Bank the licensed practitioner first, and use "modern entrepreneur" to fill the funnel

Executive summary

The case rests on four findings and one boundary.

One. Read the market through two lenses — turnover tiers to size and target, behaviour to design. Turnover splits the market into four bands the slide and the P&L are keyed on, with the smallest split on credit quality: T1a side-hustle & sub-scale micro (<$30k, risk-gated / sub-prime-leaning), T1b full-time gig & micro ($30k–$100k, bankable), T2 established solo & owner-operator ($100k–$500k), and T3 small business with a team ($500k–$2M) — profiled with their business types in §4, with medium enterprise ($2M+) held out of scope as too large for the launch and the incumbents' turf. But turnover alone groups businesses that behave nothing alike, so within the bands four behavioural segments carry the distinct jobs and values: solo service, licensed solo practitioner, micro-employer, and platform-native seller. The bands frame the wallet; the behaviours define what to build.

Two. The economics rank the four cleanly. A bottom-up three-year CLV per client — deposit margin, commercial-card interchange, and a paid workflow tier, net of cost-to-serve, acquisition, verification, and expected loss — ranks them licensed practitioner ≈ $3,765 ≈ micro-employer ≈ $3,520 > platform-native seller ≈ $2,075 ≫ solo service ≈ $615 (illustrative arithmetic on cited inputs). The practitioner and micro-employer carry the P&L; solo service is one-sixth the value of the leaders. The single largest swing on every number is whether the card ships on a commercial credential (~1.9% net; 2.00–2.35% gross scheme rate) rather than a consumer one (~0.65% net; ~0.77% gross) — worth up to two-thirds of CLV, and a launch-scope decision, not a card-operations detail.

Three. Only one segment is both worth winning and winnable. On a nine-box of attractiveness against ability-to-win, the licensed practitioner is the sole high-high: top CLV, sticky real balances, an unowned white space (trust and purpose-named accounts that no Canadian provider we reviewed offers), and the cheapest, fastest-payback acquisition channel of the four (the accountant or lawyer as paid recommender, at a cost that falls as the partnership scales). Micro-employers are highly attractive but not yet winnable — cash acceptance and payroll are owned by the Big Five and credit unions, and a branchless bank cannot take cash. Solo service is highly winnable but low-value: the doorway, not the destination. Platform-native sits in the middle on both.

Four. "Modern entrepreneur" is a positioning cross-cut, not a fifth segment. Read as leadership most likely means it — the digital-native solo founder, creator, and platform seller — it selects the two lowest-value cohorts in the model. The label wins attention and maps to the kind of consumer-brand halo a digital challenger can carry; the P&L is carried by the balance-holding segments that qualify under the behavioural reading ("the owner still does the books"). Market to the cross-cut; bank the boxes.

The boundary that makes this defensible against the incumbents. The Big Five lead the sub-five-employee market and defend the established practice hard — Scotiabank most of all, whose $115M ten-year affinity deal with the Canadian Medical Association already owns physicians. It does not defend the digital formation moment: as of August 2026, none of the Big Five distributes an account at the point of incorporation, and their published account-opening times run to five to ten days. The defensible line is behavioural — the challenger serves the owner who does their own books; the incumbents serve the business with a finance function, cash, and a credit relationship — and it holds even when both banks see the same five-person company.

Five tests to run before scope freezes:

1. Accountant-channel access for non-physician professions. The beachhead economics ride on the accountant and lawyer recommender. Scotiabank's CMA deal covers physicians; the accessible white space is dentists, lawyers, and accountants. Confirm none of their colleges or associations is already tied to a competing bank-affinity deal. This single fact gates the beachhead.
2. Commercial vs consumer card credential. Put the credential decision in front of whoever owns card economics before scope freeze. On a consumer range, two of the four segments quietly demote to acquisition-only.
3. Trust accounting — feasibility and balances. Confirm a purpose-named account can actually satisfy Canadian law-society trust rules (the winnability hook rests on it, and its only proof point, Relay, is US-regime), and that practitioner balances are operating funds — trust interest accrues to provincial Law Foundations, so a trust-dominated balance overstates the #1 CLV.
4. Professional-corporation formation counts. No public source isolates them. Close the sizing gap via provincial professional-order registries before §7 hard-commits.
5. Do the variables actually discriminate? Pull four to six weeks of onboarding data and test whether entity type, payroll intent, and payout mix split customers by real early-account behaviour — or the segmentation collapses back into firmographics.

Situation: the pool is the ~450,000 businesses formed each year, not the base that already banks

A Canadian digital challenger is preparing a business-chequing launch for October 2026, into a market the incumbent Big Five already lead. Scotiabank holds roughly 17% of Canadian small-business banking and leads the sub-five-employee band by count. Switching is rare — no Big Five bank moved more than about two points of share across the last two published survey cycles, and Canadian small-business banking satisfaction has fallen for two straight years (652 of 1,000 on J.D. Power's 2025 index, with 60% of owners financially unhealthy) without moving the market. Owners under cash pressure do not spend a week re-papering mandates. The installed base is sticky and unhappy, which means it is the wrong target. The contestable pool is the businesses that have nothing to move yet — the ones formed this month.

How large is that pool? The answer depends on which entity you count, and the honest figure spans a wide range. Three public series that do not measure the same thing:

• ISED Key Small Business Statistics 2025 confirms 1,099,521 employer businesses (December 2024), 59.1% at 1–4 employees (≈649,780), and formations averaging 105,001/year (2018–2022), 91.2% born at 1–4 employees ≈ 95,725/year. This is the ~96,000 figure the prior briefing used.
• StatCan's Entrepreneurship Indicators of Canadian Enterprises, 2023 — a newer, single-year product on a different enterprise definition — counts 83,770 enterprise births in 2023, about 20% below the KSBS 2018–2022 average. Two products, two definitions, one year against a five-year average: this is a signal to re-verify, not a confirmed decline. But it means ~96,000 may be a high-water mark rather than a current run-rate.
• The non-employer universe the prior briefing left unsized is the larger prize. StatCan's Business Register (December 2025) reports 3.67 million non-employer businesses with $30,000+ revenue — effectively a GST/HST-registrant count, since $30,000 is the CRA mandatory registration threshold. A separate person-count from the Labour Force Survey puts ≈1.9–2.0 million self-employed non-employer individuals (2023): of Canada's ~2.6M total self-employed, 46.2% are unincorporated with no employees (≈1.2M) and 25.7% are incorporated with no employees (≈682k), and together these two groups make up the ≈1.9M non-employer count. The two figures diverge because one counts entities and one counts people; for "who would open a business chequing account to run a livelihood," the LFS ≈1.9M is the better anchor, and the Business Register 3.67M is the better count of "distinct entities a bank could theoretically serve." Use both, labelled — do not sum them.

