Behavioural Solutions
Behavioural Solutions — how MBR gets people to act
Section titled “Behavioural Solutions — how MBR gets people to act”This is the SSOT for how MBR helps people act on financial strategies they already understand. It is the strategic expression of the revised USP (Unique Selling Proposition): MBR competes on behavioural barriers, not knowledge barriers.
Living document. Synthesis of Talbot’s thinking, AI research, and field evidence — add to it as new evidence or ideas arrive. Evidence backing: Behavioral-Knowing-Doing-Gap (Round 1) and Behavioral-Barriers-Playbook (Round 2), frozen in
Strategy/Research/. Every effect size below traces to one of them. Prior art: the original Gemini research that seeded this note is preserved verbatim atResearch/archive/2026-08/Gemini-Behavioural-Barriers-2026-08.md. All 19 of its strategies are carried forward below, now graded against measured evidence.
1. The Thesis
Section titled “1. The Thesis”Two foundational barriers stop people benefiting financially: knowledge and behaviour.
Many players attack knowledge — literacy content, comparison tables, calculators, rate alerts. Almost nobody attacks behaviour from the consumer’s side of the table.
MBR’s position: be the best in the world at dissolving the behavioural barrier — fast, simple, easy.
The evidence says this is not merely an underserved niche; it is where the effects actually are. It also says something more uncomfortable and more useful: the behavioural barrier is not what almost everyone assumes it is.
2. The Core Finding — three frictions, wildly unequal
Section titled “2. The Core Finding — three frictions, wildly unequal”Every natural experiment that removed one kind of friction at national scale:
| Friction removed | Experiment | Result |
|---|---|---|
| Execution — the paperwork | UK Current Account Switch Service (CASS), 12 years: guaranteed 7-day switch, automatic redirection of every payment for 36 months, auto-closure, loss guarantee. 99.6% on-time, 90% satisfaction, 11.9M switches | ≈0 net. Switches in the year to June 2025 (996,344) are below year one (~1.2M) |
| Comparison — which option is best for me | Kling et al. RCT: sent people their own current cost vs the cheapest plan, with a deadline. The same information was already free and heavily advertised | 17% → 28% |
| Decision — choosing, from a field, under uncertainty | Ofgem collective switch: personalised £ saving, one named tariff, a deadline, name + postcode only | 2.6% → 26.9%, five replications |
FCA’s own verdict on execution: “simplifying and speeding up switching in isolation can only have a limited impact on switching volumes.” Their consumer research names the real barrier — consumers “do not even consider switching… because of the lack of a trigger.” Process ranks only second.
Handel (AER 2013) closes the door on the knowledge theory: employees whose health plan became strictly dominated — worse in every possible state of the world, requiring no comparison and no judgment — 89% stayed for a year, 75% for two. Inertia cost: $2,032/employee/year.
The reframe the USP is built on: The loyalty penalty is not why people overpay. People overpay because nobody executes the renewal on their behalf — and the loyalty penalty is what the incumbent charges for that vacancy.
The FCA proved this by accident: it banned price-walking outright, nearly halved the home differential (£95.38 → £49.17) and reversed the motor penalty entirely. Switching moved 1pp — and auto-renewal at inception rose 54% → 70%. Remove the economic reason to shop, and more people go on autopilot.
Corollary — the sentence to keep in mind on every design call: removing friction from a journey nobody starts changes nothing. The pre-filled form that moved switching 3%→12% arrived, unprompted, with one recommendation attached. CASS sat waiting to be found. That difference is the entire business.
3. The Empirical Floor
Section titled “3. The Empirical Floor”Willing, eligible people. Free money. No knowledge barrier. One small action step:
- Class actions requiring a claim: median 9% file (weighted mean 4%). Paid automatically: 55% cash — 81 / 83 / 75% in the $50–100 / $100–200 / >$200 bands
- Money at stake barely matters: <$10 vs >$200 differ by 1 percentage point
- Difficulty barely matters: hardest-documentation forms −27.3pp, but only 5% of cases
- The artifact matters: notice packet with form ~10% · postcard ~6% · email ~3% · postcard with a detachable form back to ~10%
- Objections: 0.0003%. 86% of submitted claims approved — not scepticism, not ineligibility, pure inertia
- UK dormant assets replicate it independently: £2.16bn transferred, £183.3m ever reclaimed (~8.5%), despite a permanent, free, publicised right
Plan against 4–10% for anything self-serve. Plan against 55–83% for anything that arrives done.
