ELA: old numbers against the 7 October model

At a glance: year one, old against new

Old is Jon's base case as the 2 October model reproduced it: 300,000 members, AED 500 net each, 50% taking insurance. New is the 7 October Base case. Cautious and Upside are underneath each figure.

Members at the end of year 1300,00010,969Cautious 2,756 · Upside 33,031

Net revenue, year 1AED 210.0MAED6.21MCautious 1.58M · Upside 17.09M

Costs, year 1AED 4.62MAED8.00MCautious 7.58M · Upside 9.07M

Result, year 1+AED 205.38M(AED1.78M)Cautious (6.01M) · Upside +8.02M

Does the AED 7M raise cover it?Yes, "about 18 months"YesBase never dips below AED 5.15M. Cautious runs out in month 15. Upside lowest 6.70M.

Jon's own July text says AED 246M of gross revenue and AED 241.38M of profit; his 9 September summary says 265.5M and 261.88M. The 2 October model's AED 210M is the same 300,000 members after VAT and Apple's 30%. Base costs AED 8.00M include a 15% contingency of AED 1.04M.

What changed, mapped

Every assumption and every cost line from both budgets, so the totals reconcile. Cost lines are year-one AED. Rows in amber are new lines or my own assumptions: those are the ones to test.

SameChangedNewRemoved Jon's July planStevanReviewerClaude's assumption to testPublic source

