ShortPay Sifter
ShortPay Sifter ingests a small practice's ERA/835 files or claims export and returns a per-payer list of claims that paid below that practice's own historical allowed rate, packaged as a ready-to-send underpayment dispute packet.
One product, one founder, thirty days. Everything below is built to be executed inside the contest window — the official rules still decide what counts.
Decide in 20 seconds
The facts that decide whether this is worth thirty days of yours.
Owner or lead biller at a US independent medical billing company with 1-10 staff managing 5-30 small provider clients (physical therapy, chiropractic, specialty groups), triggered when a client asks why A/R is flat or when a payer contract renews.
Market demand 4.4/10
Narrow the buyer to a segment with an urgent, already-budgeted problem and reference the specific trigger event that creates willingness to pay.
Why this could work in 30 days: one founder can ship the MVP in 5 days; the plan targets a first sale on day 11; 11 sales at $99 clears $1,000 of collected revenue; strongest signal is speed to first revenue at 8.2/10.
Score and projections come from the generated report and are not guarantees. The $100K Hackathon is run by Ship or Die — read the official rules before you build.
Your first 24 hours
12 concrete actions. Tick them off as you go — nothing is saved or sent.
Product opportunity and positioning
Target customer
Owner or lead biller at a US independent medical billing company with 1-10 staff managing 5-30 small provider clients (physical therapy, chiropractic, specialty groups), triggered when a client asks why A/R is flat or when a payer contract renews.
Why now
The data already exists in every billing shop's clearinghouse export, LLM-free deterministic benchmarking is cheap and auditable, and a one-time $249 audit clears a small-company approval threshold that a platform subscription does not.
Problem
Payers short-pay, downcode and bundle claims, and small billing shops have no systematic way to find those claims because the dollars are already booked and closed; the loss only shows up as a slow drip in collections.
Solution
The biller uploads 1-12 months of 835/ERA files or a claims CSV; ShortPay Sifter builds a benchmark per payer/CPT/modifier/place-of-service from that practice's own historical allowed amounts, flags lines paid more than 5% or $25 under benchmark while excluding true denials, and generates a per-payer recovery packet (variance PDF/CSV, pre-filled reprocessing cover form, Won/Rejected/Pending tracker).
1. Opportunity and positioning
Payers underpay. A claim is adjudicated, paid and closed, and the gap between the practice's established allowed amount and what actually landed never becomes a task — it becomes a slower collections rate the billing company cannot explain to its client. The people who feel this are not hospital CFOs. They are owner-operators of independent medical billing companies with one to ten staff, each managing five to thirty small provider clients. Their trigger is concrete: a client asks why A/R is flat, or a payer contract renews, and the biller has no defensible answer because they have never had a systematic way to compare paid amounts to their own history.
ShortPay Sifter is deliberately not a revenue cycle platform. It touches no EHR, no clearinghouse, no payer portal, no clinical note. The customer already owns the data — ERA/835 files and claims exports — and already believes underpayment is real. The product takes that file, benchmarks each payer/CPT/modifier/place-of-service combination against the practice's own historical paid lines, flags current lines that landed materially below that benchmark, and returns a per-payer recovery packet: variance table, pre-filled reprocessing cover form, and a Won/Rejected/Pending tracker.
Positioning against the alternatives is the whole sale. A contract-modeling tool needs a fee schedule the customer probably has not digitized. A denials tool solves a different claim — one that paid zero. A spreadsheet pivot is what the biller does today, badly, on a Friday afternoon. ShortPay Sifter needs no contract data, no integration and no coding judgment, and produces a dollar figure in a single file. That is why a $249 one-time audit gets approved where a monthly platform does not.
The JEV score of 60/100 carries one loud signal: speed to first revenue scored 8.22. The buyer is reachable this week, the artifact is the pitch, and one person can deliver the first three audits by hand before the software is finished. The weak signals — market demand at 4.42 and operational simplicity at 4.86 — are addressed by refusing to become a platform: no PHI storage, no BAA dependency in v1, no human review step in the paid path, no 24/7 support promise. The whole business is files in, packets out, billed once.
2. Smallest sellable MVP
The smallest sellable thing is an upload page and a PDF. A biller drops in 835 files or a CSV, maps columns once, and receives a per-payer variance table plus a cover form they can print, sign and send. Everything else — tracker, monthly scans, benchmark overrides — is upside that arrives after money has changed hands.
