch uses local import of a QuickBooks ProfitAndLoss JSON report instead of pretending a live integration is connected. The importer identifies the reported net-income total, dates, basis and currency. It does not produce verified SDE or guess at add-backs.\n\nNormalize with evidence\nCheck whether the period is complete. Review cash versus accrual accounting, related-party expenses, one-time items, deferred obligations and the cost of replacing owner labor. Treat bank reconciliation, customer contracts and revenue quality as separate evidence.\n\nA short report period should not silently become annual profit. The importer displays the period and leaves valuation inputs under your control. Unexpected columns or ambiguous totals should stop the automated path and require review.\n\nConsent is a product feature\nAn owner connecting an accounting system for a readiness report has not agreed to publish the books, sell their data or distribute the business to buyers. Those are separate choices with separate retention and revocation rules.\n\nThe planned live integration should encrypt refresh tokens, allow disconnection, log report access and restrict API methods to approved GET reports. It also needs a tested account-isolation boundary before it is made available."
    },
    {
      "id": "urn:owner-thesis:article:buyer-mandate-before-marketplace",
      "url": "https://buildownsell.vercel.app/research/buyer-mandate-before-marketplace/",
      "title": "Start with a mandate, not a thousand listings",
      "summary": "A useful sourcing product learns what a buyer will actually consider—and what an owner permits you to share.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Turn preferences into a specification\n“Profitable AI businesses” is not an actionable mandate. Capture price range, minimum operating history, profit definition, desired ownership, model dependencies, customer concentration, geography, financing constraints and the work the buyer is willing to perform.\n\nExisting networks such as Axial already connect buyers with transaction opportunities. Our proposed specialization is smaller AI-enabled businesses and the additional evidence their operating model requires. This is a product thesis, not proof of unmatched market demand.\n\nSource with permission\nPublic evidence that someone owns a company does not prove they intend to sell. A sourced company profile is not a listing. An owner-approved anonymous teaser is not permission to send a confidential memorandum to every subscriber.\n\nRequire owner identity review, scope of consent, expiry and an explicit related-party disclosure. For confidential access, record which buyer is permitted to see which version of which document. Revocation must stop future access.\n\nDo not improvise a brokerage model\nBroker registration and exemption questions depend on the actual activities and transaction, not just the label on an invoice. Federal M&A broker provisions and state requirements need qualified legal review. Minority-interest or other securities transactions add questions rather than removing them.\n\nUntil the commercial workflow is reviewed, the launch offer is clearly scoped research and preparation—not negotiation, custody, transaction execution or contingent compensation. A fixed fee is not a blanket legal exemption."
    },
    {
      "id": "urn:owner-thesis:article:build-to-sell-without-building-a-job",
      "url": "https://buildownsell.vercel.app/research/build-to-sell-without-building-a-job/",
      "title": "Build to transfer, not to become indispensable",
      "summary": "A software factory creates an asset only when someone else can reliably own and operate the result.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Transfer the economic mechanism\nA useful acquisition example can reveal a customer problem, distribution channel or operating model worth studying. It does not grant a license to copy the brand, protected design, code or customer data.\n\nChange one variable as a disciplined experiment: the niche, platform, job, business model or channel. Preserve the proposed value mechanism and record what may break. App-store copycat policies and integration licensing can make a superficial clone commercially useless.\n\nTest the handoff before testing the multiple\nGive an operator the credentials and documents they would actually receive at closing. Have them fulfill a job, reconcile a payment, handle a failed model response, respond to a customer and restore a backup. Measure interventions from the founder.\n\nThe initial goal is low routine operating burden with an accountable owner, not the assertion that a legal business can have no responsible human. Refunds, incidents, tax filings, privacy requests and contracts still need ownership.\n\nLet the proof decide the factory backlog\nBefore committing to a build, seek a paying customer and a credible buyer conversation. Use cheap prototypes to find the most expensive assumption. Retire ideas that fail the test and preserve the lesson.\n\nAn exit series becomes more credible when it includes unsold assets, reused work, actual expenses and transition obligations. Publish that full ledger rather than turning a commissioned build or a deferred payout into a misleading acquisition headline."