Line	Old (2 Oct model of Jon's plan)	New, 7 Oct	Status and set by	Why
Members and acquisition
Members at the end of year 1	300,000100k / 300k / 500k	2,75610,96933,031	ChangedClaude's assumption to test	Now built month by month from ad spend, installs, conversion and churn, not a flat count for the whole year.
How members are counted	Everyone pays for all 12 months, no ramp	Monthly: new joiners minus leavers	ChangedReviewer	The reviewer called 300,000 in year 1 unrealistic. A monthly cascade shows where each member comes from.
Cost per install (ad spend per download)	None	AED 25AED 15AED 10	NewClaude's assumption to test	No UAE quote and no market figure yet: the biggest single unknown. See the stress test.
Install-to-paid conversion	None	3%5%8%	NewClaude's assumption to test	Share of downloads that end up paying. Base is 5%, not 3%: 3% is Cautious. See the stress test.
Organic, PR and creator members	None	+30%+60%+100%	NewClaude's assumption to test	Extra members on top of paid ones, as a share of paid members.
Plan mix: annual / monthly / weekly	100% annual	40% / 45% / 15%	ChangedClaude's assumption to test	The old model put everyone on the annual plan. The App Store offers all three.
Weekly payers leaving each month	None	15%12%10%	NewClaude's assumption to test	The old model had no churn at all.
Monthly payers leaving each month	None	8%6%5%	NewClaude's assumption to test	As above.
Annual plans renewed at 12 months	None	50%60%70%	NewClaude's assumption to test	Only matters from month 13.
Company seats	0 in the model75,000 corporate and SME in July text	100 a month from m7400 a month from m51,000 a month from m4	NewClaude's assumption to test	AED 500 a seat a year, no Apple fee, 85% renewal. Year 1 seats: 600 / 3,200 / 9,000. No contract signed yet.
Revenue
Annual price (VAT included)	AED 749.99	AED 749.99	SameApp Store prices	App Store prices since 22 August. Weekly AED 24.99 and monthly AED 79.99 now used too.
VAT	5%	5%	SamePublic source	UAE VAT, taken off before Apple.
Apple's commission, year 1	30%	15%15% (30% from m12)30%	ChangedPublic source	Apple's Small Business Program: 15% while App Store proceeds stay under USD 1M a year. Base passes USD 1M in month 11.
Apple's commission, years 2 and 3	Not modelled	20%20%25%	NewPublic sourceClaude's assumption to test	Blend of 30% for first-year subscribers and 15% after a year. The blend is our estimate.
Membership and seats, net, year 1	AED 150.0M	AED 1.53MAED 5.87MAED 15.47M	ChangedClaude's assumption to test	Follows from the members above. Old: 300,000 x AED 500.
Insurance take-up	50% of members	5%15%25%	ChangedClaude's assumption to test	The old 50% came from the July plan (150,000 insured). The new figures are a guess; see the insurance section.
Insurance premium per adult	AED 3,000	AED 3,000	SameAugust pricing	Set in mid-August. Jon's 9 September summary used AED 2,000.
ELA's share of the premium	10%	10%	SameJon's July plan	To confirm with Howden and AIG.
Insurance live from	Day 1	month 9month 6month 4	NewClaude's assumption to test	Time to sign the policy wording.
Insurance commission, year 1	AED 45.0M	AED 0.01MAED 0.21MAED 1.23M	ChangedClaude's assumption to test	Follows from take-up, start month and members.
Legal referral commission	On: 1% of insured claim x AED 100,000 x 10%	On: 1.5% dispatched x 60% retain x AED 25,000 x 10%	ChangedStevanClaude's assumption to test	Stevan: keep it ON until Jon confirms the law allows it (Decree-Law 34 of 2022, Art. 101). The AED 25,000 and 60% are ours to confirm with Jon.
Legal referral commission, year 1	AED 15.0M	AED 0.03MAED 0.13MAED 0.39M	ChangedClaude's assumption to test	Base year 3: AED 0.62M.
Law-firm panel membership fee	None	AED 5,000 per firm a month, 4 firms, from month 25	NewStevan	Stevan: firms join free for two years, then pay to stay. Placeholder level. AED 0.24M in year 3.
Refunds and chargebacks	None	3% of membership revenue	NewClaude's assumption to test	Counted as a member cost below.
Marketing
Paid adsold: Social media ads	AED 480,000	AED 1,800,000	ChangedReviewerClaude's assumption to test	The reviewer asked for real marketing. AED 150,000 a month in year 1 (200,000 in year 2, 250,000 in year 3) is our level, the same in all cases.
Helios AI agency	AED 480,000	AED 480,000	SameJon's July plan	AED 40,000 a month.
Public relations	AED 480,000	AED 360,000	ChangedClaude's assumption to test	AED 40,000 a month for 6 months, then 20,000.
Creators and partnerships	None	AED 240,000	NewClaude's assumption to test	AED 20,000 a month.
Marketing contingency	AED 100,000	Removed	RemovedJon's July plan	Replaced by one 15% contingency on everything.
Operations
24/7 operations deskold: Outsourced call centre	AED 140,000	AED 180,000AED 210,000AED 315,000	ChangedStevan	Stevan's tiered desk: on-call at AED 15,000 a month under 10,000 members, then staffed seats at AED 30,000 each. Base moves up in month 11.
Paying law firms	Nothing	Nothing	SameStevan	Stevan, 7 Oct: no attendance fee and no retainer. An earlier draft had AED 1,000 per dispatch plus a AED 15,000 monthly retainer; both taken out.
ID check, AED 10 per new member	None	AED 34,080AED 128,000AED 378,000	NewClaude's assumption to test	Company seats included.
Messages, codes and notifications	None	AED 25,942AED 102,517AED 315,592	NewClaude's assumption to test	AED 1.50 per active member a month.
Hosting per member	None	AED 8,647AED 34,172AED 105,197	NewClaude's assumption to test	AED 0.50 per active member a month.
Refunds and chargebacks	None	AED 45,957AED 176,128AED 464,044	NewClaude's assumption to test	3% of membership revenue.
Dispatch messaging and calls	None	AED 649AED 2,563AED 7,890	NewClaude's assumption to test	1.5% of members a year call the desk; AED 30 each.
Technology
Delivery teamold: Outsourced tech support	AED 600,000	AED 1,200,000	ChangedReviewer	The reviewer: "allow double for reality". Covers agency rates and a second engineer.
Servers, cloud, monitoring and licencesold: two lines, 100,000 + 160,000	AED 260,000	AED 300,000	ChangedClaude's assumption to test	AED 25,000 a month.
Equipment	AED 100,000	AED 100,000	SameJon's July plan	Once, in month 1.
Technical contingency	AED 100,000	Removed	RemovedJon's July plan	Replaced by one 15% contingency on everything.
Administration
Personnel (accounts, support, B2B sales)	AED 600,000	AED 600,000	SameJon's July plan	Jon's 31 July plan.
Office	AED 500,000	AED 500,004	SameJon's July plan	Jon's 31 July plan lists Office Accommodation AED 500,000. Now AED 41,667 a month.
D&O and PI insurance	AED 100,000	AED 99,996	SameJon's July plan	Jon's plan. AED 8,333 a month.
Board fees	AED 480,000	AED 480,000	SameJon's July plan	Jon's plan: AED 10,000 a month for each of 4 board members.
Legal, regulatory and accounting	None	AED 140,000	NewStevan	Stevan: AED 7,500 a month plus AED 50,000 once (legal opinions, entity, licences), not the AED 25,000 a month first drafted.
Founders' salaries	0	0	SameJon's July plan	The July plan has none. Kept at 0 as an input.
Administration contingency	AED 200,000	Removed	RemovedJon's July plan	Replaced by one 15% contingency on everything.
Contingency
Contingency on all running costs	AED 400,000three lines above	AED 989,291AED 1,043,007AED 1,182,858	ChangedReviewerClaude's assumption to test	The reviewer asked us to "allow for contingencies". 15% of every cost, ad spend included, is our level.
Funding
Raise	AED 7,000,000	AED 7,000,000	SameJon's July plan	Month 1. Opening cash 0.
Runway	About 18 months7M over 4.62M a year	Runs out in month 15Never runs out; lowest AED 5.15MNever runs out; lowest AED 6.70M	ChangedClaude's assumption to test	Now includes revenue month by month. Cautious needs the staged option in the plan.
Year-one costs, total	AED 4,620,000	AED 7,584,566AED 7,996,387AED 9,068,582	Changed	Every line above, added up. Base: +AED 3,376,387 on the July budget.