The MVP is sellable on day five because the founder runs the first three scans manually behind the same upload form. Concierge delivery is not a shortcut; it is how the benchmark logic and the reprocessing language get validated against real files before a single line of automation depends on them.
3. Technical implementation plan
Next.js on Vercel hosts a single admin and a customer upload portal. Supabase Postgres stores claim_line and packet rows under row-level security; Supabase Storage holds uploads with a 30-day auto-delete job. A serverless Node function parses 835 CLP/SVC/CAS segments; papaparse handles CSV intake with a drag-to-map column mapper. Benchmarking is deterministic SQL and JavaScript — median and 80th-percentile allowed amount per payer/CPT/modifier/POS, minimum five historical lines, manual override, percent-of-billed fallback when the sample is thin. No model sits in the core path, which keeps the variance table auditable and reproducible. Packets render with react-pdf; Resend delivers notifications; Stripe Checkout carries the $99, $249 and $99/month tiers.
The hardest technical risk is the parser, not the math. Every clearinghouse emits 835s with its own quirks — missing CAS groupings, bundled adjustment reason codes, payer IDs that alias the same payer. The mitigation is a normalizer layer with a per-format fixture test suite, grown one real customer file at a time.
4. Pricing and monetization
Three SKUs, one job. The $99 Single-Payer Scan is the entry test on one problem payer. The $249 Full Recovery Audit is the real product: all payers, twelve months, per-payer packets, 30-day tracker. The $99 monthly Short-Pay Watch is retention, not acquisition, and is excluded from the base projection.
The trust lever is the guarantee: if a scan finds under $500 in flags, the audit is free. That converts a skeptical biller at zero risk and costs the founder nothing but time. Expected first-month ARPU blends to roughly $199 because most buyers enter at $249 and a minority test at $99.
5. First-customer acquisition
One channel: direct LinkedIn outreach to owner-operators of independent medical billing companies, opened with a free single-payer scan on one real client file. No ads, no SEO, no partnerships, no purchased lists. The founder builds a list of 80 named owners, sends in waves of 25, and runs the first three scans by hand within 48 hours. Conversion happens on the delivered packet — the artifact sells itself — with a Stripe link attached to the delivery email. Day-3 and day-7 follow-ups carry the flagged dollar total in the subject line. Every closed buyer is asked for one referral name, which is the only compounding loop in the plan.
6. 30-day execution roadmap
Days 1-5 are build: schema, 835 parser, benchmark engine, packet renderer, then live Stripe and PHI stripping. Day 5 is the MVP deadline and the product must be able to produce a packet from a real file.
Days 6-11 are the first revenue push: build the 80-name list, send wave one, deliver three manual scans, post a redacted synthetic sample in two biller communities, and convert with the Stripe link. Day 11 is the first-sale target. Days 12-21 harden the parser against real clearinghouse files, ship the Won/Rejected tracker and the benchmark override, and run outreach waves two through four. Days 22-30 push volume: trade-group outreach, the monthly tier pitched to two past audit buyers, self-serve signup so no call is required, and a final accounting of collected revenue, refunds and Stripe fees.
7. Revenue and profit scenarios
All figures below are projections from stated assumptions, not verified revenue or existing customers.
Conservative: 3 customers at $99 average, $297 gross, $10 in processing fees, $25 in hosting and API cost.
Base: 12 customers blending eight audits at $249 and four scans at $99, $2,388 gross, $76 processing, one $99 refund, $70 operating cost.
Upside: 24 customers at $212 average, $5,088 gross, $163 processing, $249 refunded, $200 paid marketing, $130 operating cost.
The contest ranks collected revenue minus refunds, chargebacks, processing fees and paid marketing. Hosting and API costs are excluded from that figure and appear only in business profit. Every scenario must clear cost-to-serve per sale; nothing is priced below delivery cost.
8. Risks and validation experiments
The dominant risk is circular benchmarking: if a payer has always underpaid, the practice's own historical allowed amount is depressed and the variance goes undetected. The test is cheap — run three free scans and check whether flagged dollars exceed $500 per client. If two of three scans return nothing, the benchmark definition changes before any paid audit ships.