    },
    {
      "id": "urn:owner-thesis:article:financing-before-the-loi",
      "url": "https://buildownsell.vercel.app/research/financing-before-the-loi/",
      "title": "Finance the deal before the LOI finances your imagination.",
      "summary": "A capital stack is a set of obligations, not a collection of optimistic percentages.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Start with uses, not the headline price\nWrite down purchase consideration, fees, initial working capital and a realistic reserve. Then reconcile those uses with cash equity, senior financing and any seller obligation. A missing reserve is a funding gap, even when the purchase-price math looks neat. A seller who leaves money in the business has not supplied the same instrument as a seller who lends the buyer money.\n\nSpecify the unit and reporting period for every number. A monthly operating profit, a trailing annual adjustment, an estimated property value and a headline asking price cannot share an unlabeled spreadsheet column.\n\nSeparate three different loan statements\nSBA’s official overview lists a $5 million maximum for an individual 7(a) loan and allows eligible ownership changes among other uses. A separate May 18, 2026 announcement describes cumulative 7(a)/504 financing up to $10 million for eligible structures from July 4. The 504 program has fixed-asset purposes and excludes working capital. None of that automatically makes a $10 million goodwill-heavy acquisition financeable.\n\nBring a lender the actual transaction and ask which rules apply. The SOP edition register lists version 8.1 effective October 1, 2026. This tool does not claim to implement its detailed thresholds, and a planned closing date alone is not enough to select the controlling guidance.\n\nModel the operator after the acquisition\nStart with a documented cash-flow measure. Deduct a replacement operator only when that expense has not already been deducted. Treat maintenance capex and ongoing working-capital needs separately from discretionary growth. Then calculate scheduled debt service and a downside case. Explain the differences between your proxy and the lender’s underwriting definition.\n\nOur calculator deliberately shows a standby note’s remaining liability. A zero current payment does not erase principal, accrued interest, a balloon payment or refinancing risk. Interest-only and standby are scenarios until the lender and documents approve the actual arrangement.\n\nThe next useful action\nPrepare a one-page sources-and-uses schedule, historical statements, adjustment evidence, an operating plan, and a clear list of unresolved conditions. Ask a qualified lender to identify the three items most likely to change eligibility or structure. Resolve those before negotiating around a price that depends on fictional financing."
    },
    {
      "id": "urn:owner-thesis:article:thesis-before-target-list",
      "url": "https://buildownsell.vercel.app/research/thesis-before-target-list/",
      "title": "A list is not a thesis.",
      "summary": "The advantage is a decision you can defend, not another spreadsheet of companies.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Name the decision\nA useful search thesis says which business characteristics make ownership attractive for this buyer, which conditions disqualify a target, and what observations would change that position. “Good businesses with growth potential” has no screening power. Neither does an industry label without a customer, geography, operating model and financing context.\n\nBegin with one decision: which two subcategories deserve the next month of work? This forces comparisons and makes a research agenda more valuable than an undifferentiated queue.\n\nGive must-haves a veto\nA mandatory geography, operating capacity, revenue floor or prohibited exposure cannot be averaged away by attractive website copy. Separate hard gates from preferences. Define what evidence can satisfy each gate, how old it may be, and what happens when it is absent.\n\nUnknown should have its own state. A company can look attractive on the two fields we know and still require evidence on the eight fields that determine whether it fits. Keep overall coverage beside the weighted score.\n\nWrite the thesis twice\nFirst build a bottom-up thesis from sourced conversations and observations, without giving the synthesizer the principal’s existing answer. Then compare that output with the working thesis. Preserve disagreement; do not request a harmonized story merely because the original thesis came from a senior person.\n\nFor every observation retain who said it, when, what they were in a position to know, conviction, relevance and whether it supports or challenges the thesis. Repeated hearsay is not independent corroboration.\n\nThe deliverable is a test\nConclude with a shortlist of research priorities, explicit exclusions, and the cheapest falsification test for each. Our recommended rule: every strong recommendation must identify the fact most likely to overturn it. The Search Desk exports the evidence and the working thesis in separate phases so the comparison is possible without quietly leaking the answer into step one."