Old cost lines are Jon's 31 July budget, which the 2 October model kept line for line. The July plan put the call centre under Technical; it sits under Operations here so the groups match the new model.

Where the year-one costs went

From AED 4.62M to AED 8.00M (Base). Most of the AED 3.38M increase is marketing and contingency, both asked for by the reviewer. Administration barely moved: it's Jon's lines plus Stevan's AED 140,000 for legal and accounting.

Group	July plan	7 Oct, Base	Difference	Mostly because
Marketing	1.44M	2.88M	+1.44M	Paid ads up from AED 0.48M to 1.80M (reviewer: allow for marketing).
Operations and member costs	0.14M	0.65M	+0.51M	Member costs that grow with members (AED 0.44M) and the tiered desk.
Technology	0.96M	1.60M	+0.64M	Delivery team doubled to AED 1.2M (reviewer).
Administration	1.68M	1.82M	+0.14M	Legal, regulatory and accounting added (Stevan). Office, board, personnel unchanged.
Contingency	0.40M	1.04M	+0.64M	One 15% contingency on all costs replaces three fixed lines.
Total	4.62M	8.00M	+3.38M	Cautious AED 7.58M, Upside AED 9.07M: only member costs, the desk and contingency move with members.

How sure are we? The stress test

Ten drivers decide the numbers. The chart shows how far each one moves the Base case on its own, from its Cautious value to its Upside value. The measure is cash at month 36: AED 19.8M in Base.

Cautious valueUpside value

Company seats−8.5M+17.6M

Install-to-paid conversion−10.2M+14.9M

Cost per install−10.2M+12.4M

Organic share−4.7M+6.2M

Weekly and monthly churn−1.8M+1.2M

Insurance take-up and start−1.5M+1.5M

Annual renewal−0.8M+0.8M

Referral fees (off)−1.1M0

Three drivers matter far more than the rest: company seats, conversion and cost per install. Insurance, renewal and churn each move it by under AED 2M. Each run uses sim.py with only that driver changed; referral is shown switched off.