Second risk: false positives from coordination-of-benefits, modifier and sequestration adjustments. Mitigation is a suppression rule set built from the first three real files, validated by asking each biller to mark which flags they would actually dispute. Kill criterion: more than 30% of a packet's flags rejected by the biller as not real.
Third risk: billers refuse to upload anything without a BAA. Mitigation is the claim-ID-only file spec published on the upload page and a manual, file-free path for the first scans. If the first ten conversations show BAA is a hard blocker, the free scan becomes fully concierge and the upload portal is deprioritized.
Fourth risk: payers reject reprocessing requests without a contract citation, so recovered dollars trail flagged dollars. The tracker measures this directly; if recovery rate sits under 15% across the first two paid audits, the packet gains a contract-reference field and the pitch shifts from recovered dollars to documented variance.
Fifth risk: the buyer decides an Excel pivot is good enough. The counter is delivery speed — a packet in 48 hours versus an afternoon of pivot work, priced at $249 against a flagged total in the thousands.
9. First 24-hour action plan
Hours 1-2: create the repo, Supabase project and Stripe test account; write the claim_line schema. Hours 3-5: build the 835 parser against a synthetic fixture and confirm CLP/SVC/CAS output. Hours 6-8: wire the benchmark and variance query and run it on the fixture. Hours 9-11: build the upload page and skeleton PDF cover form. Hours 12-14: assemble the first 40 billing-company owner names from LinkedIn and draft the outreach message. Hours 15-18: send the first ten messages with the free single-payer scan offer, then commit the packet template. The day ends with a deployed upload form, a working variance query and ten live conversations.
Smallest sellable MVP
An upload page that accepts 835/ERA files or a mapped CSV and returns a per-payer PDF/CSV variance table plus a pre-filled reprocessing cover form, with the first three audits delivered by the founder by hand.
Must-have features
- Upload portal accepting 835/EDI files, ERA PDFs and a CSV with drag-to-map column mapping
- 835 parser for CLP/SVC/CAS segments plus CSV normalizer into a unified claim_line table
- Self-benchmark engine: median and 80th-percentile allowed amount per payer/CPT/modifier/POS, minimum 5 historical lines
- Variance flagger ranking recoverable dollars and excluding zero-pay and denial-coded lines
- Per-payer recovery packet: variance table in PDF and CSV plus a pre-filled reprocessing cover form with submission checklist
- Stripe Checkout for the $99 scan, $249 audit and $99 monthly tiers
- Automatic PHI stripping (patient name/DOB dropped at intake) and 30-day file auto-delete
Explicitly out of scope
- No EHR, PM system, clearinghouse or payer portal integration
- No automated faxing or electronic payer submission
- No denial appeals and no payer policy RAG corpus
- No clinical coding advice or CPT/ICD recommendation engine
- No multi-role permissions, SSO or enterprise admin console
- No BAA-gated PHI storage in v1
- No multi-client dashboard in the first five days
Demo flow (what the buyer sees)
- Biller uploads three months of 835 files for one client
- Column mapper confirms payer, CPT, modifier, POS, billed, allowed, paid and denial code fields
- PHI fields are dropped and the file is stored with a 30-day delete timer
- Benchmark engine computes allowed-amount distributions per payer/CPT/modifier/POS
- Variance flagger ranks the current lines that paid more than 5% or $25 under benchmark
- PDF packet and CSV render per payer with the cover form pre-filled
- Packet is emailed to the biller with a Stripe link for the full audit
- Biller marks packet lines Won, Rejected or Pending in the tracker
Technical implementation plan
Stack
Architecture
A Next.js app on Vercel serves the upload portal and an internal admin view. Uploaded files land in Supabase Storage; a serverless function parses 835 segments or a normalized CSV into a claim_line table in Postgres. Benchmark and variance logic run as deterministic SQL views over claim_line, scoped by an RLS policy on the uploading account. The packet renderer reads the variance query and produces a PDF cover form plus a CSV line detail, stores both in Storage, and Resend emails the links. Stripe Checkout issues the entitlement that unlocks full-payer output; files auto-delete after 30 days.