Reconciled into a single market frame (TAM / SAM / SOM):

The revision to the prior briefing is directional and material: the contestable pool is not ~96,000 employer formations. It is roughly 400,000–550,000 formations and entries a year — four to five times larger — once the non-employer and self-employed cohort is counted, and that cohort sits almost entirely in the solo-service and platform-native segments. The non-employer entry range (300,000–450,000) is derived by applying StatCan population-transition rates to the ≈1.9M stock and is explicitly illustrative, not a clean formation count. The employer-formation figures are hard public data; the non-employer figures are order-of-magnitude.

Two conditions shape what can be sold. The segment is under cost pressure — inflation is the top concern for 54% of Canadian small businesses, tariffs newly at 40% — so a proposition built on saving money and admin time fits, and one built on premium features does not. And two market changes land alongside the launch: Payments Canada's Real-Time Rail (instant account-to-account payments) and the first phase of Canada's Consumer-Driven Banking regime (Bill C-15, Royal Assent 26 March 2026), whose read-access phase lets a challenger pull an incumbent account to accelerate switching. One competitive fact overshadows the rest: EQ Bank launched a business account in October 2025 — licensed, CDIC-insured, sole-trader-eligible on day one, no fees, 2.5% interest at launch (2.25% by early 2026, tracking the policy rate). That is the closest thing to the challenger's own launch position, and it is already live.

Two lenses: turnover tiers frame the wallet, behaviour defines the need

The segmentation runs on two lenses that answer different questions, and the launch needs both. Turnover tiers (§4) size the wallet and reach the customer — they set how large a banking relationship can become, gate credit appetite, and are the axis leadership, the slide, and the public data are keyed on. Behavioural segments define what to build — because turnover alone groups businesses that behave nothing alike. A sole proprietor selling handmade goods online and a five-person manufacturer can occupy adjacent turnover bands yet share almost none of the same banking needs: one needs a payout feed from Shopify reconciled against one set of books; the other needs payroll, cash deposit, and a line of credit. Segment on turnover to size and target; segment on behaviour to design the proposition.

The behavioural lens is needs-based: group owners by the job they are hiring a bank to do, and assess profitability only after the groups are drawn — not before, and not as the sorting variable. This is the jobs-to-be-done discipline, and it is the right basis for a product segmentation whose output is a roadmap, because behaviours map to needs and needs point to what to build.

Within the behavioural lens, four inputs do distinct jobs, so the two-lens frame does not collapse into a single cut:

• Needs / jobs-to-be-done defines the behavioural segments. Each is a distinct cluster of jobs: get paid; do not get caught by an unreserved tax bill; keep trust money separate; make payroll; reconcile platform payouts.
• Behavioural variables operationalise them — the observable proxies (entity type, payroll presence, payout sources, balance pattern) the challenger can capture at onboarding and in-account to assign a real customer to a segment.
• Turnover tiers and firmographics (the sizing lens) size and reach the segments — turnover band, entity type, province, formation stage — but never define the proposition, because public data and acquisition channels are keyed on them.
• Value / economics is the prioritisation overlay — applied after the segments are drawn, to decide which to serve first.

Every candidate segment is held to the classic five-part test: Measurable (countable in public data), Substantial (big enough to matter to the launch P&L), Accessible (a real channel reaches it), Differentiable (its needs differ enough to warrant distinct treatment), and Actionable (the challenger can build, price, or serve it differently). A candidate that fails Differentiable or Actionable is a descriptor, not a segment — the test we apply to "modern entrepreneur" in §8.

One design principle carries through the whole report: one account, many segments. The challenger builds a single account against the union of behaviours and runs separate acquisition motions per segment. The segmentation drives the roadmap priority of capabilities, where the paid tier is priced, the verification path, and the acquisition channel — not four different products.

The jobs owners hire a bank to do: remove the admin, and make cash flow legible

The average Canadian small-business owner spends about five hours a week — 260+ hours a year — managing finances, and 40% rank administrative burden a top-three business challenge, ahead of finding customers. Only 6% use a bookkeeper; the owner is the bookkeeper. Separately, Canadian small businesses wait ~29.8 days to be paid, invoices land ~11.6 days late, and 43% name cash flow their single biggest financial challenge. The bank is being hired against two master jobs — remove the admin, and make cash flow legible and timely — and neither is predicted by revenue or headcount. The segments differ in which jobs dominate and how intensely.

Judged against what a one-to-five-person business can actually get from a Canadian provider today (not what US or UK specialists ship), the jobs map is this:

Segment by segment, the defining job is clear. Solo service is hired to get paid and not acquire a tax bill I didn't reserve for — the surprise April bill is the shock every self-employment specialist abroad (Found, Lili) built its lead feature against, each skimming a set percentage of every deposit into a tax bucket on arrival. The licensed practitioner's defining, non-negotiable job is to keep trust money separate from operating money — for lawyers this is a regulatory obligation with a mandatory monthly three-way reconciliation and province-specific deadlines (Ontario 25 days, BC 30 days plus an annual trust report), a heavy, recurring, willingness-to-pay job that no Canadian digital provider maps to. The micro-employer's jobs are make payroll on time and know what's left after payroll and tax are reserved — plus cash deposit, which decides whether a branchless account can be primary at all. The platform-native seller's defining job is to reconcile multi-gateway payouts that arrive separately, netted of fees and refunds, on different timings — an estimated five to seven staff-days a month of manual bookkeeping.

Five jobs are underserved by every Canadian provider we reviewed at once — the roadmap's white space: (1) automatic GST/HST + income-tax set-aside; (2) purpose-named / restricted accounts (the "what's spendable" and trust-separation jobs); (3) day-one onboarding integrated with that workflow layer — challengers now offer day-one, sole-trader-eligible eligibility (§6: EQ Bank, Venn, Float), but none pairs it with tax set-aside and purpose-named accounts, and the Big Five still take five to ten days; (4) native platform-payout ingestion and reconciliation; (5) cash acceptance — a structural gap that excludes the highest-balance micro-employers and needs a partner answer, not a build.

Ranking the jobs by intensity, frequency, and willingness-to-pay puts two at the top: tax set-aside (broad, high-frequency, universal) and trust / purpose-account separation (narrow but regulatory, highest willingness-to-pay). Day-one onboarding scores low as a paid feature — it is free and one-time — but it gates everything: nothing else can be sold until it is solved, so it ranks first operationally.

