Data security · reviewed Sep 2026
Your accounts. Our access. Every rule, written down.
The page for whoever in your company asks who exactly gets access to what. Short answers first, details below.
- Security contact: audits@marginfix.ai
- DPA and mutual NDA by email within one business day
- Signed off by Cristian Bragau
- Median finding $58k a year across 74 audits
Three rules before any account is shared
The short answers. Everything below is the detail behind them.
- Rule 01
NDA before anything
Signed before a single account is shared. Yours or ours, signed within one business day.
- Rule 02
Read only, scoped, logged
Viewer and analyst roles only, limited to the accounts in scope. We never hold admin rights and never make changes.
- Rule 03
Gone at handover
Access revoked at readout; working data deleted within 30 days. Client data is not used for model training.
Four steps. No copies left behind.
The data flow, end to end, from the roles you grant to the deletion you can verify.
- Step 01
Your accounts
You grant viewer or analyst roles from your side. We never hold your passwords.
- Step 02
Read only pull
Scoped API reads of the accounts in scope: aggregates and reports, never customer PII.
- Step 03
Analysis
Audit files are maintained in MarginFix’s local working environment. Anonymized audit information is processed through the Claude API to assist analysis. Client data is not used for model training. A senior auditor reviews and approves every finding.
- Step 04
Revoke and delete
Access revoked at your readout; working data deleted within 30 days after close. You can verify both.
What we access, exactly
Systems accessed during the audit, the access level and what each is used for.
| System | Access level | Used for |
|---|---|---|
| GA4 / analytics | Viewer | Funnel and attribution |
| Meta and Google Ads | Analyst (read only) | Spend efficiency |
| Email / CRM (e.g. Klaviyo, HubSpot) | Read only | Lifecycle coverage |
| Revenue (e.g. Shopify, Stripe, exports) | Reports only | Margin verification |
| Customer PII | Never requested | Aggregates only |
- ✓NDA before data; DPA available on request
- ✓Read only, scoped access: no admin rights, no changes made
- ✓Client data is not used for model training
- ✓Working data deleted within 30 days after close
- ✓CCPA and GDPR rights honored at audits@marginfix.ai
- ✓Subprocessor table below; the full legal list ships with the DPA
A MarginFix audit works NDA first with read only scoped access, anonymized analysis through the Claude API, no training on client data and deletion within 30 days.MarginFix, marginfix.ai/security/
Who touches the data, and under what terms
Every processor that can see anything you share with us, and the terms it works under.
| Subprocessor | Purpose | Location | Terms |
|---|---|---|---|
| Netlify, Inc. | Website hosting and form intake | Global CDN, US | DPA and SCCs |
| Zoho Corporation | CRM: lead and client records | EU (crm.zoho.eu) | DPA and SCCs |
| Anthropic (Claude API) | AI assisted analysis of anonymized audit information | Per provider terms | Commercial API terms, no training on client data |
| Google Workspace | Report delivery, client communication and call scheduling (Calendar and Meet) | EU / US | DPA and SCCs |
| Google Tag Manager and Google Analytics 4 | Site analytics, after you accept optional cookies | EU / US | DPA, consent gated |
| Microsoft Clarity | Session analytics, after you accept optional cookies | EU / US | DPA, consent gated |
| Meta Platforms (Meta Pixel) | Advertising measurement, after you accept optional cookies | EU / US | Meta data processing terms, consent gated |
Current as of Sep 2026 · changes announced 30 days ahead to active clients · the full legal list ships with the DPA.
This website
The controls on the site itself. Each one is a response header you can read, not a claim.
- Strict-Transport-SecurityHTTPS only, one year, subdomains included, on the browser preload list.
- Content-Security-PolicyHash strict: a script runs only if the build signed it. No inline handlers, frames from other sites refused, forms post to this origin only.
- Permissions-PolicyCamera, microphone, geolocation, payment and the other device APIs are switched off for every page.
- Cross-Origin policiesOpener and resource policies set to same origin; MIME sniffing off; referrers trimmed to the origin.