Build steps
- Create Next.js app, Supabase project, Stripe test account and schema for claim_line, benchmark, packet
- Write synthetic 835 fixtures covering CLP/SVC/CAS including multiple CAS adjustment groups
- Implement the 835 parser and CSV normalizer into a single claim_line shape
- Build the benchmark SQL view with minimum-sample and manual-override rules
- Build the variance flagger with denial-code exclusions and dollar ranking
- Build the upload UI and column mapper with PHI field stripping
- Render the per-payer PDF cover form and CSV detail with react-pdf
- Wire Stripe Checkout, entitlement gating and Resend delivery
- Add the Won/Rejected/Pending tracker and the quarterly recovered-dollar summary
- Ship the 30-day storage auto-delete and publish the claim-ID-only file spec
AI components
- No LLM in the core benchmark or variance path by design
- Optional LLM assist only for mapping unfamiliar CSV column headers to the canonical schema
- Nothing clinical or coding-related touches a model
Pricing and monetization
Single-Payer Scan
Entry offerFor Biller testing the tool on one problem payer
- One payer, up to 12 months of data
- Variance table CSV
- One-page PDF dispute cover form
Full Recovery Audit
For Billing company owner auditing one client across all payers
- All payers, up to 12 months of data
- Prioritized variance table (PDF + CSV)
- Per-payer pre-filled dispute packets
- 30-day Won/Rejected tracker
Monthly Short-Pay Watch
For Billing companies running ongoing monthly scans per client
- Monthly scan on new 835s
- New flags only
- Recovered-dollar rollup per client
Model: One-time per-audit pricing with an optional low monthly scan tier for retention.
First revenue offer: Free single-payer Short-Pay Scan on one real client file, delivered in 48 hours with the exact dollar amount short-paid; $249 Full Recovery Audit only if the scan flags more than $500.
Expected ACV: $199
Upsell path: Convert the $249 audit buyer to the $99 monthly Short-Pay Watch per client after the first packet, then expand to their second and third client files.
First-customer acquisition
Direct LinkedIn outreach to owner-operators of independent medical billing companies, opened with a free single-payer scan on one real client file.
Where to find them
- LinkedIn search: 'medical billing' owner or founder, United States, 1-10 employees
- LinkedIn search: 'billing manager' at independent billing companies serving physical therapy and chiropractic
- Facebook and LinkedIn groups for independent medical billing company owners
- Physical therapy and chiropractic practice-owner communities where their billers are referenced
- Referral names requested from every biller who receives a delivered packet
Exact outreach motions
- Send wave one of 25 LinkedIn connection notes with the free single-payer scan offer and the claim-ID-only file spec
- Run the first three scans by hand within 48 hours and deliver the variance table with the flagged dollar total in the subject line
- Day-3 follow-up: packet summary plus one line explaining what the largest flag means
- Day-7 follow-up: the $249 audit offer with the under-$500 refundable guarantee and a Stripe link
- Post one redacted synthetic-data sample packet in two biller communities and answer questions in the thread
- Ask every delivered-scan recipient for one other biller who would want the same scan
Message template
Hi [Name] — I run ShortPay Sifter. Send me 3 months of one client's 835/ERA files (or a claims CSV, claim IDs only, no patient names) and I'll send back the exact dollar amount that payer short-paid versus your own historical allowed rates, free, in 48 hours. No integration, no BAA needed for the scan. Worth a look?