Segmentation by turnover tier: four bands, and the businesses inside each

Leadership and the launch slide need a segmentation that both sizes the wallet and reaches the customer — and both are keyed on turnover. This section cuts the addressable market into four turnover bands, profiles the real business types inside each, and then overlays the four behavioural segments (§2) that discriminate within the bands. Turnover frames how large the banking relationship can become; behaviour defines what to build for it. The two are complementary, not rival: a band sets the ceiling on deposits, card spend, and credit appetite; the behavioural cut names the job and points to the proposition. Medium enterprise ($2M+, a finance function forms) is deliberately out of scope — too large for the launch and the incumbents' turf — so the analysis runs from the smallest micro band up to small businesses with a team. And because the smallest band is not one thing, it is split on revenue into a sub-scale, higher-risk half and a bankable half — the credit-quality split leadership should see before funding a funnel.

The turnover bands (annual turnover, Canadian context; GST/HST registration becomes mandatory at $30,000 of revenue — the line used to split the smallest band):

Turnover bands and populations are order-of-magnitude and illustrative, not a census; they overlap at the edges. T1a, T1b and T2 partition the ~4.3M-business micro TAM (§1); T3 sits just above it as the up-market sequence, and the medium-enterprise / finance-function band above $2M is out of scope entirely (the Scotiabank boundary). A unit caveat travels with the table: the T1 bands are person-counts (Labour Force Survey), while T2–T3 are entity-counts (Business Register), so the bands are an approximate reconciliation, not a clean sum (§1 warns the two bases "do not sum").

T1a — Side-hustle & sub-scale micro (turnover < ~$30k)

Part-time and supplemental earners below the GST-registration line — a second income, not yet a business. Balances are thin and volatile, and a large share of accounts go dormant within months. Instant, sole-trader onboarding concentrates its adverse-selection and fraud risk here, so this is the sub-prime-leaning half of the funnel: much of it will not clear acquisition cost and expected loss. Serve it on a risk-tiered flow (velocity limits, delayed full-limit activation), acquire it at zero marginal cost, and treat it as optionality, not revenue.

So what — T1a. The sub-prime-leaning half: high volume, high dormancy, negative expected value after acquisition cost and loss on much of it. Its role is optionality — a near-free top-of-funnel from which a minority graduate into T1b and T2 — not revenue. Gate it with risk-tiered onboarding and delayed limits; never underwrite it or fund the launch against it. The hypothesis to test in-market: what share of T1a accounts activate, stay active past six months, and cross $30k — that share is the only part worth an acquisition dollar.

T1b — Full-time gig & micro (turnover ~$30k–$100k)

GST-registered, primary-livelihood earners — a real business, run full-time. Deposits are steadier, the tax-set-aside and payout-reconciliation jobs are genuine and recurring, and a meaningful share will incorporate and cross into T2. This is the bankable half of the funnel — the "actually good" customers worth acquiring for value, not just optionality.

So what — T1b. The good half. Steadier deposits, real recurring jobs (tax set-aside, payout reconciliation), and — most important — the segment that graduates into the T2 beachhead. Its true value is as the on-ramp to the practitioner and owner-operator bands, so acquire it deliberately (tax set-aside + payout ingestion + a clean incorporation path) and instrument for the graduation signal — rising balances, incorporation, a first invoice over $X — that promotes a T1b account into T2. Both T1 bands map to solo-service and platform-native (~$615–$2,075 CLV); the difference is that T1b earns its acquisition dollar and T1a does not.

T2 — Established solo & owner-operator (turnover ~$100k–$500k)

Full-time, one principal, real and retained balances, often just incorporated. This is the beachhead band: the newly-incorporating licensed practitioner lives here, alongside the established sole trader and the scaling solo brand. Willingness to pay for admin automation is real because the owner is still the bookkeeper but the numbers now matter.

So what — T2. The economic core of the launch. Real retained balances, day-one-servable, and reachable through a paid recommender (the accountant/lawyer) at a CAC that falls with scale. The licensed practitioner (rows 1–3) is the beachhead; the rest of the band is the sequence to grow into. Price the paid workflow tier here.

T3 — Small business with a team (turnover ~$500k–$2M)

Two-to-fifteen staff, payroll running, often cash in the till, a credit relationship forming. The highest-balance micro-segment — but the one where a branchless, no-cash, no-credit launch is only a partial fit until the cash-acceptance and payroll answers exist. This is the up-market sequence, not the Year-1 beachhead. The band splits cleanly in two: the licensed professional practices (below), whose money is largely non-cash and which therefore carry T3 value at T2 winnability; and the wider cash/payroll businesses, which do not.

The professional-practice exception — the winnable slice of T3:

The wider cash/payroll businesses (one shared profile). The rest of T3 — cafés and restaurants, salons and spas, boutique retail, staffed trades (HVAC, roofing, auto-repair, construction subcontractors), gyms and daycares, wholesalers, specialty food producers, staffed e-commerce, marketing agencies, and property-management firms — shares one banking profile: payroll to run, cash or card in the till, and a forming credit line. All read Partial at launch (Partial (cash) for the till-heavy trades and retail; Partial (credit) for the credit-led wholesalers and subcontractors), blocked by the same two capabilities — cash acceptance and payroll — until the partner answers exist. The two software-adjacent exceptions (staffed e-commerce, agencies) rate High, since their money is already digital.

So what — T3. The highest-value band per client (maps to micro-employer, ~$3,520 CLV, and the upper practitioner) but the least winnable at launch: cash acceptance and payroll are owned by the Big Five and credit unions, and a branchless bank cannot take cash. Serve the wider band in Year 2–3 by following T2 owners up as they hire, and only once the cash/payroll partner answers exist. The five professional practices above are the exception — they carry T3 value with T2 winnability because their money is largely non-cash.

Above T3 is out of scope. The medium-enterprise band ($2M+, where a finance function, treasury needs, and a real credit relationship form) is the launch's explicit ceiling and the cleanest expression of the Scotiabank boundary — the incumbents' relationship model and credit balance sheet win there, and a day-one deposit account is not the product. The challenger's job is to bank these owners before they cross that line and to hand off cleanly when they do, not to compete for the relationship-banked mid-market.