Responsible disclosure: /.well-known/security.txt names the channel and the policy. Check any response header in your browser's developer tools; nothing here depends on a badge.
Documents for your review
Three documents cover most security reviews. Request any of them and it arrives by email within one business day.
- Mutual NDA (template)Yours or ours. Signed within one business day.Request it →
- Data Processing AgreementWith the full subprocessor list and the retention schedule.Request it →
- Access and revocation checklistWhat we ask for, per system, and how you confirm it is gone.Request it →
Every request lands on the contact form with the topic preset, or write to audits@marginfix.ai. Reply within one business day.
In their words
What clients said after the readout. Every statement sits on a published case; open the case for the audited figure.
$134k a year found in 6 days
Contribution margin on the two top campaigns: −7% → +9%
“Revenue grew and cash didn't. Victoria rebuilt margin per order after shipping, returns and payment fees, and our two biggest prospecting campaigns were losing money on every sale. We stopped scaling them.”
$118k a year of wasted spend cut in 7 days
Rebates paid on volume never actually earned: 19% → 1%
“Rebate expense outran distributor volume and we could not see why. Cristian reconciled rebates paid against volume actually shipped and paid for, tier by tier: accruals had run off gross purchase orders. Quarterly true ups now catch it.”
$96k a year of wasted spend cut in 7 days
App installs that funded a real account: 24% → 63%
“Installs looked great and funded accounts stagnated. David followed sign ups through to funding, and the bonus was buying registrations from people who claimed it and never transacted. Incentives now pay on activation, not the download.”
$92k a year found in 7 days
Enterprise deal margin lost to quarter end: 5 pts lost → 0 pts lost
“Bookings hit target while margin slipped. Cristian split our discounting by timing and rep: concessions spiked in the last days of every quarter because buyers had learned to wait. Approval thresholds fixed it without losing deals.”
$88k a year of wasted spend cut in 7 days
Paid social proven to lift store visits: ~1 in 3 → 9 in 10
“Our CFO doubted that online ads drove store visits. David's matched market geo lift test proved the doubt right: footfall barely changed where the ads were off. The budget moved to markets and channels that passed the test.”
−27% risk adjusted CAC in 7 days
Risk adjusted acquisition cost per loan: $540 → $400
“Our cheapest channel was our worst credit. David split default rates by acquisition source: the low CAC channels brought borrowers who defaulted materially more often. Budget now follows risk adjusted cost per loan, not just CAC.”
$88k a year of wasted spend cut in 7 days
Congress spend with a tracked return: 8% → 88%
“Six figures a year in congresses and KOL sponsorships, renewed on habit. Leslie tried to connect it to share of voice, engagement or prescribing and found no instrumentation at all. The unmeasurable part now funds engagement we can track.”
$83k a year of wasted spend cut in 7 days
Referral rewards on genuine, nongamed volume: 67% → 98%
“Referral payouts grew faster than real users. David examined the volume behind the rewards: self referrals through second accounts, circular trades, wash volume. Integrity checks now run before any payout; gamed rewards were clawed back.”
$83k a year found in 7 days
Interchange margin on the worst channel: +0.4% → +2.4%
“Every channel cost the same per merchant, so we treated them as interchangeable. David split profitability down to the interchange line: some channels brought merchants who earned us a fraction. Budget now follows margin, not just CAC.”
$79k a year found in 7 days
Contract margin vs live input costs: 9% → 16%
“Volume was stable and contract margin kept eroding. Cristian compared resin, board and energy cost movements against the surcharges we actually passed through, and the lag was costing us every month. An indexed pass through fixed it.”
$78k a year of wasted spend cut in 7 days
High intent visitors paid for twice: ~1 in 3 → ~1 in 30
“Two teams, two search programs, one visitor bought twice. Leslie looked at branded and disease awareness search together for the first time: about one high intent visitor in three was paid for twice. Shared negatives and one plan ended it.”