Cost per install (ad spend per download)

AED 25AED 15AED 10

Evidence: No public UAE iOS figure exists; the AppsFlyer and Liftoff benchmark that has one is gated. Nearest: Apple Search Ads averaged USD 3.76 a download in 2025 (about AED 14), but USD 13.28 (about AED 49) for Finance (SplitMetrics). Business of Apps puts iOS in EMEA at "$2.00 - $4.00" (AED 7 to 15), an editorial range. Liftoff saw finance apps at USD 4.35 on iOS in 2021 to 2022 (about AED 16).
What it changes: Base at AED 25: 7,861 members in year 1, year-2 result AED 1.03M (not 4.49M), lowest cash AED 3.39M. At AED 50: 5,531 members, lowest cash AED 0.79M, and the first profitable month slips to 29. Cash never runs out on Base settings.
How we'll test it: The paid test at launch: AED 25,000 at AED 15 buys about 1,700 installs, enough to tell 3% from 8% conversion. Ask the three agencies, Helios included, for a UAE CPI quote by channel.

VerdictBase AED 15 sits inside the public ranges. "Legal" keywords may price like Finance, so Cautious AED 25 is not the worst case. Keep the values; test first.

Install-to-paid conversion

3%5%8%

Evidence: RevenueCat 2026, hard paywalls (pay before using): median 10.7% of installs pay within 35 days; the "floor" is 4.2% and the top quartile above 20%. Freemium apps: 2.1%. Business apps, all paywall types: 2.6%. In the Middle East and Africa, 63.5% of those who pay do so on the day they install.
What it changes: Base at 6%: 12,523 members, year-2 result AED 6.24M (+1.75M). At 8%: 15,630 members and a year-1 profit of AED 0.29M. At 10%: 18,738. At 12%: 21,845, year-2 result AED 16.65M. Cautious at 5%: cash runs out in month 18, not 15.
How we'll test it: TestFlight and soft-launch funnel: installs, each of the 10 sign-up steps, ID check passed, paywall views, purchases. Then the paid test at launch.

VerdictStevan is partly right. Base is 5%, not 3% (3% is Cautious), and the evidence supports about 6%. See the next section.

Organic, PR and creator members (share of paid)

+30%+60%+100%

Evidence: No credible public benchmark found for this ratio in a new category. It depends on PR, word of mouth and partner channels, none of which have launched.
What it changes: Base at +30%: 9,512 members, year-2 result AED 2.85M. At +100%: 12,911 members, AED 6.67M.
How we'll test it: Tag every install by source from day one (attribution links for PR, creators, partners). Compare organic joiners to paid ones each month.

Churn: weekly and monthly payers leaving each month

15% / 8%12% / 6%10% / 5%

Evidence: RevenueCat 2026: after 12 months, a median 8% of monthly subscribers are still paying (9% with a hard paywall; top quartile 11 to 25%); weekly ~1%. Adapty 2025: 17% of monthly plans still active after a year. Our Base 6% a month keeps 48% of monthly payers for a year; 12% a month keeps 22% of weekly payers.
What it changes: Benchmark churn and renewal together (monthly 15%, weekly 30%, annual renewal 40%): Base year 1 has 9,493 members, year-2 result AED 2.04M, and cash at month 36 falls from AED 19.8M to 12.5M.
How we'll test it: Cohort tables from the soft launch: share of each month's joiners still paying at 1, 2, 3 and 6 months, by plan.

VerdictToo optimistic. Monthly churn should be about 15% (Base); weekly about 30%.

Annual plans renewed at 12 months

50%60%70%

Evidence: RevenueCat 2026: annual plans renewing, median 28%; 23% for high-priced apps. First annual renewal by category: Business median 40%, top quartile 53%; Utilities (with Finance) 35%. 35% of annual cancellations happen in month 1.
What it changes: Only bites from month 13. Base at 40%: year-3 result AED 9.07M, not 10.11M. The full benchmark set is in the churn card above.
How we'll test it: Can't be measured before month 13. Proxy: share of annual buyers who turn off auto-renew in month 1 (App Store Connect shows it).

VerdictOur Base 60% is above the best category's top quartile. A protection product may renew better than a typical app, but nothing public shows it. Use 30 / 40 / 55%.