4. The Strategy Catalog
Section titled “4. The Strategy Catalog”Talbot’s original 19 (Gemini, 2026-08) carried forward and graded against measured field evidence. ✅ supported · ⚠️ conditional · ❌ evidence absent or against · 🚫 rejected on client-first grounds
A. Supply the decision — highest value
Section titled “A. Supply the decision — highest value”| Strategy | Verdict | Evidence |
|---|---|---|
| Curated defaults — cut choice to ONE recommendation | ✅ Strongest | Ofgem’s arm without the single named tariff underperformed by 5–6pp. Original note said “max 3”; the evidence says one |
| Trigger on a real dated event | ✅ | FCA: absence of a trigger is the primary barrier. Canada supplies the mortgage renewal date free; Ofgem had to manufacture one |
| Deadline / expiry | ✅ | A 1-week deadline beat 3-week (p=0.01) and was statistically indistinguishable from paying people |
| Collapse the plan date to “today” | ✅ | “Switch today” intenders followed through +22.5pp vs those planning in 3–6 months; this acceleration explained most of the treatment effect |
| Second ask to non-responders | ✅ | 14% vs 2%. Few recalled the first contact; none objected |
| Active choice / forced decision | ⚠️ | Directionally sound, but weaker than a pre-started default (below) |
| Asymmetric decoy (“ugly brother”) | 🚫 Reject | No supporting evidence retrieved, and it is manipulation by construction — structurally incompatible with the Anti-Pitch and the fiduciary moat |
B. Own the default
Section titled “B. Own the default”| Strategy | Verdict | Evidence |
|---|---|---|
| Smart defaults (opt-out not opt-in) | ⚠️ Cannot deploy directly | The field’s highest-effect intervention — and MBR has no custody, so it cannot set a default on anyone’s account. Open banking Phase 1 is read-only |
| Opt-out inside MBR’s own flow | ✅ The substitute | Opt-out beat opt-in 22% vs 16% — and the advantage was flat across baseline motivation. The switch arrives already started; the user declines rather than begins |
| Pre-commitment / “auto-refinance if rates drop” | ✅ Best single build | Whoever owns the default owns the outcome: 15pp attrition swing on identical products. Opt-out escalation took sign-up 6% → 49%. Becomes the opt-out annual re-shop, booked to the renewal date |
| Scheduled proactive re-contact | ✅ | Treatment receipt 61% vs 33%, AOR 3.85. The active ingredient is the calendar, not the message |
C. Remove steps — necessary, not sufficient
Section titled “C. Remove steps — necessary, not sufficient”| Strategy | Verdict | Evidence |
|---|---|---|
| Pre-filled paperwork, finishing in-session | ✅ | FAFSA assistance +15.7pp filing (enrolment 34%→42%) at 8 minutes and $88/participant; pre-filled return form 3% → 12% |
| Never end a session with homework | ✅ | FAFSA’s near-complete forms mailed home “were not actually filed unless applicants followed up.” Exit criterion is a submitted application |
| Chunking / micro-steps | ⚠️ | Right direction, wrong variable — the binding constraint is total fields, not their arrangement. US checkouts average 23.48 form elements vs an ideal of 12–14. Hard-cap MBR at 14 |
| Frictionless data portability / open banking | ⚠️ | Phase 1 is read-only and screen scraping is legislated to end. But prefill acceptance is near-universal when offered (>90%) |
| Reduce identity friction | ✅ | IRS Direct File: account creation cost −37.5% of users already told they qualified; form completion cost a further −59.3%. Defer KYC past value delivery. Passkeys: 93% vs 63% sign-in success |
| Fogg B = MAP (raise Ability, not Motivation) | ⚠️ | Correct framing, weak citation — a 2009 conference paper applied as a design model, with no effect size behind the claim. Cite the model; never quote it |
D. Framing and content
Section titled “D. Framing and content”| Strategy | Verdict | Evidence |
|---|---|---|
| Implementation intentions (if-then) | ✅ Ship it | d = .65 overall, d = .61 specifically on failures to get started. Plan capture beat a confirmation tap +4.1pp vs +2.0pp (ns); date and time +4.2pp vs date-only n.s. Use the literal if-then format |
| Goal-specific message content | ✅ | Reminders naming the specific goal were ~2× more effective than those that didn’t |
| Plain feedback + written next steps, no counselling layer | ✅ | Cochrane: longer counselling adds little. SIPS: 20 minutes of counselling vs a leaflet — OR 0.78, null |
| Loss-aversion framing (“you’re paying an extra $150/mo”) | ⚠️ | Free, and worth testing — but Karlan found gain vs loss framing not significant in reminders. Test, don’t assume |
| Temporal reframing (“pennies a day”) | ⚠️ | Dollar framing beats percentages directionally, but at MBR’s tier the evidence is weak: personalised dollar gain plus urgency at n=124,000 could not be distinguished from zero |