    },
    {
      "id": "urn:owner-thesis:article:off-market-does-not-mean-for-sale",
      "url": "https://buildownsell.vercel.app/research/off-market-does-not-mean-for-sale/",
      "title": "Off-market does not mean for sale.",
      "summary": "Three states that a trustworthy acquisition product must never merge.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "A research universe is not a marketplace\nA public locations page can establish that a company claims to operate in a region. It cannot establish that the owner wants to sell. A review count cannot establish revenue; a recent website redesign cannot establish expansion funding. Start with modest claims that the source actually supports.\n\nUse three distinct records: a researched company, a reviewed relationship/contact path, and an owner-authorized opportunity. Each promotion between states needs an explicit reason and permission.\n\nMake outreach useful before making it frequent\nExplain who you are, why this particular business is relevant, and what a low-pressure next conversation would be about. Do not say you represent a buyer, have funding or have a mandate unless that is true. Asking to learn how an operator thinks about succession can be more honest than pretending every inquiry is a purchase offer.\n\nA warm introduction can add context and accountability. Ask a banker, fractional CFO or operator for an introduction they are allowed to make—not private borrower information. A lender relationship is not permission to infer or disclose distress.\n\nTreat a decline as data, not friction\nRecord the response, scope and date. Do not turn an owner’s reply into newsletter consent. Do not keep moving an opted-out person into new sequences under different campaign names. The draft tool records a suppression flag and stops the export path for a suppressed recipient. A production sender must verify the actual suppression service rather than trust a local checkbox.\n\nThe FTC’s commercial-email guidance addresses truthful headers, subjects, identification, postal address and opt-out handling. B2B commercial messages have no blanket exemption. Whether a specific acquisition inquiry is covered depends on its primary purpose; obtain appropriate legal review instead of labeling every message compliant by template.\n\nImprove the conversation record\nAfter a permitted conversation, record timing, ownership objectives, the owner’s boundaries and the next agreed action. “Not now” is neither “for sale” nor permission to contact monthly. The system should make a respectful follow-up easier, not manufacture urgency."
    },
    {
      "id": "urn:owner-thesis:article:market-map-without-invented-revenue",
      "url": "https://buildownsell.vercel.app/research/market-map-without-invented-revenue/",
      "title": "Build a market map without inventing private revenue.",
      "summary": "Use geography and operating structure to ask better questions—not manufacture precise targets.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Define the denominator\nStart with a geography, business model and industry definition. Explain what is included and excluded, and whether the count represents businesses, establishments, franchise brands or physical sites. A five-location brand is not automatically five independent acquisition opportunities.\n\nKeep local customer demand, supply density and the buyer’s ability to operate in the region as different questions. Combining them too early produces a single score with no clear interpretation.\n\nGive each source a specific job\nCensus County Business Patterns supplies aggregate establishment, employment and payroll context. BLS QCEW supplies industry and geographic labor information. Neither is an individual company’s P&L. Directory metadata can supplement a company’s published footprint, but it may be incomplete or stale.\n\nThe CBP API documentation now requires a key for requests. Store the dataset vintage, release date and geography alongside results. A page checked in 2026 may describe an older observation year. Suppression and missing categories remain missing—not zero.\n\nMake the map show uncertainty\nFor each sector-geography cell, record observed supply, source coverage, operating questions, direct competitors, evidence of customer demand and available relationship paths. Leave profitability, ownership structure and seller intent blank unless actual sources establish them.\n\nAn apparently sparse market may represent unmet demand, poor classification, unavailable data or economics that discourage entry. Those explanations imply different tests. Do not label a data gap an opportunity before distinguishing them.\n\nChoose the next field test\nOur recommended output is a ranked research agenda: which subcategory, why it fits the buyer, what contradicts the thesis, which operator or intermediary could clarify it, and which evidence would stop the search. Save a short decision memo with the source snapshot so new data can change the recommendation rather than overwrite history."
    },
    {
      "id": "urn:owner-thesis:article:private-market-data-is-not-the-moat",
      "url": "https://buildownsell.vercel.app/research/private-market-data-is-not-the-moat/",
      "title": "Data is consolidating. Your decision record should get better.",
      "summary": "What Datasite’s Valu8 acquisition suggests—and what it does not prove.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "What is reported\nDatasite announced its acquisition of Valu8 on May 8, 2026, describing stronger private-market financial data and ownership visibility within its sourcing and execution ecosystem. Terms were not disclosed in the announcement. That is a strategic capability story, not a usable numerical valuation multiple for a new app.\n\nOur interpretation\nThe implication for a new entrant is not “build a bigger company database.” A stronger starting point is to help a buyer turn the sources they can access into a defensible decision: which target fits, which evidence is missing, what changed and why the next action is justified.\n\nA durable decision record combines declared criteria, source observations, rejected alternatives, conversations, changes of mind and actual outcomes. This is our product thesis, not a claim that the acquisition proves it.\n\nWhat an owner can learn\nAn asset may be valuable because it improves an existing workflow, adds difficult-to-recreate data rights, or fits a buyer’s distribution. Reverse engineering that mechanism means studying rights, usage and integration—not reproducing the acquired product’s interface.\n\nDo not infer the acquirer will buy your company. A historical transaction supplies a research lead. Current strategy, budget, authority, integration constraints and evidence requirements need their own conversations.\n\nA useful experiment\nTake one actual acquisition announcement and annotate three columns: what the buyer explicitly said, what we infer, and what remains unknown. Then generate one adjacent hypothesis with one changed variable. Ask a relevant potential buyer what they would have to see to consider owning it. Until they answer, the acquisition hypothesis is unvalidated."