Company seats

100 a month from m7400 a month from m51,000 a month from m4

Evidence: No public benchmark. No contract is signed. The July plan's text expected 75,000 corporate and SME members, but neither model used that figure.
What it changes: The biggest swing in the model. Base with no seats: 7,769 members, year-2 result AED 0.75M, first profitable month 15. At 200 a month: 9,369 members, AED 2.59M. At 1,000: 15,769, AED 10.18M.
How we'll test it: Pipeline conversations: named employers, seats discussed, price and start date. Set Base from signed letters of intent, not hopes.

VerdictLeast evidence for the most weight. Keep 400 a month only if, by launch, the pipeline shows most of Base's 3,200 year-one seats in named conversations.

Insurance take-up

5%15%25%

Evidence: Nobody else sells this cover in the UAE (see Insurance). Abroad, cheap legal add-ons reach about 40% opt-in and 80% pre-ticked (FCA, 2014), but they cost a few per cent of the main product. ELA's costs four times the membership.
What it changes: Each 5 points adds about AED 0.26M to Base year 2 (AED 257,240). At the July plan's 50%: year-2 result AED 6.29M, not 4.49M.
How we'll test it: Howden and AIG data on take-up for comparable add-ons; offer the cover in the soft launch and count quotes and purchases.

Insurance commission (ELA's share of premium)

10% in all cases

Evidence: Jon's plan; to confirm with Howden and AIG. No public UAE figure for affinity commissions found.
What it changes: At 15%: +AED 0.39M in Base year 2. At 20%: +AED 0.77M.
How we'll test it: Put it in the Howden and AIG term sheet.

Legal referral fees and panel fee

On in all cases

Evidence: No benchmark. The legal question (Decree-Law 34 of 2022, Art. 101) is open with Jon. AED 25,000 of fees per retained case and 60% retaining the panel firm are ours.
What it changes: Referral off: Base year-2 result AED 4.10M, not 4.49M. Fees at AED 10,000 a case: AED 4.26M. The panel fee adds AED 0.24M, in year 3 only.
How we'll test it: Jon's legal view; ask two panel firms what a typical retained arrest case bills.

24/7 operations desk

AED 15k a month, then 30k a seat (all cases)

Evidence: Stevan's tiered design. No quote yet for either stage.
What it changes: Small. A staffed seat from month 1 instead of the on-call desk: Base year-1 result AED 1.96M loss instead of 1.78M.
How we'll test it: Two quotes for each stage, one UAE-based, one offshore.

"97% download the app and don't sign up?"

A fair doubt, and partly right. Three things first: Base is 5%, so 95% don't pay; 3% is the Cautious case. The 5% applies only to paid-ad installs; organic members come on top. And for most apps the figure is lower still.

Out of 100 installs, how many pay?

Median hard-paywall app (RevenueCat 2026)10.7

Same, after an ID step that 60 to 85% pass6.4 to 9.1

Our Upside8

Recommended Base6

Our Base today5

Our Cautious3

Business apps, all paywall types2.6

Median freemium app2.1

Row 2 is arithmetic: 10.7% times the 60 to 85% KYC completion an ID-check vendor quotes (Didit, no method given). Every other row is a quoted figure or a model input.

Why ELA sits below the hard-paywall median

Ad installs are low-intent. RevenueCat's median counts every install, organic ones included. People who search for an app convert better than people who tapped an ad.
Ten steps, an ID document and a paywall before any protection starts. Each step loses people. In Signicat's 2022 survey, 38% of Europeans had abandoned a financial application because they lacked the right ID, and 21% over the amount of personal information asked.
The region pays less. RevenueCat puts trial-to-paid at a 19.5% median in the Middle East and Africa, against 34.2% in North America.
But a hard paywall helps. Its floor, 4.2%, is twice the freemium median. And in the region, 63.5% of payers pay on day one: if the sign-up is quick, the paywall catches people while they care.