| Vividness / future-self connection | ❌ Unevidenced | Nothing retrieved measures this in a financial switching context. A hypothesis, not a lever |
| Social proof / descriptive norms | ⚠️ Constrained | A truthful national norm says most Canadians don’t switch (12–18%) — which backfires. Use in-cohort only: “of MBR users who saw this alert, N% moved” |
| Cognitive salience of an immediate win | ⚠️ | Immediacy and certainty of confirmation transfer from contingency management; the payment does not — CM’s effect rests on objective verification MBR cannot obtain without custody |
E. Timing and momentum
Section titled “E. Timing and momentum”| Strategy | Verdict | Evidence |
|---|---|---|
| Fresh-start effect | ✅ | Gym visits +33.4% at a new week, +14.4% new month, +7.5% after a birthday |
| Messenger identity | ✅ Worth ~2× | Identical letter, identical offer: supplier-branded 26.9% vs regulator-branded 15.0%. MBR is neither — the weakest position in that experiment. This is the quantitative case for the Chilton partnership |
| Embed the ask in an adjacent transaction | ✅ | Legacy ask placed inside will-writing: 4.9% → 15.4%. Canadian analogues: mortgage renewal, tax filing, closing, insurance renewal |
| Endowed progress (“start at step 3 of 7”) | ❌ | No field effect size on a financial outcome retrievable — loyalty-card evidence only |
| Micro-incentives for milestone completion | ❌ | Gamification showed no significant difference in weekly saving; gamified investing platforms increased risk-taking |
| Sunk-cost momentum (“you’ve completed 2 of 3”) | 🚫 Reject | Deliberately exploiting a fallacy against the user — incompatible with client-first. The legitimate version is Hawthorne framing (+2.5pp): “we’re tracking your follow-through” |
5. The Kill List
Section titled “5. The Kill List”Evidence says do not build:
A savings calculator as the product · a “yes, I’m interested” tap as the alert’s terminus (structurally the self-prediction arm — ATT +2.0pp, ns) · any conversational persuasion layer · motivational interviewing as core design (MBR cannot deliver the multi-session dose; the warmth hypothesis was tested directly and not supported) · stages-of-change segmentation (zero demonstrations that stage-tailoring improves outcomes) · anything requiring return visits to an app (median 4.0% daily open, 3.3% 30-day retention across 93 real apps — healthy consumers, not patients) · habit / streak mechanics (no substrate — the behaviour recurs every few years) · public pledge features (announcing an identity-relevant intention can reduce action) · confidence-building copy (self-efficacy does not consistently moderate) · social-pressure disclosure (the highest ROI measured anywhere — ~8pp at $1.93/vote — and rejected: it generated complaint calls, and MBR has no civic mandate).
6. F.A.S.T., Revised
Section titled “6. F.A.S.T., Revised”F.A.S.T. is canonically Fast, Adaptive, Simple, Tailored — SSOT: Core/Misc/Glossary (Talbot, 2026-08-15). “Fast” (quick and easy) is the behavioural-solution KSF; a 2026-08-14 pass in this file drifted the “F” to Frictionless — corrected here. The archived Gemini research used a different earlier expansion (Fast, Adoptable, …) — that file is frozen prior art and deliberately not edited; treat its wording as superseded, not as an alternative.
Against BIT’s evidence-derived EAST framework:
| Element | Maps to | Status |
|---|---|---|
| Fast | EAST Easy — reduce effort | ✅ Direct — the strongest external validation MBR has |
| Simple | EAST Easy — simplify messages | ✅ Direct |
| Adaptive | EAST Timely, and extends it | ✅ Genuine advance — firing at an individual’s threshold crossing, not a population calendar event |
| Tailored | EAST Attractive (personalisation) | ⚠️ Largely duplicates it; make the advance claim for Adaptive alone |
| — | EAST Defaults | ❌ Largest gap, and structural — the highest-effect sub-principle, unavailable without custody |
| — | EAST Social | ❌ No F.A.S.T. counterpart |
| — | MINDSPACE Messenger | ❌ No counterpart — and worth ~2× (§4E) |
Honest assessment: two of four F.A.S.T. elements map onto sub-principles of one of EAST’s four. Half the framework covers a quarter of the evidence base. Defensible as a deliberate bet — friction is the stated diagnosis of the target user — but state it as a bet, not as coverage.
Talbot’s defence, from BIT’s own text: BIT abandoned MINDSPACE because nine elements were too many for busy people to hold in mind, and found that “removing even the tiniest amount of friction” outperformed what the academic literature emphasised. F.A.S.T. independently landed where BIT landed. To make it evidentially defensible, MBR needs a named measurable outcome per element, in-product holdout capability, and a revision log recording where a principle failed and was changed.