The behavioural cross-cut within the tiers

Turnover bands size and reach; they do not, by themselves, tell the challenger what to build. Within the bands, four behavioural segments — solo service, licensed solo practitioner, micro-employer, platform-native seller — carry the distinct jobs, KYB paths, and values that drive the roadmap. Solo service and platform-native populate T1a–T2; the licensed practitioner concentrates in T2 (and carries into T3); the micro-employer is the T3 core. The cross-cut and how the challenger tells the segments apart, sizes them, and verifies them:

Three variables carry almost all the discriminatory power, one per segment, with solo service as the residual default. Per the start-simple discipline, the challenger does not need a dozen:

1. Entity type + licensing declaration (day one) isolates the licensed practitioner at the point of application, with no in-account data needed.
2. Payroll intent (day one), confirmed by payroll-run detection (later) isolates the micro-employer — the declaration routes onboarding, the in-account signal corrects under-declaration.
3. Payout-source mix (in-account only) isolates the platform-native seller; there is no reliable day-one proxy, because "I sell on a platform" is a self-report a consultant could also tick.
4. Solo service is whoever remains — unincorporated or simply incorporated, no payroll, no platform-payout signature.

That four-variable core clears the good-segment test: each variable is measurable in data the challenger will already hold, each segment maps to real formation volume, each routes at the exact moment KYB already touches the customer, each variable is near-absent in the other three segments, and each changes the KYB path or the first-session experience. Critically, the same onboarding form does three jobs at once — it satisfies FINTRAC's entity-verification and beneficial-ownership requirements, it routes KYB depth per segment, and it produces a defensible day-one segment tag without waiting for account history. That tag is provisional: within three to six months, in-account signals will confirm or reassign a meaningful share of accounts, most visibly in the platform-native and micro-employer groups where self-declaration is weakest.

KYB depth is a segment attribute, not an afterthought. It ranges from individual-ID-only (solo service) to enhanced, beneficial-ownership-heavy (multi-owner micro-employer), and it feeds directly into the cost-to-verify line in §5 and the ability-to-win axis in §7. Day-one onboarding feasibility is not uniform across segments — which is one reason the highest-value micro-employer is not the easiest to win.

A caveat on the sizing. No public dataset tags businesses by these four behavioural labels; every figure above is a proxy with unavoidable overlap, and the licensed-practitioner count is the softest — no public source isolates professional-corporation formations from all incorporated solo operators, so the ≈75,000–100,000 figure is an over-count. Closing that gap via provincial professional-order registries (law societies, colleges of physicians and dentists) is a named week-one action before the report hard-commits to the practitioner as the beachhead.

Economics and value per segment: the practitioner and micro-employer carry the P&L; the card credential swings everything

Value per client is a bottom-up three-year CLV: deposit balance × margin, plus card spend × interchange, plus paid-tier revenue, times a per-segment retention overlay, minus cost-to-serve, acquisition, verification, and expected loss. This is the single most decision-relevant number in the report — it is the attractiveness axis of the prioritisation — and every figure is illustrative arithmetic on cited inputs, not observed the challenger data.

The inputs that matter and their basis: operating balances anchored on the prior briefing's C$5,000–15,000 band and Starling's ~£12,800/month SME average, set higher for the practitioner (real retained balances) and lower for the platform seller (swept, volatile); a Canadian deposit net interest margin of ~2.0% on non-interest-bearing operating (chequing) balances, held just below the Bank of Canada policy rate (2.25%) that caps near-risk-free deployment — well below Starling's 4.12% UK figure. Rate-shopping savings balances of the kind EQ Bank pays ~2.25% on are out of scope; the model assumes operating funds a primary account holds, not deposits won on rate; commercial interchange netted to ~1.9% (against a ~0.65% consumer alternative); a ~$14/month paid tier; and a per-segment annual retention rate that is the reason the ranking differs from a naïve balance ranking.

The ranking: licensed practitioner (~$3,765) ≈ micro-employer (~$3,520) > platform-native seller (~$2,075) ≫ solo service (~$615). All four clear the 3:1 CLV:CAC bar, but solo service does so at one-sixth the value of the leaders.

Mapped onto the turnover bands (§4), value rises the same way turnover does, with one deliberate exception. T1a and T1b (gig & micro) are the ~$615–$2,075 funnel — T1b the bankable half that earns its acquisition dollar, T1a the risk-gated half that mostly does not; T2 (established solo & owner-operator) is where the ~$3,765 practitioner CLV is booked; T3 (small business with a team) is the ~$3,520 micro-employer band. The exception is what makes the beachhead call: the practitioner carries near-T3 value at T2 turnover, because a newly-incorporating professional holds real balances without yet needing the cash, payroll, and credit that make T3 hard to win.

The top two are a near-tie decided by retention, and this is a conflict worth stating plainly. On raw annual net contribution the micro-employer ($1,586) out-earns the practitioner ($1,507). The practitioner takes first place only because of its higher modelled retention (90% vs 80%, giving 2.71 vs 2.44 active-years). Retention by behavioural segment is itself a named data gap — approximated from the briefing's "digitally active clients churn less" and general business-survival patterns. So the #1/#2 order is genuinely soft. It does not matter for the beachhead call, because the micro-employer's low ability-to-win (cash gap, incumbent-defended — §7) keeps the practitioner the beachhead regardless of which of the two tops attractiveness. A second caveat sits on the practitioner deposit line specifically: Canadian lawyers' mixed-trust interest typically accrues to provincial Law Foundations, not the bank, so the modelled balance must be operating (non-trust) funds. If practitioner balances turn out to be trust-dominated, the deposit line and the #1 ranking are overstated — target operating funds, and verify the mix before committing (test 3 in the executive summary).

Three swing factors, ranked by how much they move the numbers:

1. Card credential — commercial vs consumer — is the dominant swing. Dropping to a consumer credential cuts interchange revenue by roughly two-thirds across all four segments. Because interchange is a different share of each segment's value, total CLV falls unevenly — practitioner −36%, micro-employer −52%, solo service −63%, platform-native −66%. Platform-native economics collapse without a commercial credential — from a P&L-relevant $2,075 to an acquisition-only $705. This is a launch-scope decision, and it belongs in front of whoever owns card economics before scope freeze.
2. Primary-account share hits the balance-led segments hardest. Only a primary account holds the modelled operating balance; Starling runs at 56% primary among SME. Haircutting the deposit line to 56% costs the practitioner −24% and the micro-employer −15%. The practitioner is worth roughly a quarter less if the launch opens accounts but does not win the primary relationship — so the build must be designed to win the primary account, not merely open one.
3. Paid-tier attach is the non-swing. Flexing attach by ±15 points moves CLV by only ~2% in the balance-led segments (practitioner, micro-employer, platform-native), because the paid tier is small beside deposits and interchange; it moves solo service more (~8%), where the tier is a larger share of a smaller CLV. The pricing debate that will consume the most oxygen moves the P&L the least — the tier's job is attachment and primary-account behaviour, not margin.