30-day execution roadmap
Week 1 · Build & validate · 7 days
Infrastructure
- Create Next.js repo, Supabase project and Stripe test account
- Define claim_line, benchmark and packet tables with RLS
Deliverable: Deployed skeleton app with schema migrated
835 parser
- Write synthetic CLP/SVC/CAS fixtures
- Implement serverless 835 parser and CSV normalizer
Deliverable: Parser emits claim_line rows from test files
Benchmark and variance
- Build median/80th-percentile benchmark SQL view
- Add denial-code exclusion and dollar-ranked variance flagger
Deliverable: Variance table produced from fixture data
Packet rendering
- Build upload UI with column mapper
- Render per-payer PDF cover form and CSV detail
Deliverable: End-to-end packet from a single upload
Monetization and safety
- Wire Stripe Checkout and entitlement gating
- Add PHI stripping, Resend delivery, 30-day auto-delete
Deliverable: MVP live with paid checkout
Build the list
- Assemble 80 billing-company owners from LinkedIn
- Draft and test the outreach message
Deliverable: 80-row prospect sheet and final message copy
Wave one
- Send 25 outreach messages
- Publish the claim-ID-only file spec page
Deliverable: 25 messages sent, spec live
Week 2 · Launch & first sale · 7 days
First real scan
- Run scan one by hand on a real client file
- Deliver variance table and capture the biller's reprocessing wording
Deliverable: First packet delivered, wording bank started
Scans two and three
- Run two more manual scans
- Log every false positive for the suppression rules
Deliverable: Three delivered scans and a false-positive list
Public artifact
- Post redacted synthetic sample packet in two biller communities
- Send day-3 follow-ups to all scan recipients
Deliverable: Two community posts live, follow-ups sent
First sale
- Offer $249 audit with under-$500 guarantee to scan recipients
- Send Stripe links and confirm payment
Deliverable: First paid Full Recovery Audit
Parser fixes
- Add support for two new clearinghouse layouts
- Add job-level parse error surfacing for the uploader
Deliverable: Parser passes new format fixtures
Wave two
- Send 25 outreach messages
- Ask every delivered recipient for one referral name
Deliverable: 50 cumulative messages, referral names collected
Week-two metrics
- Review reply rate, scan acceptance and converted dollars
- Tighten guarantee copy based on objections
Deliverable: Metrics snapshot and revised offer copy
Week 3 · Repeat & systematize · 7 days
Recovery tracker
- Ship Won/Rejected/Pending status per packet line
- Add recovered-dollar rollup per client
Deliverable: Tracker live in the portal
Wave three
- Send 25 outreach messages
- Follow up on all day-7 pending threads
Deliverable: 75 cumulative conversations started
Paid audits
- Deliver two paid audits
- Record which flags the buyers actually disputed
Deliverable: Two paid packets delivered
Benchmark control
- Ship manual benchmark override UI
- Ship percent-of-billed fallback for thin samples
Deliverable: Override and fallback live
Wave four
- Send 25 outreach messages
- Post one community update with a synthetic before/after
Deliverable: 100 cumulative messages sent
Testimonials
- Ask three buyers for a one-line quote and a referral
- Add verified quotes to the landing page
Deliverable: At least two quotes published
Client-facing summary
- Ship the one-page recovered-this-quarter summary
- Add per-payer packet email delivery
Deliverable: Client summary PDF shipped
Week 4 · Scale what works · 7 days
Trade groups
- Target billing-company trade groups and associations
- Send 25 messages to the new segment
Deliverable: 25 new qualified contacts
False positives
- Tune suppression rules for COB, modifier and sequestration cases
- Re-run two delivered packets under the new rules
Deliverable: Suppression rule set v2
Monthly tier
- Pitch the $99 monthly Short-Pay Watch to two audit buyers
- Send Stripe subscription links
Deliverable: At least one monthly subscription started
Case format
- Publish a case-format writeup using synthetic numbers
- Link it from the outreach message
Deliverable: One public case-format post
Wave six
- Send 25 messages with the case link
- Run day-7 follow-ups on all open threads
Deliverable: 130 cumulative messages sent
Self-serve
- Ship self-serve scan signup with no call required
- Add a Stripe link on the post-scan result page
Deliverable: Self-serve signup live
Upload reliability
- Harden upload validation and error messages
- Reduce failed-upload support requests
Deliverable: Error handling shipped
Days 29-30 · Consolidate · 2 days
Open deals
- Final outreach push and close every open proposal
- Deliver any remaining paid audits
Deliverable: All open conversations answered
Accounting and wrap
- Reconcile collected revenue, refunds and Stripe fees
- Publish a wrap post with flagged and recovered dollars per client
Deliverable: Final accounting and public wrap
Revenue and profit scenarios
How the numbers are defined
Hackathon profit — Collected revenue − refunds − chargebacks − payment processing fees − paid marketing spend.
Business profit — Hackathon profit − operating costs (hosting, domains, tooling, AI/API delivery costs).
The contest ranks entrants on hackathon profit only. Hosting, tooling and AI/API delivery costs are not deducted by the organisers, but each sale still has to cover them — so we show both. Definitions verified on 2026-11-15 against the official rules.
Stated assumptions
- All scenarios are projections from stated assumptions, not verified revenue, customers or traction.
- Conservative assumes the free-scan offer converts slowly and every buyer enters at the $99 Single-Payer Scan.
- Base assumes 12 buyers blending eight $249 Full Recovery Audits and four $99 scans, giving an average of $199 per customer.