7. What MBR Actually Ships
Section titled “7. What MBR Actually Ships”Ranked by effect after replication discount × build cost × deployability with no custody:
- Opt-out annual re-shop, booked to the renewal date — converts MBR from an alert product into a default-owning one
- Assisted completion ending in a submission, never in homework
- One named recommendation + your dollar figure + a short real deadline + minimal input (the Ofgem recipe)
- Fire on the mortgage renewal date
- Plan capture — day, time, what you’re doing beforehand — instead of a confirmation tap
- Second ask to non-responders, on a schedule
- Attack identity friction, then form length — Stage-1 no-KYC MVP is the critical path; cap forms at 14 elements
- Instrument completion, not consent — north star is dollars of rate differential realised on a live account
Requires a partner: the messenger (Chilton) · the adjacent-transaction moment · embedded distribution · true defaults · write-access execution (open banking Phase 2).
8. Standing Cautions
Section titled “8. Standing Cautions”- Replication discount. Published behavioural effects run ~4× inflated (8.7pp in journals vs 1.4pp in nudge units, with statistical power explaining the entire gap). Plan on 1–3pp absolute / 5–15% relative for anything communication-only. Bias-corrected, the “information” and “assistance” categories fall to d = 0.04 — and finance is a domain specifically flagged for left-truncation.
- The registered-intention illusion. “User tapped yes” is not an outcome. Wills: 35% want a legacy, 7% have one. Wales: 38% registered, consent didn’t improve for a decade. Danish electricity: intention +27.5pp, behaviour +1.3pp.
- Spread dependence is the existential risk. Flipper — UK, no custody, £30/yr, fee charged only when it found ≥£50 of savings, zero supplier commission — is MBR’s structural twin, and it closed in September 2021 when price dispersion collapsed. The commission-funded competitor died at the same time of the same cause. The funding model is exonerated; dependence on a spread MBR does not control is not. Model the subscription against the narrowest historical Canadian dispersion, not today’s.
- Three named MBR mechanisms are currently unevidenced — Guaranteed Benefit, Anti-Pitch, Net Worth Thermometer. Keep them (they are genuine commitments, and the Thermometer has a good second justification as a non-interruptive presence layer), but label them hypotheses in investor material rather than selling them as conversion levers.
- “Save More Tomorrow won a Nobel” is false as stated. Prizes go to people, not programs. Safe copy: “…developed by Richard Thaler, who received the 2017 Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel for his contributions to behavioural economics.”
Related
Section titled “Related”- Behavioral-Knowing-Doing-Gap — Round 1 evidence base: 50 strategies, replication-discount rules, 79 sources
- Behavioral-Barriers-Playbook — Round 2 cross-industry playbook: 48 mechanisms, funnel analysis, kill list
- behavioural-solution-research — the research task thread and open decisions
- Core-Philosophy · Key Success Factors · Unique Selling Proposition · Core Value Propositions
Acronyms and terms used above — every occurrence links here rather than repeating the definition inline:
- USP — Unique Selling Proposition
- CASS — Current Account Switch Service (the UK bank-switching guarantee scheme)
- RCT — Randomized Controlled Trial
- Ofgem — UK’s energy market regulator (ran the collective-switching trials cited here)
- FCA — UK Financial Conduct Authority (financial-services regulator)
- AER — American Economic Review (academic journal, citation only)
- AOR — Adjusted Odds Ratio (a statistic measuring effect size, adjusted for confounders)
- KYC — Know Your Customer (mandatory identity-verification step for financial accounts)
- Fogg B = MAP — behaviour-design model stating Behavior happens when Motivation, Ability, and a Prompt converge (B.J. Fogg, Stanford)
- SIPS — the cited UK primary-care brief-alcohol-intervention trial (Kaner et al., BMJ 2013); exact acronym expansion not verified, cited by trial name only
- OR — Odds Ratio (a statistic measuring effect size; 1.0 = no effect)
- ATT — Average Treatment effect on the Treated (a causal-inference statistic)
- pp — percentage points (an absolute, not relative, difference)
- ns — not statistically significant
- BIT — UK Behavioural Insights Team (“Nudge Unit”) — originated MINDSPACE, then EAST
- EAST — BIT’s current framework: make it Easy, Attractive, Social, Timely
- MINDSPACE — BIT’s earlier, retired 9-lever framework: Messenger, Incentives, Norms, Defaults, Salience, Priming, Affect, Commitments, Ego
- MVP — Minimum Viable Product
- FAFSA — Free Application for Federal Student Aid (US) — the Bettinger/H&R Block study cited throughout
- CM — Contingency Management (addiction-treatment technique: pay/reward a target behaviour directly, verified objectively)
- Hawthorne framing — telling someone “we’re tracking this” changes behaviour on its own, independent of any other intervention (named for the 1920s-30s Hawthorne Works productivity studies)
- Flipper — UK savings-switching app, no custody, £30/yr, commission-free, closed September 2021 when price dispersion collapsed — MBR’s closest structural precedent (§8)