    },
    {
      "id": "urn:owner-thesis:article:om-triage-before-diligence",
      "url": "https://buildownsell.vercel.app/research/om-triage-before-diligence/",
      "title": "Read an OM without inheriting its conclusion.",
      "summary": "A first-pass screen should preserve the claims it cannot yet substantiate.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Treat the memorandum as an input\nAn offering memorandum can organize a seller’s story. It does not replace the underlying evidence. Read it against your own criteria before its suggested valuation or narrative becomes the anchor for the process.\n\nCapture every important number with its period, definition and source. Revenue, adjusted EBITDA, owner cash flow, bookings and run-rate are different measures. Flag unaligned periods rather than silently reconciling them into a precise-looking multiple.\n\nAsk what changes after the founder leaves\nIdentify the tasks the owner performs, who owns customer relationships, the availability and cost of replacement labor, licensing, key-person risks and transition obligations. A claim of low operating effort needs logs and an independent handoff test, not a paragraph in a deck.\n\nFor physical locations, surface lease term, renewal options, landlord consent, site condition and deferred capital needs. These are diligence questions, not assumptions to fill from a model’s general knowledge.\n\nReturn a decision packet\nThe first pass should say proceed, stop, or evidence needed; show the mandatory gates; and identify the few documents most likely to change the decision. A list of fifty generic questions is less useful than five questions tied to specific claims and constraints.\n\nFor a financed acquisition, separate business attractiveness from lender eligibility and affordability. SBA’s overview itself places repayment ability and lender review at the center of its program. A promising OM is not a loan approval.\n\nKeep the claims separate from the record\nMaintain a source snapshot and an issue log. Request missing material through the authorized channel and respect confidentiality. Our local CIM Lens only flags topic mentions and missing categories; it does not perform a quality-of-earnings engagement or independently verify a document."
    },
    {
      "id": "urn:owner-thesis:article:owner-to-buyer-evidence-path",
      "url": "https://buildownsell.vercel.app/research/owner-to-buyer-evidence-path/",
      "title": "The owner’s evidence becomes the buyer’s confidence.",
      "summary": "Build the business for someone else to understand without building your whole life around an exit.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Choose an outcome before a multiple\nAn owner might want liquidity, a partner, a transition, growth capital or simply a more independent business. Those are different objectives. A minority investment, asset sale, full acquisition and founder rollover can produce very different control and cash outcomes even when the headline company value is similar.\n\nMake one operating claim falsifiable\nPick the claim most likely to matter to a buyer: customers renew, the delivery works without the founder, the data can legally transfer, or margins survive current model and support costs. Define the evidence, owner, observation window and disconfirming result.\n\nDo not wait until a buyer arrives to discover that the contractor agreement lacks a usable assignment or that a key integration cannot transfer. Record those uncertainties now and obtain professional review where needed.\n\nUse AI to make the evidence cheaper, not to fake it\nAutomation can help classify records, draft documents, monitor exceptions and prepare comparisons. Its outputs remain proposals until checked against the relevant source. A model-written customer story is not a reference, and a synthetic buyer interview is not demand.\n\nGive the next operator a working playbook, explicit failure modes and ownership of incidents. An automated routine still needs someone accountable for refunds, privacy requests, outages and other obligations.\n\nA thirty-day owner move\nChoose one workflow. Record its current cost and failure rate, document the inputs and decisions, and ask someone else to run it with appropriate supervision. Save what broke and what changed. Whether the eventual decision is to sell or keep the business, reduced dependency and clearer evidence can make the ownership decision better."