$74k a year of wasted spend cut in 7 days
Share of all impressions hitting no convert inventory: 40% → 6%
“Forty percent of our impressions landed where buyers never convert. Cristian's delivery teardown found placement, dayparting and geo settings untouched since launch on a seven figure budget. We rebuilt them.”
$74k a year of wasted spend cut in 7 days
Courier bonuses paid into already full zones: 22% → 3%
“Incentive spend rose while service levels stayed flat. Alex matched bonus payouts against real time courier availability by zone: in dense areas we paid couriers who would have worked anyway. Bonuses now fire only where coverage is short.”
$72k a year found in 7 days
Portfolio margin across the product mix: 11% → 17%
“One flat commission across very different products steered brokers to the wrong mix. David set commission against each product's margin and loss ratio; thin lines were being sold near unprofitable. Commission now follows the margin band.”
$72k a year found in 7 days
Landed margin on the flat pack lines: −1% → +6%
“We held prices for two years and freight did not. Cristian rebuilt landed cost from current carrier invoices: oversized carton surcharges had erased the margin on every flat pack line. We repriced and redesigned the worst packaging.”
$71k a year of wasted spend cut in 7 days
Loyalty program margin vs its liability: −8% → +6%
“Loyalty was untouchable, so nobody had priced it. David set our accrued point liability against the incremental margin the program drove, and it was underwater. We restructured earn and burn without gutting the rewards members value.”
$71k a year of wasted spend cut in 7 days
Acquisition cost vs reimbursement per test: 122% → 78%
“Acquisition and reimbursement sat with different teams, so nobody compared them. Cristian reconciled cost per patient against reimbursement per test: several high volume tests lost money on every patient we marketed in. We capped those.”
$70k a year found in 6 days
Billable usage actually invoiced: 92% → 99%
“I expected a slide deck. Alex reconciled metered usage against invoices and found a whole category of overages recorded for analytics but never billed. A config fix, no price change, and customers expected to pay for it anyway.”
$69k a year of wasted spend cut in 7 days
Online margin once its returns are charged in: 2% → 9%
“Online looked as profitable as our stores, which felt wrong. David charged returns processing to the channel that generated each return, and online's margin fell from flattering to honest. We stopped scaling the wrong channel.”
$69k a year of wasted spend cut in 7 days
Co-op budget spent promoting rival brands: 28% → 3%
“Our co-op money was promoting our competitors. Cristian reviewed what distributors actually spent it on: campaigns featured rival brands next to ours, sometimes more prominently. Reimbursement now needs proof of brand presence.”
$68k a year found in 6 days
Full price sell through: 61% → 78%
“Our best customers had learned to wait for a code. Victoria showed that a large share of discounted orders would have happened at full price anyway. We moved to a tiered structure for new customers and bigger baskets, and volume held.”
$66k a year of wasted spend cut in 6 days
Paid clicks the site already won organically: ~1 in 3 → ~1 in 30
“We were bidding on terms we already ranked for. Alex paused paid search on those terms in a test and bookings barely moved. Roughly one paid click in three we would have won for free. That spend now goes where paid actually adds demand.”
$64k a year of wasted spend cut in 7 days
Detailing spend that moved no prescribing: 28% → 4%
“Reach and call counts looked healthy. Leslie matched detailing and samples to prescribing change per HCP, and 28% of the effort landed on physicians it never moved. Field time now follows responsiveness, not coverage targets.”
$64k a year of wasted spend cut in 6 days
Budget lost to the two teams overbidding: 22% → 3%
“Our HCP and patient agencies were bidding against each other for the same impressions. Cristian examined both programs together for the first time and merged audience planning. Reach held, and the internal premium was gone.”
$64k a year found in 6 days
Dealer margin on the engineered range: 14% → 20%
“One dealer discount across the whole catalog felt simple. Cristian set it against margin at product level: commodity fittings got a generous cut while our engineered range subsidized them. Discount tiers by margin band fixed it.”