How the field segments: every winner entered narrow at the solo end and sequenced up — and the account layer has already commoditised in Canada

The benchmark evidence points to one rule. Every SME neobank in our benchmark set (the seven below) started narrow — at the self-employed / solo end that incumbents served worst — and sequenced up-market, never down; none of the seven launched broad or led with an interest rate or with lending. They chose a beachhead they could serve better than anyone, won primary-account status there, then widened the product to follow customers as they grew.

The closest analogue is Mettle, NatWest's free, app-first, separate-brand digital challenger for UK sole traders and small limited companies. NatWest built it on separate systems and a separate brand explicitly to defend the freelancer/micro band against challengers — you cannot open a Mettle account in a NatWest branch. Its proposition — free, bundled FreeAgent accounting, rolling in-app tax calculation — is almost exactly the recommended challenger build. The lesson for a digital challenger positioned against the incumbents: target the segment the incumbents serve worst (self-employed, phone-native, wants admin automated), not compete with the branch network for the relationship-banked base. Two cautions travel with it: Mettle does not accept cash (the same structural gap a branchless challenger inherits), and a separate-brand build only amortises if the beachhead is real.

The Canadian field looks different from the prior briefing, and one correction is load-bearing. Relay is a Toronto-headquartered neobank whose product serves US-registered businesses only — its own eligibility pages require "an operating presence in the U.S." and support US corporations, LLCs, partnerships, and sole proprietors. Its disclosed scale (150,000+ businesses, USD 1.3bn deposits) is US customers. In Canada, Relay contests only the narrow cross-border-seller slice. It should be treated as a product model to copy — purpose-named/trust accounts are its proven hook — not as an entrenched Canadian incumbent. Any earlier framing of Relay as a Canadian challenger is wrong.

That correction reshapes the real Canadian set — EQ Bank, Venn, Float, Keep, Wealthsimple, plus the Big Five and credit unions:

Entity flexibility has largely arrived — a Canadian sole trader can already open a free, high-interest, day-one account at EQ Bank, Venn, or Float. What is still empty is not who can open but what the account does once open. Three acquisition doors are staffed — rate (EQ), accountant (Venn), card/credit (Float, Keep) — but no one owns the formation-moment door (registries, incorporation services, payroll sign-up) as a primary channel. And critically, no Canadian provider we reviewed ships the workflow layer: EQ Bank has the account, rate, and day-one eligibility but no tax logic, no purpose-named accounts, no payroll; Venn and Float sync to accounting software but do not set aside tax.

The white space, ranked: (1) a free same-day account with automatic GST/HST + income-tax set-aside and bookkeeping, for the solo-service sole trader — no Canadian provider offers the combination; (2) purpose-named / trust accounts plus scoped accountant access for licensed practitioners — the Relay model, which no Canadian player replicates; (3) a cash-accepting primary account for micro-employers — a structural gap for every branchless challenger, owned by the Big Five and credit unions; (4) formation-moment onboarding for the pre-revenue business.

Segment prioritisation: the licensed practitioner is the sole high-high beachhead, robust to re-weighting, gated on three facts

Scoring every segment on attractiveness (worth winning) against ability-to-win (winnable by October, for a digital challenger), each axis a weighted set of sub-criteria scored 1–5, weights set economics-first — value dominates attractiveness; onboarding and channel reality dominate ability.

Attractiveness (CLV 40%, contestable pool 20%, growth 15%, deposit-led fit 25%):

Ability-to-win (onboarding/platform fit 25%, channel access 25%, capability-gap 15%, contestability 15%, incumbent boundary 20%):

The attractiveness order reproduces the economics ranking exactly, so the framework is anchored, not free-floating. Plotted on the nine-box (bands: Low 1.00–2.33, Medium 2.34–3.66, High 3.67–5.00):

The map is unusually clean: one segment — the practitioner — is the sole high-high, and the other three each fail exactly one axis. Solo service is highly winnable but low-value (the doorway); micro-employer is highly attractive but not yet winnable (cash and payroll gap, incumbent-defended); platform-native sits in the middle on both.

The beachhead is the newly-incorporating licensed solo practitioner — the professional-corporation formation, a T2-turnover business ($100k–$500k). It is the only high-high: top CLV (~$3,765), sticky real balances, the accountant/lawyer recommender channel with the cheapest, falling-with-scale acquisition cost of the four (§9), and a genuine white space no benchmark or Canadian challenger we reviewed owns (trust and purpose-named accounts). The target is the practitioner at formation — Scotiabank defends the established practice (the CMA/MD deal, the Scotia Professional Plan) but not the newly-incorporating one.

The three-year sequence — earn the right up, never down:

• Year 1 — win the practitioner beachhead, acquire through the solo-service doorway. The same October scope (day-one onboarding + tax set-aside + purpose-named accounts) serves both, so solo service is the top-of-funnel acquired at near-zero marginal cost. Win primary-account status among practitioners — the −24% swing variable — not signups.
• Year 2 — consolidate solo; follow practitioners up into micro-employer as a growing practice hires its first employee. This is when the cash-acceptance / payroll partner answer must be solved — the single capability blocking the high-attractiveness micro-employer box.
• Year 3 — micro-employer proper (once the cash answer exists) and selective platform-native seller — but only on a commercial card credential (the −66% P&L gate) plus a multi-platform reconciliation hook no single platform's own banking product can match.

The beachhead survives sensible re-weighting decisively. Under a growth-first weighting (CLV 25% / growth 30%), the practitioner still tops attractiveness (4.30) and remains the only high-high. Under a brand-fit-first weighting that maximises fit toward the modern-entrepreneur solo+platform cohort, solo service's CLV score of 1 caps it at medium attractiveness — it becomes a stronger doorway, never the beachhead. Remove the Scotiabank boundary entirely and the practitioner gets more robustly high-high, because the boundary is the only thing holding its ability score back. The prioritisation is robust to the one soft attractiveness assumption (the retention-driven #1/#2 order) because the micro-employer's low ability keeps the practitioner the beachhead either way.