    },
    {
      "id": "urn:owner-thesis:article:relationship-led-sourcing",
      "url": "https://buildownsell.vercel.app/research/relationship-led-sourcing/",
      "title": "The best sourcing workflow earns introductions.",
      "summary": "Turn a specific thesis into a useful request for the people who already know the market.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Trade specificity for trust\nAn intermediary cannot responsibly help with “send me good businesses.” Give them the category, geography, business model, scale and exclusions, plus why the buyer could be a credible next owner. Explain what you do not need and what information should remain confidential.\n\nOffer a small useful artifact: a sourced market brief, owner-readiness checklist or clean one-page criteria summary. The aim is a relevant conversation, not disguising an advertisement as an independent recommendation.\n\nBuild a permissioned referral loop\nStart with operators, bankers, fractional CFOs, accountants, transaction counsel and industry specialists. Ask what they are seeing in aggregate and whether they can introduce someone who would welcome the discussion. Do not ask them to expose client financials or confidential problems.\n\nRecord introducer, permission scope, next agreed action and any conflict. Paid referral arrangements need legal and disclosure review; a revenue target is not a reason to improvise transaction-compensation rules.\n\nUse direct channels with restraint\nResearched email, calls and direct mail are hypotheses to test against response quality and permission—not simply volume. For commercial email, use the FTC guidance as a baseline and obtain jurisdiction-specific advice where appropriate. Keep unsubscribes and owner declines actionable across every campaign.\n\nThe first commercial test\nOur suggested experiment is one narrow paid research pilot with a real decision maker: one mandate, two short decision briefs and a criteria-change log. Success is a better decision and a willingness to renew. It is not a guaranteed deal count, a scrape volume or a claim that every identified company is for sale."
    },
    {
      "id": "urn:owner-thesis:article:contribution-not-deal-size",
      "url": "https://buildownsell.vercel.app/research/contribution-not-deal-size/",
      "title": "Your contribution is not the deal size",
      "summary": "A defensible track record starts with what you actually owned, not the employer’s headline transaction value.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "The size belongs to the transaction\nThe same deal can involve a founder who created the product, an operator who rebuilt delivery, an advisor who led negotiations and a specialist who handled one workstream. Those are different achievements. Record them separately.\n\nA smaller outcome can reflect deeper responsibility\nOur proposed Contribution Index weights decision ownership, execution, stewardship, durable outcomes and handoff. Deal value contributes no points. A well-run small company can demonstrate all five. This is an explicit product rubric, not a scientific score of a person.\n\nA better question for your next milestone\nWhat could a customer, buyer or operating colleague substantiate about the part you personally delivered? Start with one clear statement and its evidence."
    },
    {
      "id": "urn:owner-thesis:article:badge-is-a-claim",
      "url": "https://buildownsell.vercel.app/research/badge-is-a-claim/",
      "title": "A badge is a claim. Make it inspectable.",
      "summary": "Identity, role, outcome and amount verification are different promises.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "More than an image\nOpen Badges describes issuer, recipient, achievement criteria and evidence. The W3C credential model distinguishes claims, holders and verifiers. A good-looking PNG alone does not supply those properties.\n\nSay what was reviewed\nDo not replace several distinct checks with a universal verified label. A confirmed identity does not establish an exit amount. A signed statement protects its integrity; it does not make its underlying assertion true.\n\nThe initial release\nBuild Own Sell’s current card studio produces unsigned, self-reported visual summaries. Future reviewed milestones need scoped reviewer receipts and a correction or revocation path. Free and paid users must meet the same criteria."
    },
    {
      "id": "urn:owner-thesis:article:profiles-without-surveillance",
      "url": "https://buildownsell.vercel.app/research/profiles-without-surveillance/",
      "title": "Professional profiles without a surveillance business",
      "summary": "Research can be useful without inventing a person’s private story.",
      "date_published": "2026-09-15T12:00:00Z",
      "content_text": "Resolve identity before attribution\nA name match alone is unreliable. Use at least two professional anchors, such as company and role or company and a primary profile reference. Leave ambiguous candidates unmerged.\n\nA social URL is not a scraping license\nLinkedIn explicitly prohibits unauthorized scraping and automation. Start with permitted company announcements, founder pages, filings, authorized uploads and licensed interfaces. Preserve claim-level provenance instead of copying unnecessary personal information.\n\nLet the subject correct the record\nKeep unclaimed research private and unscored in this product. Never infer family wealth, net worth, hardship, personal proceeds or willingness to sell from an acquisition article. Give people a clear route to correct, dispute or withdraw a published claim."
    }
  ]
}