$63k a year of wasted spend cut in 6 days
Supply incentives hitting a real shortage: 35% → 94%
“Our onboarding bonus solved a shortage we no longer had. Alex checked every incentive against current supply by category, and we were still paying to recruit providers in saturated ones. We kept bonuses only where supply is genuinely short.”
$63k a year found in 6 days
Active users actually being billed: 93% → 99%
“Expansion lagged what we saw inside accounts. Victoria reconciled active users against billed seats: many customers used more seats than they paid for. Automated true ups now capture it, and nobody pushed back.”
$62k a year found in 6 days
Qualified inbound calls that became bookings: 2 in 3 → 9 in 10
“Marketing hit every target and revenue stalled. Leslie followed our qualified calls into intake and about a third were dropped at busy times or never followed up. Fixing staffing and call handling converted demand we had already paid for.”
$61k a year found in 6 days
Net margin on hero ASINs, fees & returns in: −2% → +6%
“ACoS looked efficient while our hero ASINs lost money. Victoria rebuilt contribution per ASIN with FBA fees, storage, referral fees and returns loaded in. We repriced and repacked the affected lines.”
$61k a year found in 6 days
Support cost vs fee on small accounts: 140% → 85%
“Small accounts used as much planner time as large ones for a fraction of the fee. David modeled cost to serve by tier: below a certain balance the support ran at a loss. Human planning now sits with the tiers whose fee covers it.”
+26% media efficiency in 7 days
Media weight landing in real demand windows: 45% → 88%
“Our flighting calendar was older than our customers' habits. Cristian compared spend timing against when the category actually converts: our heaviest weight landed in weak windows. Same budget, real demand windows, efficiency up 26%.”
$59k a year found in 6 days
Contribution margin on the three bestsellers: −2% → +7%
“The more our bestsellers sold, the flatter profit looked. Victoria rebuilt landed cost from the supplier invoice up, and our three top SKUs were priced below cost after freight and packaging. We repriced in stages and volume held.”
$58k a year found in 6 days
Effective take rate after the old waivers: 9.5% → 12.5%
“Our effective take rate sat below our stated one for years. Alex traced it cohort by cohort to launch era fee waivers that had never been switched off. We retired them on matured cohorts and kept relief only for new participants.”
$58k a year found in 6 days
Margin on big ticket orders, delivery in: −1% → +6%
“Free delivery and assembly were eating our biggest orders. David rebuilt the true cost with two person crews, vehicle time and failed deliveries, and some lines were near break even. Tiered service pricing on the largest items fixed it.”
$58k a year of wasted spend cut in 6 days
Media spend outside eligible patient windows: 24% → 2%
“Reach looked healthy, but a quarter of our media reached people who could not act on it. Leslie compared delivery against the eligible population and its windows: 24% of spend fell outside them. Flighting now follows eligibility.”
$57k a year found in 6 days
Annual discount vs the churn it prevents: 1.8× over → right sized
“We were discounting loyalty we already had. Victoria compared the annual discount against the churn it actually prevented: a large share of members taking it would have stayed anyway. We right sized it and kept the deep offer for real churn risk.”
$57k a year found in 6 days
Margin on the smallest orders: −3% → +6%
“Our smallest orders cost more to pick, pack and ship than they earned. Cristian rebuilt handling cost per order and set a minimum that covers it. A small order charge and consolidation incentives did the rest.”
$57k a year found in 6 days
Margin on configured upsell orders: −3% → +8%
“Configured orders always carried thinner margins than manual quotes. Cristian priced real configurations down to installed cost: default discounts were selling premium finishes at a loss on every order. A margin floor in the tool ended it.”
$54k a year found in 6 days
Fulfillment margin on click and collect orders: −2% → +5%
“Click and collect was never free. David rebuilt the cost of a collection order, the picking, staging and handover labor hiding in store payroll, and on many order types it matched the delivery it replaced. We repriced those.”
$54k a year found in 7 days
Retainers priced below their delivery cost: 6 of 34 → 0 of 34
“Our blended margin said every account was fine. Leslie broke it apart account by account: six of thirty four retainers were priced below delivery cost, funded by the profitable ones. We repriced or exited them, our own advice.”