So: raise Base to 6%, not to 10%

Evidence supports a little more than 5%, not the 10.7% median, because our funnel is longer than most. Base at 6% gives 12,523 members in year 1 and adds AED 1.75M to the year-2 result. At 8%: 15,630 members and a year-1 profit of AED 0.29M. At 10%: 18,738. The soft-launch funnel will settle it within weeks, which is the real answer to "how do we stress test it".

Insurance: unique, but take-up is a guess

Is there another arrest insurance for UAE residents? No.

Neither our 1 October competitor notes nor a fresh web check on 7 October, in English and Arabic, found any insurance, takaful or card benefit that pays for a lawyer or defence after an arrest at home in the UAE.

Sukoon, GIG, AIG Travel Guard, Orient, ADNIC, Liva and Salama pay bail or legal fees on trips only. GIG's wording: "Any Trip solely within the Country of Residence is not insured."
Visa Infinite (Emirates NBD) and Amex Platinum cards: trip cover, mostly for "false arrest", and Amex's is an advance charged back to the card.
Howden UAE's own list of 39 products has no legal expenses line. Its only detention cover is kidnap and ransom, which excludes arrests under local law.
ARAG and D.A.S., the big legal expenses insurers, don't operate in the UAE.
Legal Cover (AED 95 to 399 a month) is an advice subscription, not insurance, and "does not provide court representation".

Limits: eight smaller UAE insurers' wordings weren't read, and a broker could place a one-off policy for a wealthy client without publishing it.

Why take-up still matters less than it looks

ELA earns 10% of a AED 3,000 premium: AED 300 a year per insured member. So each extra 5 points of take-up adds about AED 0.26M (AED 257,240) to Base year 2. One extra point of conversion (5% to 6%) adds AED 1.75M. Even the July plan's 50% take-up only lifts the Base year-2 result from AED 4.49M to 6.29M.

Is 5% too low?

5% is Cautious; Base is 15%. Being unique raises demand, but the price pulls the other way. The add-on costs AED 3,000, four times the AED 749.99 membership, and is bought after joining. The high take-up figures abroad come from cheap add-ons sold at checkout: UK motor legal expenses cover reached 80% when pre-ticked and about 40% when customers opted in (FCA, 2014). LegalShield's USD 1 arrest rider reached about 72% of members. No public figure exists for an add-on that costs more than the product it sits on. So 15% Base is not cautious, and I wouldn't raise it.

What would raise it: the price, not the pitch

A bundled tier. Membership with a capped defence benefit at a few hundred dirhams more, with the full AED 3,000 cover as the upgrade. Note that cutting the premium alone doesn't pay: at AED 1,500, take-up must reach 40% just to add the same AED 0.26M as going from 15% to 20% at AED 3,000.
Company-paid cover. Put the insurance into company-seat contracts, where the employer pays and every seat is insured.
Offer it at sign-up, as one choice on the paywall, not as a later upsell.
Ask Howden and AIG for take-up data on comparable add-ons, and price a capped tier.

What I'd change in the model now

A recommendation only: no file has changed. The evidence pushes conversion up a little and retention down a lot. Together they leave year 1 about where it is and take a quarter off year 3.

Assumption	Now: C / B / U	Recommended: C / B / U	Why
Install-to-paid conversion	3 / 5 / 8%	3 / 6 / 10%	Hard-paywall median 10.7%, less our longer funnel and ID step.
Monthly payers leaving each month	8 / 6 / 5%	20 / 15 / 10%	Median 8 to 9% still paying after a year; top quartile up to 25%. 15% a month leaves 14%.
Weekly payers leaving each month	15 / 12 / 10%	35 / 30 / 25%	About 1% of weekly payers still pay after a year.
Annual renewal at 12 months	50 / 60 / 70%	30 / 40 / 55%	Business median 40%, top quartile 53%. All apps 28%.
Cost per install	AED 25 / 15 / 10	No change	Inside the public ranges. Add AED 50 to the sensitivity grid: Apple Search Ads Finance averages about AED 49.
Insurance take-up	5 / 15 / 25%	No change	Don't raise it. Work on the price and a bundled tier instead.
Company seats	100 / 400 / 1,000 a month	No change yet	Biggest lever, no evidence. Gate Base on named pipeline before launch.

