Research notes (capture-layer): the affirmative, inward decision-edge case for data intelligence — information asymmetry applied to pricing, demand, risk, retention, targeting (June 2026)

Status: Capture-layer (Deliverable 1 only) per author framing — not a KB entry or article draft. Affirmative, inward decision-edge counterpart to the cautionary trio Research brief: Public data as a private moat — building proprietary intelligence from government open data (piece 11 of 15) (public-data moat), Research brief: The Dataset is the Product — when a service business should own its data (piece 12 of 15) (own vs rent), and Research brief: live data and data-driven tools for SMBs — when it's an edge, when it's overkill (June 2026) (edge or overkill).

The seam in one line

Prior briefs answer "should you build/own it?" (mostly: be careful, often no). This piece answers "given you can see something the market cannot, which inward decisions does that change, and why is an uninformed competitor disadvantaged?" The unifying lens is information asymmetry applied to the firm's own decisions (Akerlof 1970 — "The Market for Lemons"; asymmetric information can collapse markets (Nobel 2001), Stigler 1961 — "The Economics of Information"; costly search produces price dispersion (JPE 69:213-225)) — not generic analytics or dashboards.

Five validated decision domains

Candidate set validated against independent evidence: (Synthesis: the five inward decisions a proprietary information advantage actually changes)

1. Pricing — SURVIVED (strongest, anchor). Information economics (Stigler 1961 — "The Economics of Information"; costly search produces price dispersion (JPE 69:213-225)) + airline yield management (American Airlines yield management — $1.4B over 3 years, ~5-7pct incremental revenue, Franz Edelman Award 1991 (Smith Leimkuhler Darrow, Interfaces 1992)) + Progressive telematics (Progressive Snapshot — first UBI ("Autograph") 1996; modern Snapshot 2010; surcharge capability 2014).
2. Demand & timing — SURVIVED, MERGE NOTE. Cleanest documented evidence overlaps with pricing (RM is demand-forecasting applied to price) and with risk (leading indicators). Distinct domain but flag the overlap. SMB-relevant leading-indicator evidence (permits, releases, weather) NOT independently sourced this pass.
3. Risk & exposure — SURVIVED. Alt-data credit scoring explicitly framed as information-asymmetry reduction (NBER WP 29840 (Di Maggio, Ratnadiwakara, Carmichael, 2022) — "Invisible Primes: Fintech Lending with Alternative Data", IFC / World Bank — "Cracking the Credit Code: Alternative Data and AI for Financial Inclusion" (2026), J-PAL Agarwal et al. 2019 India — mobile / social-footprint ML predicts loan defaults more effectively than credit-only models) + Progressive's pricing-as-risk-selection (Progressive CIO Ray Voelker — Snapshot "given us access to segments of the auto insurance markets that we normally did not attract").
4. Retention — SURVIVED, evidence thinner / vendor-tainted. Capability is real (INFORMS Analytics Magazine — churn modelled as binary classification on RFM / engagement signals (logistic regression / decision trees / ensembles)). Magnitude claims (5x-25x cheaper to retain, 5pct retention → 25-95pct profit) are heavily vendor-recycled — see Quarantine: "5x-25x cheaper to retain," "5pct retention → 25-95pct profit," "80pct profits from 20pct customers," "AI churn → 20-30pct retention improvement" — vendor-recycled, untraced to primary.
5. Targeting & market selection — PROVISIONALLY SURVIVED, WEAKEST. Tesco "Lifestyles" segmentation (Tesco "Lifestyles" — behavioural segmentation system used to target Clubcard mailings (SS, weakest evidence in the targeting domain)) and Progressive supports targeting within an existing base. Independent primary-source evidence for site selection / territory / lookalike as a distinct SMB-relevant edge was NOT secured this pass. See Gaps: information-asymmetry decision-edge brief (June 2026) — designated next-pass research.