$53k a year of wasted spend cut in 7 days
Incremental ROAS: original campaign vs replacement channels: 0.2× → 2.1×
“Our agency's hero campaign was the biggest loser. Victoria ran a geo holdout and sales barely moved when it was off. Two years of attributed revenue, almost none of it caused by the ad. The budget now goes to channels that pass that test.”
$52k a year found in 6 days
Free account cost vs the plan revenue it earns: 130% → 78%
“Free felt cheap until Alex modeled cost to serve. Support tickets, infrastructure and engineering time made a free account cost more than our entry plan earned. We reset free tier limits and repriced the entry plan.”
$52k a year found in 5 days
Involuntary churn from failed payments: 4.1% → 1.6%
“A slice of our churn was never churn. Victoria split voluntary from involuntary and found expired cards getting one token retry before we wrote the subscriber off. Smarter retries and a card updater cut involuntary churn from 4.1% to 1.6%.”
$52k a year found in 6 days
Margin on the fixed fee engagements: 9% → 17%
“Busy, revenue holding, and profit per partner hour falling. Leslie reconciled logged hours against every fixed fee package: complexity had pushed many below cost. We repriced and rescoped them, with a scope trigger so it stays fixed.”
$49k a year found in 6 days
Stock marked down before demand softened: ~40% → ~8%
“Our markdown calendar was cutting prices on lines still selling at full price. David compared markdown timing against each SKU's real demand curve. Discounts now trigger on live sell through instead of a date, and clearance still clears.”
$49k a year found in 6 days
Margin on the thin vet channel lines: −2% → +5%
“One discount for the whole range looked competitive. Leslie checked it against margin by product line, and on our thinner lines it pushed them below cost. Discount tiers now follow margin band, and the channel kept its competitive pricing.”
$48k a year found in 5 days
Catalog margin after the dead SKU tail: 11% → 16%
“A broad catalog felt like an asset. Cristian ranked every SKU by velocity against the cost it carried, and a long tail of dead products was dragging the whole range's margin. We retired them and set a velocity threshold to stay listed.”
$47k a year found in 5 days
Orders shipped free below break even: 1 in 3 → ~1 in 20
“One order in three shipped free at a loss. Our threshold was set years ago and never revisited, and the cost sat in logistics where nobody looked. Victoria reset it, and average order value went up, not down.”
$47k a year of wasted spend cut in 6 days
Referral fees paid on self sourced deals: ~24% → ~2%
“Partner fees grew faster than partner pipeline. Alex reconciled partner attributed deals against their true origin: many were sourced by our own team with a partner tagged late. Fees now require documented sourcing.”
$47k a year found in 6 days
Would be pauses lost to a full cancel: ~4 in 10 → ~1 in 10
“Pause was buried and cancel was one tap away. Victoria mapped the pause, skip and cancel journeys and about four in ten would be pauses ended as full cancellations. We rebuilt the flow and offer pause first inside the cancel journey now.”
$47k a year found in 6 days
Delivery hours actually billed to clients: ~80% → 98%
“Every project ran thinner than its day rate. Leslie reconciled real effort against billing and found travel, kickoff and early onboarding absorbed as goodwill. Those hours are written into every scope now.”
$44k a year of wasted spend cut in 5 days
Commissions paid on sales you already owned: ~29% → ~2%
“Affiliate revenue kept climbing and margin didn't. Victoria tested incrementality and coupon partners were collecting commission on customers who arrived through branded search and direct. We now pay for demand partners actually create.”
$44k a year found in 7 days
Take rate actually kept vs the 15% contracted: 11.1% → 14.0%
“Finance couldn't explain why we kept less than our contracted take rate. Alex traced every dollar from gross booking to net revenue and found it in refunds nobody owned. Clear rules and one accountable owner closed the gap.”
$44k a year found in 5 days
Job margin once callbacks are counted: −1% → +5%
“Jobs kept landing below what we quoted. Leslie loaded callbacks and warranty visits back onto the jobs that caused them, and the quoting model had never counted them. Quotes now price the whole job, not the first visit.”