Three gates the launch must clear before this hard-commits — the report states them rather than hides them:

1. Accountant-channel access outside the CMA physician deal — the load-bearing gate. The beachhead economics ride on the accountant/lawyer recommender, but Scotiabank's own $115M CMA/MD affinity deal already owns physicians. The accessible white space is dentists, lawyers, and accountants — reached through their colleges and associations, none of which was confirmed free of an existing bank-affinity deal. If those professions are already tied up, the practitioner channel narrows sharply. This single fact, not the channel's general attractiveness, gates the beachhead.
2. The trust-account-interest caveat. The #1 CLV assumes operating balances; lawyers' mixed-trust interest accrues to Law Foundations. If practitioner balances are trust-dominated, the deposit line and the ranking are overstated. Target operating funds; verify before committing.
3. Practitioner sizing is directional. No public source isolates professional-corporation formations; the contestable-pool score is a proxy off "incorporated solo self-employed," an over-count. Close via provincial professional-order registries in week one.

The "modern entrepreneur" decision: adopt it as the acquisition narrative, not as the first segment

Leadership named the "modern entrepreneur" as the first target but has not defined it. Defining it — and testing whether it is the right first focus — is part of the job. There is no single external definition; there is a cluster, and the cluster's shape is the finding.

Analysts converge on a behavioural core — solo/lean, digital-native, platform- or AI-enabled, formed online — and diverge on who. ICSB's lead 2026 MSME trend is "solo founders go big." The solopreneur literature sizes ~29.8M US solopreneurs generating ~$1.7T, with the demographic surprise that ~64% of US solopreneurs are over 45 — "modern" is not synonymous with "young." Creator-economy data counts 200M+ creators worldwide but average income of only ~$44,000. Canadian generational data pulls the other way: a TD survey found 73% of Canadian Gen Z find starting a business appealing, and ~25.5% of Canadian private-sector businesses are immigrant-owned. The sources agree on the behavioural axis (how the owner wants to be served) and split on the demographic and channel axes (who they are and how they earn). Tellingly, no neobank markets "modern entrepreneur" as a segment — each operationalises it as an entity type (sole trader / freelancer / limited company) plus a job (tax, reconciliation, spend control). The word "modern" does the marketing work; account eligibility is drawn on entity type.

Three candidate definitions, assessed:

• A — the digital-native solo / creator / gig / e-commerce operator. Strongest external support; maps to solo service + platform-native seller. Weakness: it selects the lowest-balance, most-volatile cohorts — the two lowest CLVs in §5 (~$615 and ~$2,075).
• B — a behavioural cross-cut ("wants banking done the modern way"). Defined by digital engagement and appetite for admin automation, regardless of entity or industry; a "modern" dentist and a "modern" café owner both qualify. Strength: the only reading that includes the P&L-bearing segments, and the axis every source shares. Weakness: by itself it is a descriptor, not a countable list.
• C — a generational / newcomer / diverse cohort. Millennial/Gen-Z, newcomer, women- and BIPOC-founded. Strength: a real acquisition-and-inclusion narrative that over-indexes on formations. Weakness: it cross-cuts all four behavioural segments rather than defining one, and targeting on protected characteristics carries fair-treatment risk (§10).

Applying the good-segment test to the standalone-segment reading (A treated as a fifth segment) settles it. It is Measurable only through proxies, Substantial on formations, and Accessible at the formation moment — but it fails Differentiable (its needs are the same tax, reconciliation, and onboarding jobs already split across solo-service and platform-native) and fails Actionable (the challenger would build one account and run separate acquisition motions — exactly what the four-segment model already prescribes; there is nothing it builds or prices differently for "modern entrepreneur"). By the methodology's own rule, a candidate that fails Differentiable or Actionable is a descriptor, not a segment.

Recommended working definition, to carry as a cross-cut:

The modern entrepreneur is the owner-operator who is simultaneously founder, bookkeeper, and payroll clerk — who forms the business digitally, expects to bank from a phone on day one, and wants the administrative burden (tax set-aside, categorisation, reconciliation) automated away.

It includes sole proprietors and single-director corporations where the owner does the books, freelancers and solo practitioners incorporating now, platform sellers needing payout reconciliation, and 2–5-person micro-employers where the owner still runs payroll personally (qualifying behaviourally, not by size). It excludes businesses with a dedicated finance function (the Scotiabank boundary), cash-intensive trades that need branch cash acceptance, funded startups with a controller, and pure demographic membership without the behavioural traits (the fair-treatment guardrail). The include/exclude line is drawn on behaviour and entity type, never on the demographic label alone.

The digital-native deep-dive confirms the economics warning. The Canadian digital-native cohort is real but low-value: the closest public balance proxy (Lili's ~$1,500 average monthly deposit) is an order of magnitude below the practitioner/micro-employer balances, income is volatile and swept out to cover living costs, and a meaningful share of the StatCan-verified 667,000 platform earners and the broader ~6M self-reported gig participants are not running a business that needs a dedicated account at all. Its distinguishing feature is intensity of payout fragmentation, not a new job. The specialist hooks that win it — native payout ingestion, tax set-aside, near-interbank FX for the cross-border minority — are worth building early because they are cheap relative to their acquisition value and they defend against Found/Lili/Novo encroachment, but the launch must not be scoped or funded as if this cohort clears payback on its own.

The demographic cohorts are acquisition and inclusion lenses, never segmentation or pricing inputs — and this is a hard line. Newcomer/immigrant founders are the highest-weight lens: ~816,000 immigrant-owned businesses, immigrant-owned incorporations were over half of all net-new incorporations in 2010–2020, and the cohort carries a specific banking friction — a thin or absent Canadian credit file (79% of recent newcomers report difficulty building Canadian credit) — that a day-one-eligible, statement-underwritten account is structurally positioned to solve. That is a product insight (accept non-Canadian credit history for every applicant), not a targeting rule. Millennial/Gen-Z founders confirm the digital-native cross-cut and concentrate in solo-service and platform-native; their distinctive insight is behavioural (67% do not proactively seek bank support, so distribution must sit inside the platforms and registries where they already are). Women-owned and Black/BIPOC-owned founders are acquisition-channel and inclusion priorities — both concentrate in the lowest-balance solo-service segment, so their economics run through the solo-service line, treated as upside, not P&L drivers. None of these cohorts is a fifth segment; each cross-cuts the four, and none may enter KYB-routing, pricing, or eligibility logic.

Acquisition by segment: the accountant channel wins the beachhead — if it is accessible past the physician deal

Distribution beats brand spend at this revenue-per-account, and no benchmark built its book primarily on paid media — all layered a small number of high-intent, formation-adjacent channels. The channel that wins each segment differs because each passes through a different formation ritual.