- Upside assumes broader community reach plus $200 of paid sponsored posts and 24 buyers averaging $212.
- Payment processing is modeled at approximately 3.2% of gross (Stripe card pricing) across all scenarios.
- Paid marketing is $0 in conservative and base because acquisition is organic LinkedIn outreach.
- Operating costs cover Vercel, Supabase, storage, PDF rendering and email delivery; no LLM inference cost sits in the core path.
- The $99 monthly Short-Pay Watch is excluded from base; any subscriptions are upside.
- Every sale is priced above cost-to-serve, so no scenario relies on selling below delivery cost.
Conservative
3 customers- Refunds
- $0
- Chargebacks
- $0
- Payment fees
- $10
- Paid marketing
- $0
- Hackathon profit
- $287
- Operating costs
- $25
- Business profit
- $262
Base case
12 customers- Refunds
- $99
- Chargebacks
- $0
- Payment fees
- $76
- Paid marketing
- $0
- Hackathon profit
- $2,213
- Operating costs
- $70
- Business profit
- $2,143
Optimistic
24 customers- Refunds
- $249
- Chargebacks
- $0
- Payment fees
- $163
- Paid marketing
- $200
- Hackathon profit
- $4,476
- Operating costs
- $130
- Business profit
- $4,346
Model your own numbers
Runs in your browser · nothing is saved- Collected revenue
- $1,188
- Payment fees (3.9%)
- $46
- Paid marketing
- $0
- Hackathon profit
- $1,142
- Business profit
- $1,072
Hypothetical arithmetic, not a forecast. The fee rate is a flat 3.9% assumption, refunds and chargebacks are left at zero here, and none of this is tracked data.
Scenario models from stated assumptions, not verified revenue. Hackathon profit excludes hosting and API costs; business profit deducts them.
Main risks and validation experiments
Risk 1Circular benchmarking: if a payer has always underpaid, the practice's own historical allowed amount is depressed and the variance never surfaces.
Publish the benchmark methodology, add a percent-of-billed fallback for thin samples, and let the biller override any benchmark.
Run three free scans this week and record flagged dollars per client; two of three must exceed $500.
If two of three scans return under $200 in flags, redefine the benchmark or abandon the wedge.
Risk 2False positives from coordination-of-benefits, modifier and sequestration adjustments destroy trust in the packet.
Suppression rules built from real files, plus a biller-marked 'would dispute' feedback loop on every delivered packet.
Ask each of the first three billers to mark which flags they would actually send, and compute the reject rate.
Abandon the current flag logic if more than 30% of a packet's flags are rejected as not real.
Risk 3Billers refuse to upload files without a BAA, stalling the 30-day clock.
Publish the claim-ID-only file spec, strip patient identifiers at intake, auto-delete after 30 days, and offer a fully manual scan path with no upload.
Track BAA objections across the first ten outreach conversations and count how many convert on the no-upload path.
If all ten conversations block on BAA, drop the portal from the paid path and sell concierge audits only.
Risk 4Payers reject reprocessing requests that lack a contract citation, so recovered dollars trail flagged dollars badly.
Add a contract-reference field to the cover form and set expectations with a flagged-versus-recovered split in the tracker.
Measure recovery rate on the first two paid audits over their 30-day tracker window.
If recovery sits under 15% across two audits, reposition from recovered dollars to documented variance.
Risk 5A sophisticated biller decides an Excel pivot replicates the logic and never pays.
Compete on turnaround, not features: packet in 48 hours, pre-filled cover form, and a clean per-payer artifact that a pivot does not produce.
Offer the free scan to five billers who explicitly mention doing this manually today.
If fewer than one of five converts to a paid audit, the price or the artifact is wrong.