What the prior briefs already own (do NOT re-argue)

• Research brief: Public data as a private moat — building proprietary intelligence from government open data (piece 11 of 15) owns: how a normalised corpus built from government/open data becomes defensible; mechanics of moat-building from public sources. This piece treats a normalised public-data corpus as one valid form of proprietary information advantage and points to the seam — does not re-derive the moat framework.
• Research brief: The Dataset is the Product — when a service business should own its data (piece 12 of 15) owns: the data-maturity curve, the spreadsheet ceiling, when a service business should own its data, build/buy decision. This piece assumes own-vs-rent is settled and focuses on what the information does to specific decisions.
• Research brief: live data and data-driven tools for SMBs — when it's an edge, when it's overkill (June 2026) owns: the argument that most SMBs should NOT build a data product; the overkill case. This piece is the affirmative counterpart — the global "don't build it" argument stays with the existing brief and is not restated here as the article spine; the counter-case lives inside the threshold-conditions discipline of this brief, not as its thesis.

Anchor claims (V = verified, IC = industry-consensus, SS = single-source, DS = directional-self-report)

Theory (V, non-commercial):

• Stigler 1961 — "The Economics of Information"; costly search produces price dispersion (JPE 69:213-225) — information is a costly economic good; lower search cost → captured surplus.
• Akerlof 1970 — "The Market for Lemons"; asymmetric information can collapse markets (Nobel 2001) — the spine: an information advantage is structurally the counterparty's information disadvantage; Nobel 2001 (with Spence, Stiglitz).
• Levin (Stanford) — making private information public unambiguously improves trade; the edge erodes as data diffuses — important nuance: making private information public unambiguously improves trade — i.e., information edges decay as data diffuses.

Named cases:

• American Airlines yield management — $1.4B over 3 years, ~5-7pct incremental revenue, Franz Edelman Award 1991 (Smith Leimkuhler Darrow, Interfaces 1992) — $1.4B over 3 years, ~5-7pct incremental revenue, Edelman Award, peer-reviewed in Interfaces. The most defensible single magnitude figure in the corpus.
• Revenue management — the four preconditions that generalise yield management beyond airlines — fixed/perishable resource + time-limited sale + heterogeneous WTP + controllable segmentation; generalises beyond airlines.
• Tesco Clubcard 1994 trial — 9-14 stores over 3 months, >50 million transactions; dunnhumby analysed ~10pct sample + Tesco Clubcard — overtook Sainsbury's in 1995 within a year of launch; helped double grocery market share within ~3 years + Lord MacLaurin epigraph — "you know more about my customers after three months than I know after 30 years" (Tesco Clubcard, 1994) — strongest documented loyalty-card case.
• Progressive Snapshot — first UBI ("Autograph") 1996; modern Snapshot 2010; surcharge capability 2014 + Progressive CIO Ray Voelker — Snapshot "given us access to segments of the auto insurance markets that we normally did not attract" + Progressive — pricing accuracy framed as individuals vs actuarial class ("customers want to be treated as individuals — not as members of an actuarial class") — pricing-as-risk-selection through proprietary observation.
• NBER WP 29840 (Di Maggio, Ratnadiwakara, Carmichael, 2022) — "Invisible Primes: Fintech Lending with Alternative Data" + IFC / World Bank — "Cracking the Credit Code: Alternative Data and AI for Financial Inclusion" (2026) + J-PAL Agarwal et al. 2019 India — mobile / social-footprint ML predicts loan defaults more effectively than credit-only models + PMC peer-reviewed — excluding alternative data led to "a significant decline in model performance" (PMC11108212) — alt-data credit as information-asymmetry reduction.

Capability without outcome (Canadian SME context):

• BDC digital-maturity research — "only one in five Canadian businesses has achieved a high level of digital maturity, while more than half show low levels" — under one quarter of Canadian SMEs are high/very-high digital maturity; mature firms grow faster, profit more, export more.
• BDC / MIT framework — "investing in digital technologies drives revenue, but transformation management capabilities drive profits" (capability ≠ outcome) — investing in tech drives revenue, but transformation-management capabilities drive profit. Capability ≠ outcome.