The accountant channel is the one with the clearest, most quotable economics and the best fit for the beachhead. Relay's accountant Partner Program pays $50 per approved account (a firm with 1–2 clients) up to $300 per account (50+ clients), plus a 5–10% monthly revenue share on interest earned — a structure where CAC falls as the partner scales, the opposite of paid media, and where the accountant pre-qualifies the client (already incorporated, already has a real operating need), lowering both CAC and expected verification cost. Tide runs the same structure in the UK. The channel matters because the accountant is the second decision-maker: US survey data shows 47% of small businesses lean on accountants and advisors, and 27% turn to accountants specifically when applying for credit — no confirmed Canadian equivalent exists, a gap to close, but the direction is consistent with the practitioner's top job ("hand clean books to my accountant").

The load-bearing risk sits inside this channel. The single most directly relevant Canadian precedent for the practitioner segment is not a neobank — it is Scotiabank's own ten-year, $115M affinity partnership with the Canadian Medical Association and MD Financial Management, the Scotiabank Healthcare+ Physician Banking Program. An incumbent already runs the highest-value affinity channel into exactly the segment named as the beachhead. A challenger cannot easily contest physicians against that entrenched, decade-long deal; the accessible white space is dentists, lawyers, and accountants — professions the CMA deal does not cover — reached through their regulatory colleges or provincial associations, none confirmed free of an existing bank-affinity deal. This is the week-one test that gates the beachhead (test 1 in the executive summary).

The other channels are weaker fits. The incorporation-bounty channel that would suit micro-employers is the tightest Canadian precedent — Ownr/RBC pays $49 (sole prop) / $300 (corporation) on a funded account — but Ownr is RBC-owned, so the challenger cannot buy distribution through it and would compete against a Big Five bank's captured channel. The payroll channel (Wagepoint, Payworks) is currently a pass-through, not a funnel — a referral or default-account partnership is unclaimed white space, timed to the RP-account moment. The platform channel is worsening: Shopify Balance and Square Checking are moving to own the payout account themselves, so the challenger competes for the payout default against the platform's own product. Its narrow, defensible wedge is multi-platform reconciliation across Shopify + Stripe + Square + marketplaces in one place — which no single platform's own banking product can offer, since each sees only its own payouts.

Risks, regulatory, and data constraints: instant onboarding is compliant only if it is risk-tiered, and demographics may never enter the KYB or pricing logic

Four constraint sets bound how the challenger may segment, target, and price — and what could invalidate the segmentation.

AML/KYB — verification depth is a segment attribute, and speed must come from automating a check, not skipping one. Every account sits under the PCMLTFA and FINTRAC guidance. Entity verification differs by legal form — the primary KYB variable behind the segmentation — from verifying the individual owner of a sole proprietorship (lightest) to confirming beneficial owners with 25%+ control of a multi-owner micro-employer (heaviest, and now requiring discrepancy flags against public corporate registries since October 2025). OSFI Guideline E-13 and the PCMLTFA require a documented, risk-based program: every client risk-rated with due diligence proportionate to the rating. Instant, sole-trader-eligible onboarding is compliant only if the speed comes from automating a risk-appropriate check — a risk-tiered flow, not one fast lane for everyone. Adverse selection at instant onboarding is a named, foreseeable risk (fast opening attracts accounts used to move proceeds before genuine volume arrives); the mitigation is behavioural — velocity limits and delayed full-limit activation keyed to each segment's expected pattern.

Privacy — PIPEDA and the Consumer-Driven Banking Act scope what data may build segments. Under PIPEDA, the behavioural variables the segmentation proposes (transaction mix, payout sources, engagement frequency) must be disclosed as part of the account's stated purpose at opening, not retrofitted, and building a propensity model for a new purpose likely needs a fresh consent event. The Consumer-Driven Banking Act (Bill C-15, Royal Assent March 2026) adds a constraint on data the bank did not originate: data received via open banking — for example, to pre-fill a switch — is scoped to that request's consented purpose and cannot be repurposed to assign a marketing segment. Consent scope must be tracked per data source: account-opening data, in-account behavioural data, and open-banking data carry different legal bases, and Phase 2 (mid-2027, write access) widens what becomes possible.

Fair treatment — segment on behaviour and need, never on protected characteristics, even where they correlate. The Canadian Human Rights Act s.5 prohibits adverse differentiation on prohibited grounds and applies to federally regulated banks; the Financial Consumer Protection Framework adds a positive duty to offer products appropriate to a customer's needs. The demographic cohorts in §8 are legitimate as an acquisition and inclusion narrative. They become a legal and reputational risk the moment they are used as an input to segment assignment, pricing, or verification depth — a KYB flow that runs slower because a name, formation channel, or ID document correlates with newcomer status is indirect discrimination even without intent. The discipline: every candidate behavioural variable must be checked for whether it plausibly proxies age, gender, ethnicity, or newcomer status before it routes onboarding speed, pricing, or credit. Entity type, payroll presence, and payout source are defensible; anything correlated with name origin, community-specific formation service, or ID-document country is not, however predictive.

Segmentation-invalidation risks and the tests that close them:

Implications and week-one tests: what to build, price, and prove before October

What to build — the roadmap order the jobs map and economics dictate:

1. The formation gate (launch-critical). Day-one, sole-trader-eligible onboarding from a phone, with business/personal separation, delivered as a risk-tiered flow. This gates every segment and is the acquisition claim. No account, no launch.
2. The two paid-tier anchors. Automatic GST/HST + income-tax set-aside (broad, universal, and Canada-specific logic a vanilla core platform will not ship natively), and purpose-named / restricted accounts (which serve "what's spendable" for everyone and the regulatory trust-separation job that makes the beachhead winnable — subject to confirming a purpose-named account can actually satisfy law-society trust-accounting rules, not merely label a sub-account; see test 3). Price the ~$14/month tier here.
3. The segment-specific hooks, sequenced. Native platform-payout ingestion for platform sellers (cheap relative to acquisition value; defends against US specialists), and a cash-acceptance partner answer for micro-employers (a Year-2 partner decision, not a build, and the gate on the highest-attractiveness box).

What to price. Give the account away; charge for the workflow tier above it, anchored on the two paid jobs. The tier's job is attachment and primary-account behaviour, not margin — the pricing debate moves the P&L by ~2%. The credential decision, by contrast, moves it by up to two-thirds: confirm the card ships commercial before scope freeze.

How to verify segment assignment. Build the day-one form (entity type, licensing declaration, payroll intent, requested account structure) and its routing logic for October; instrument the in-account confirming variables from day one even though they score no one yet; treat the first six to nine months as the validation set for the Phase-2 analytics model.