Hour-by-hour: the first 24 hours
Hours 1-2
- Create the Next.js repo, Supabase project and Stripe test account
- Write the claim_line, benchmark and packet schema with RLS
Output: Deployed skeleton app with migrated schema
Hours 3-5
- Build synthetic 835 fixtures with CLP/SVC/CAS segments
- Implement the serverless 835 parser and confirm output rows
Output: Parser producing claim_line rows from fixtures
Hours 6-8
- Build the benchmark SQL view with a 5-line minimum
- Add the dollar-ranked variance flagger with denial exclusions
Output: Working variance query on fixture data
Hours 9-11
- Build the upload page with the drag-to-map column mapper
- Render a skeleton PDF cover form and CSV detail
Output: End-to-end upload to packet on test data
Hours 12-14
- Assemble the first 40 billing-company owner names from LinkedIn
- Finalize the free single-payer scan message
Output: 40-row prospect sheet and final outreach copy
Hours 15-18
- Send the first ten outreach messages
- Publish the claim-ID-only file spec page
Output: Ten live conversations and a public data spec
Full 30-day launch report
Read the full generated report
ShortPay Sifter — the 30-day solo launch
1. Product opportunity and positioning
Payers underpay. A claim is adjudicated, paid and closed, and the gap between the practice's established allowed amount and what actually landed never becomes a task — it becomes a slower collections rate the billing company cannot explain to its client. The people who feel this are not hospital CFOs. They are owner-operators of independent medical billing companies with one to ten staff, each managing five to thirty small provider clients. Their trigger is concrete: a client asks why A/R is flat, or a payer contract renews, and the biller has no defensible answer because they have never had a systematic way to compare paid amounts against their own history.
ShortPay Sifter is not a revenue cycle platform. It touches no EHR, no clearinghouse, no payer portal, no clinical note. The customer already owns the data — ERA/835 files and claims exports — and already believes underpayment is real. The product benchmarks each payer/CPT/modifier/place-of-service combination against the practice's own historical paid lines, flags current lines that landed materially below benchmark, and returns a per-payer recovery packet: variance table, pre-filled reprocessing cover form, Won/Rejected/Pending tracker.
Positioning against alternatives is the whole sale. A contract-modeling tool needs a fee schedule the customer probably has not digitized. A denials tool solves a different claim — one that paid zero. A spreadsheet pivot is what the biller does today, badly, on a Friday afternoon. ShortPay Sifter needs no contract data, no integration and no coding judgment, and produces a dollar figure in a single file. That is why a $249 one-time audit clears approval where a monthly platform does not.
The JEV score of 60/100 carries one loud signal: speed to first revenue at 8.22. The buyer is reachable this week, the artifact is the pitch, and one person can deliver the first three audits by hand before the software is finished. The weak signals — market demand at 4.42 and operational simplicity at 4.86 — are handled by refusing to become a platform: no PHI storage, no BAA dependency in v1, no human review in the paid path, no 24/7 support promise.
2. Smallest sellable MVP
The smallest sellable thing is an upload page and a PDF. A biller drops in 835 files or a mapped CSV and receives a per-payer variance table plus a cover form they can print, sign and send. The tracker, monthly scans and benchmark overrides arrive only after money has changed hands.
The MVP is sellable on day five because the founder runs the first three scans manually behind that same upload form. Concierge delivery is not a shortcut; it is how the benchmark logic and reprocessing language get validated against real files before automation depends on them.
3. Technical implementation plan
Next.js on Vercel hosts the portal and an internal admin. Supabase Postgres stores claim_line and packet rows under row-level security; Supabase Storage holds uploads with a 30-day auto-delete job. A serverless Node function parses 835 CLP/SVC/CAS segments; papaparse and a drag-to-map column mapper handle CSV intake. Benchmarking is deterministic SQL and JavaScript — median and 80th-percentile allowed amount per payer/CPT/modifier/POS, minimum five historical lines, manual override, percent-of-billed fallback for thin samples. No model sits in the core path, which keeps the variance table reproducible. Packets render with react-pdf, Resend delivers notification emails, Stripe Checkout carries the three SKUs.
The hardest technical risk is the parser, not the math. Every clearinghouse emits 835s with its own quirks — bundled adjustment reason codes, aliased payer IDs, inconsistent allowed-amount placement. Mitigation is a normalizer layer with per-format fixture tests, grown one real customer file at a time.
4. Pricing and monetization
Three SKUs, one job. The $99 Single-Payer Scan is the entry test on one problem payer. The $249 Full Recovery Audit is the real product: all payers, twelve months, per-payer packets, 30-day tracker. The $99 monthly Short-Pay Watch is retention and is excluded from the base projection.
The trust lever is the guarantee: if a scan finds under $500 in flags, the audit is free. That converts a skeptical biller at zero risk and costs the founder only time. Expected first-month average revenue per customer blends to roughly $199.