What the four changes do together

Case	Members, end of year 1	Result with the changes, years 1 / 2 / 3 (AED M)	Cash at month 36	Cash runs out?
Cautious	2,756 → 2,313	(6.18) / (5.45) / (4.95)	(7.64)M → (9.57)M	Month 14, a month earlier
Base	10,969 → 10,752	(1.68) / 3.30 / 7.39	19.82M → 16.01M	No; lowest AED 5.32M
Upside	33,031 → 34,830	9.39 / 28.98 / 44.43	92.48M → 89.81M	No

Today's Base result is (1.78) / 4.49 / 10.11. Recommended Base with only 200 seats a month: 9,152 members and a lowest cash of AED 4.56M. Also worth watching: buyers in the region favour monthly plans (52 to 55% of plans sold, RevenueCat), against our 45%.

The tests, in order

TestFlight and soft-launch funnelInstalls, each sign-up step, ID passed, paywall views, purchases, by source. Settles conversion and shows which step to fix.
Paid test at launch, AED 20,000 to 30,000Real CPI and paid conversion in the UAE. At AED 15 a download, AED 25,000 buys about 1,700 installs.
Agency proposals, three agencies including HeliosCPI quotes by channel, before the AED 150,000 a month starts.
Howden and AIGTake-up data on comparable add-ons, the commission, and a capped tier's price.
Company-seat pipelineNamed employers and seat counts. This decides more than any other input.
Cohort retention from month 2Monthly and weekly churn by cohort; the share of annual buyers who switch off auto-renew in month 1.

Sources

Our own files

Old: Jon's business plan v3, 31 July 2026 (budget and revenue, extracted/plan-v3-2026-07-31-with-shaz.md); his 9 September summary (extracted/teaser.md); the 2 October model (draft/ELA-financial-model-2026-10-02.xlsx).
New: the 7 October model (draft/ELA-financial-model-2026-10-07.xlsx), its report (rebuild-2026-10-07/REPORT.md) and brief (rebuild-2026-10-07/prompt.txt, with prompt-v1.txt for what Stevan changed).
What-ifs: draft/build/model-2026-10-07/sim.py, unchanged, run from diff-2026-10-07/work/ (whatif.py, tornado.py, revised.py).
Competitors: research/competitors.md and research/scratch/competitors-notes-*.md, 1 October; fresh check in diff-2026-10-07/research-insurance.md. Benchmarks with exact quotes: diff-2026-10-07/research-benchmarks.md.

Benchmarks

RevenueCat, State of Subscription Apps 2026 115,000+ apps, 2025 data. Conversion, retention, regional figures.
RevenueCat, renewal rates by app category (2026 benchmarks) First annual and monthly renewal, by category.
SplitMetrics, Apple Ads cost benchmarks 2025 data. Cost per download, by category.
Business of Apps, Cost Per Install rates (2025) Editorial ranges; no dataset given.
Liftoff, 2022 Mobile Finance Apps Report Via press coverage; June 2021 to June 2022 data.
Adapty, State of in-app subscriptions 2025 Monthly plan retention.
Signicat, The Battle to Onboard 2022 Survey of 7,600 European consumers.
Didit, identity verification conversion rates Vendor blog; low confidence.

Insurance

FCA, General insurance add-ons, MS14/1 (March 2014) Opt-out 80% against opt-in about 40%.
Pre-Paid Legal Services, 10-K for 2010 LegalShield arrest rider; attach rate is our arithmetic.
GDV, Insurance Flash 2026/01 49.8% of German households hold legal expenses cover (2023).
GIG Gulf travel terms Trips within the country of residence not insured.
Sukoon Travel Easy terms Bail and legal cover on trips only.
Howden UAE, full product range No legal expenses line.
Howden UAE, kidnap and ransom The only detention-related cover.
Emirates NBD Visa Infinite travel terms False arrest, card-paid trips only.
Legal Cover terms Advice subscription; no court representation.
ARAG locations None in the GCC.