Edge decay:

• Tesco / Computing on dunnhumby sale — "Using CRM data from loyalty cards is now common practice rather than the unique differentiator that it once used to be" — Tesco's own data chief now says loyalty-card CRM is "common practice rather than a unique differentiator." Treat any information edge as temporary rent.

Caveats (quarantined claims and known limitations)

See Caveats: information-asymmetry decision-edge brief (June 2026) — vendor-recycled magnitudes + modeled projections for the full quarantine list. High-priority exclusions:

• Progressive consumer-savings marketing figures ($169/$322/$328 avg, up-to-30pct, $1.2B discounts) describe consumer savings, not the firm's information edge — exclude from any draft as outcome evidence (Progressive consumer-savings marketing ($169/$322/$328 avg, "up to 30pct," $1.2B discounts) — EXCLUDED from any article as edge evidence).
• The recycled "5x-25x cheaper to retain" / "5pct retention → 25-95pct profit" magnitudes circulate vendor-to-vendor without primary source (Quarantine: "5x-25x cheaper to retain," "5pct retention → 25-95pct profit," "80pct profits from 20pct customers," "AI churn → 20-30pct retention improvement" — vendor-recycled, untraced to primary).
• The IDC "2.5 hours/day searching for information" figure is a 2001 intranet-era estimate, widely mythologised (IDC "2.5 hours/day searching for information" — 2001 intranet-era estimate, widely mythologised).
• BDC's 38pct productivity / 14pct GDP / $350B upside figure is a modeled projection, not realised result (BDC modeled projection — "if all Canadian SMEs reached very high digital maturity, productivity could rise ~38pct, GDP ~14pct, ~$350B" (DS, modeled not realised)).
• dunnhumby co-founder claim of "extra £60bn of sales over 10 years" — incentive-flagged, directional only (dunnhumby co-founder Edwina Dunn — "over the next 10 years Clubcard and dunnhumby made an extra £60bn of sales" (DS, quarantine)).

Gaps designated for the next research pass

See Gaps: information-asymmetry decision-edge brief (June 2026) — designated next-pass research:

1. Targeting / site-selection / lookalike independent SMB evidence — the thinnest domain.
2. Leading-indicator demand (permits, economic releases, weather) tied to a firm-level decision.
3. Overfitting / small-n statistical-significance primary methods reference.
4. Retention magnitude primary sources (likely Reichheld/Bain; Gupta-Lehmann CLV) — or drop the numbers.
5. Verbatim confirmation of the three prior briefs to guarantee no seam duplication.
6. Re-run the subagent → enrich_draft → complete sequence the source session's turn-limit pre-empted.

Rules distilled (R1-R8)

• Rule: treat an information advantage as the counterparty's information disadvantage — that is the source of the edge
• Rule: treat any documented information edge as temporary rent, not permanent moat
• Rule: quarantine the recycled retention magnitudes (5x-25x, 25-95pct, 80/20, 20-30pct AI churn) until primary sourced
• Rule: capability does not equal outcome — without transformation management, tech investment drives revenue but not profit
• Rule: three thresholds before claiming an information edge — volume to clear noise, clean data, an actual decision someone will act on (and timeliness before commoditisation)
• Rule: the mechanism generalises, the magnitudes do not — SMBs cannot extract the same uplift Tesco / AA / Progressive did
• Rule: prefer peer-reviewed / award-vetted magnitudes (Edelman, NBER, INFORMS) over vendor-recycled figures
• Rule: the affirmative info-asymmetry article's seam is inward decisions, not build-vs-own — that is the prior briefs' job

Companion briefs in the June 2026 SMB-advice series

Affirmative complement to the cautionary trio. Adjacent sisters: Research brief: customer-facing calculators & tools for SMBs — the honest case (June 2026) (customer-facing calculators), Research brief: client portals for SMBs — the honest case (June 2026) (client portals), Research brief: dashboards for SMBs — what's worth showing, and when an embedded one earns its keep (June 2026) (dashboards), Research brief: why interactive tools deepen a business's relationship with its audience — a mechanism-level research package (June 2026) (interactive-tool mechanisms).