5. First-customer acquisition
One channel: direct LinkedIn outreach to owner-operators of independent billing companies, opened with a free single-payer scan on one real client file. No ads, no SEO, no partnerships, no purchased lists. The founder builds a list of 80 named owners, sends in waves of 25, and runs the first three scans by hand within 48 hours. Conversion happens on the delivered packet — the artifact sells itself — with a Stripe link in the delivery email. Day-3 and day-7 follow-ups carry the flagged dollar total in the subject line. Every buyer is asked for one referral name, the only compounding loop in the plan.
6. 30-day execution roadmap
Days 1-5 are build: schema, 835 parser, benchmark engine, packet renderer, then live Stripe and PHI stripping. Day 5 is the MVP deadline; the product must produce a packet from a real file.
Days 6-11 are the first revenue push: assemble the 80-name list, send wave one, deliver three manual scans, post a redacted synthetic sample in two biller communities, convert with Stripe links. Day 11 is the first-sale target.
Days 12-21 harden the parser against real clearinghouse files, ship the Won/Rejected tracker and benchmark override, and run outreach waves two through four while delivering paid audits as they land.
Days 22-30 are volume: trade-group outreach, the monthly tier pitched to two past audit buyers, self-serve signup so no call is required, upload reliability work, then a final reconciliation of collected revenue, refunds and processing fees.
7. Revenue and profit scenarios
Every figure below is a projection from stated assumptions, not verified revenue, existing customers or real traction.
Conservative: three customers at $99 average, $297 gross, $10 processing, $25 hosting and API cost.
Base: twelve customers blending eight audits at $249 and four scans at $99, $2,388 gross, $76 processing, one $99 refund, $70 operating cost.
Upside: twenty-four customers at $212 average, $5,088 gross, $163 processing, $249 refunded, $200 paid marketing, $130 operating cost.
The contest ranks collected revenue minus refunds, chargebacks, processing fees and paid marketing. Hosting and API costs sit outside that figure and appear only in business profit. No scenario prices a sale below its delivery cost.
8. Risks and validation experiments
Circular benchmarking is the dominant risk: if a payer has always underpaid, the practice's own history is depressed and nothing flags. Cheap test — three free scans this week, each must exceed $500 in flagged variance; if two of three fall under $200, the benchmark definition changes.
False positives from coordination-of-benefits, modifier and sequestration adjustments erode trust. Mitigation is a suppression rule set built from real files and biller-marked 'would dispute' feedback. Kill criterion: more than 30% of a packet's flags rejected as not real.
BAA objections could stall uploads entirely. Mitigation is the claim-ID-only file spec, identifier stripping at intake, 30-day auto-delete, and a manual no-upload scan path. If all first ten conversations block on BAA, the portal leaves the paid path.
Payer rejection of reprocessing requests without a contract citation is measured directly by the tracker. If recovery sits under 15% across two paid audits, the pitch shifts from recovered dollars to documented variance.
The status-quo competitor is an Excel pivot. The counter is turnaround: a completed packet in 48 hours with a pre-filled cover form, priced at $249 against four-figure flagged totals.
9. First 24-hour action plan
Hours 1-2: repo, Supabase project, Stripe test account, claim_line schema. Hours 3-5: 835 parser against synthetic fixtures. Hours 6-8: benchmark view and variance flagger. Hours 9-11: upload page and skeleton PDF cover form. Hours 12-14: first 40 prospect names and final outreach copy. Hours 15-18: send ten messages and publish the claim-ID-only file spec. The day ends with a deployed upload form, a working variance query and ten live conversations.
This brief was generated from public data about established startups, then narrowed and scored for a solo 30-day launch. Revenue and profit figures are scenario models, not verified results, guarantees or evidence of existing customers. Validate demand yourself before you spend a month on it — and check the contest rules for what counts toward your score.
Related ideas
All ideas →Standby Desk
A private shift-fill desk for restaurants, hotels and clinics that turns their existing list of past extras into one form, one blast, one accept page, and one mission sheet PDF.
Denial Autopsy
Turns a denied behavioral-health insurance claim into a payer-specific appeal letter and proof checklist, so a solo therapist can overturn a denial in 10 minutes instead of writing it off.
DoseNote Coach
DoseNote Coach turns a 12-question client intake into a white-labeled, citation-footnoted supplement plan PDF in under two minutes, so independent health coaches stop hand-writing plans in Google Docs.