The organisational question the launch must answer. Mettle (separate brand, separate systems) and Monzo Business (same core, same brand) are the two models. Which one the challenger picks determines whether the defensible boundary with the incumbents is drawn cleanly. Whichever it is, the boundary is behavioural — the owner who does their own books — not a size line.

The five tests that gate the beachhead before scope freeze (restated from the executive summary because they are the near-term equivalent of the briefing's "hypothesis to test in week one" discipline):

1. Accountant-channel access for dentists, lawyers, and accountants — is the white space past the CMA physician deal actually open, or already tied to a competing bank? Gates the beachhead.
2. Commercial vs consumer card credential — a launch-scope decision worth two-thirds of two segments' CLV.
3. Trust accounting — feasibility and balance mix — can a purpose-named account meet Canadian law-society trust rules (monthly three-way reconciliation, statutory deadlines, annual trust report), and does the #1 deposit line survive the Law Foundation caveat (balances must be operating, not trust)?
4. Professional-corporation formation counts — close the sizing gap via provincial professional-order registries.
5. Do the variables discriminate? — test on four to six weeks of real onboarding data before the segmentation drives the roadmap.

Appendix

A. Methodology

Basis. Two complementary lenses. Turnover bands (T1a <$30k, T1b $30k–$100k, T2 $100k–$500k, T3 $500k–$2M; medium enterprise $2M+ out of scope) size and reach the market and populate the persona tables in §4. Within them, needs-based / jobs-to-be-done defines the four behavioural segments (solo service, licensed solo practitioner, micro-employer, platform-native seller); behavioural variables operationalise them; turnover band and firmographics size and reach them; value/economics prioritises them. Segments are drawn before profitability is assessed, then scored on value — the deliberate sequence that prevents pre-sorting customers into "profitable" and "unprofitable" before the behavioural groups exist.

Good-segment test (applied to every candidate, including "modern entrepreneur"): Measurable, Substantial, Accessible, Differentiable, Actionable. Failure on Differentiable or Actionable makes a candidate a descriptor, not a segment.

Prioritisation. GE-McKinsey nine-box at segment level — Attractiveness (CLV 40%, contestable pool 20%, growth 15%, deposit-led fit 25%) against Ability-to-Win (onboarding/platform fit 25%, channel access 25%, capability-gap 15%, contestability 15%, incumbent boundary 20%), scored 1–5, built economics-first. Weights are stated so the map can be re-weighted; the beachhead was tested against growth-first and brand-fit-first weightings and holds.

Nature. Phase 1 public-data behavioural segmentation. Phase 2 — analytics-driven micro-segmentation — is deferred until nine to twelve months of in-account behaviour exists. Every recommendation is a hypothesis to validate in-market.

B. Sizing workings

• Employer base: 1,099,521 (ISED KSBS 2025, Dec 2024); 59.1% at 1–4 employees (≈649,780). Formations 105,001/yr (2018–2022 avg), 91.2% at 1–4 (≈95,725). Discrepancy: StatCan Entrepreneurship Indicators 2023 counts 83,770 births (about −20%), a different definition and a single year — re-verify, do not treat as a confirmed decline.
• Non-employer: 3.67M entities >$30k revenue (StatCan Business Register, Dec 2025); ≈1.9–2.0M self-employed non-employer persons (StatCan LFS, 2023). The two measure entities vs people and are not summed.
• SAM: ≈400,000–550,000 formations/entries per year (≈70,000–96,000 employer 1–4 + ≈300,000–450,000 non-employer, illustrative and churn-inclusive). SOM: ≈4,000–15,000 year-1 accounts (1–3% of SAM, illustrative — no Canadian benchmark).
• CLV inputs: balances anchored on Starling SME ~£12,800/mo and the briefing's C$5–15k band; deposit net interest margin ~2.0%, held just below the BoC policy rate (2.25%) and net of a competitive deposit rate; commercial interchange netted ~1.9% (vs consumer ~0.65%); paid tier $14/mo; per-segment retention 65% / 90% / 80% / 70%. All CLV figures are illustrative arithmetic on cited inputs, not observed the challenger data.

C. Named data gaps to close in week one

Professional-corporation formation counts (provincial professional-order registries); formations by entity type × province (StatCan custom request or ON/QC/BC registry aggregation); current CRA payroll-account and GST/HST-registrant totals; Canadian Stripe/Square merchant counts; Canadian operating-balance, card-spend, and retention by behavioural segment; Canadian digital-engagement-to-value series; Canadian accountant-influence-on-banking-choice data; a Canadian Square Checking equivalent (confirm/deny); channel-specific CAC (all figures are bounds or proxies); the exact StatCan population-projection figure behind "80% vs ~100% of growth is newcomer-driven"; and legal confirmation of the tiered-verification design and small-business proxy-discrimination doctrine.

D. Sources

Consolidated from the twelve research workstreams. Primary public sources include: ISED Key Small Business Statistics 2025 and SME Profile: Ownership Demographics; StatCan (Business Register Dec 2025, Entrepreneurship Indicators 2023, self-employment and platform-work studies, Black-owned and immigrant-entrepreneur studies, Labour Force Survey); Bank of Canada policy and deposit rates (Aug 2026); Visa/Mastercard Canada interchange schedules (May 2026); Starling FY26 results; Business of Apps, Demandsage, storeleads.app (platform merchant counts); Clio Canada and the Law Society of Ontario (trust-accounting rules); North Ledger/Zenbooks and Xero Small Business Insights (admin-burden and payment-timing data); J.D. Power 2025 Canada Small Business Banking Satisfaction Study; ICSB Top 10 MSME Trends 2026; company product, pricing, and eligibility pages for Starling, Tide, Qonto, Mettle, Monzo, Found, Lili, Mercury, Relay, EQ Bank, Venn, Float, Keep, Wealthsimple, Ownr, Wagepoint, Payworks, Shopify, Square; CMA/Scotiabank affinity announcement and Scotiabank Healthcare+; FINTRAC client-identification and beneficial-ownership guidance; OSFI Guideline E-13; OPC/PIPEDA principles; the Consumer-Driven Banking Act (Bill C-15) and Finance Canada regulations; the Canadian Human Rights Act s.5; FCAC consumer-protection guidance; BDC and Business Data Lab reports on immigrant and women entrepreneurs; TD/Harris Poll Gen-Z entrepreneurship survey. Companion briefing: Canadian SME banking — competitive briefing (17 August 2026). Full per-claim citations are held in the twelve workstream files.