$41k annual budget redeployment identified in 5 days
Share of all paid spend that was incremental: 28% → 82%
“Our safest lines were the least incremental. Alex showed brand term search and retargeting were intercepting demand from sales calls and word of mouth, then invoicing for it. We cut both to a defensive minimum.”
$41k a year found in 6 days
Shipping margin on the monthly cohort: −4% → +8%
“A flat shipping tier felt fair. Victoria mapped shipping cost against cohort value and our most loyal subscribers were subsidizing the least loyal ones. Tiers now follow order frequency, so loyalty stops paying for churn.”
$41k a year found in 5 days
Promo discount covered by vendor funding: 78% → 96%
“The vendor pays for it, we assumed. David reconciled every cofunded promotion against the support we actually received, and the discounts routinely ran deeper than the funding. Renegotiated terms and promo guardrails closed it.”
$41k a year of wasted spend cut in 6 days
Trade show spend tied to a booked order: 6% → 90%
“Six figures a year on trade shows because we always had. Cristian traced the spend to booked orders and found no link at all; in our specifier led category the events reached the wrong buyers. The budget now follows traceable pipeline.”
$39k a year of wasted spend cut in 5 days
Search cost per click across the clinics: $2.40 → $1.78
“Our cost per click climbed with no new competitor in sight. Leslie found the competitor was us: clinics bidding against each other on the same keywords in the same auctions. One geo partitioned account structure ended it.”
$36k a year of wasted spend cut in 7 days
Content spend with a traceable path to revenue: 9% → 92%
“Eighteen months of content, not one sourced opportunity. Cristian traced the program forward through the funnel and could not connect it to pipeline. We kept the assets sales actually used and moved the rest of the budget to demand capture.”
$134k a year found in 6 days
Contribution margin on the two top campaigns: −7% → +9%
“Revenue grew and cash didn't. Victoria rebuilt margin per order after shipping, returns and payment fees, and our two biggest prospecting campaigns were losing money on every sale. We stopped scaling them.”
$118k a year of wasted spend cut in 7 days
Rebates paid on volume never actually earned: 19% → 1%
“Rebate expense outran distributor volume and we could not see why. Cristian reconciled rebates paid against volume actually shipped and paid for, tier by tier: accruals had run off gross purchase orders. Quarterly true ups now catch it.”
$96k a year of wasted spend cut in 7 days
App installs that funded a real account: 24% → 63%
“Installs looked great and funded accounts stagnated. David followed sign ups through to funding, and the bonus was buying registrations from people who claimed it and never transacted. Incentives now pay on activation, not the download.”
$92k a year found in 7 days
Enterprise deal margin lost to quarter end: 5 pts lost → 0 pts lost
“Bookings hit target while margin slipped. Cristian split our discounting by timing and rep: concessions spiked in the last days of every quarter because buyers had learned to wait. Approval thresholds fixed it without losing deals.”
$88k a year of wasted spend cut in 7 days
Paid social proven to lift store visits: ~1 in 3 → 9 in 10
“Our CFO doubted that online ads drove store visits. David's matched market geo lift test proved the doubt right: footfall barely changed where the ads were off. The budget moved to markets and channels that passed the test.”
−27% risk adjusted CAC in 7 days
Risk adjusted acquisition cost per loan: $540 → $400
“Our cheapest channel was our worst credit. David split default rates by acquisition source: the low CAC channels brought borrowers who defaulted materially more often. Budget now follows risk adjusted cost per loan, not just CAC.”
$88k a year of wasted spend cut in 7 days
Congress spend with a tracked return: 8% → 88%
“Six figures a year in congresses and KOL sponsorships, renewed on habit. Leslie tried to connect it to share of voice, engagement or prescribing and found no instrumentation at all. The unmeasurable part now funds engagement we can track.”
$83k a year of wasted spend cut in 7 days
Referral rewards on genuine, nongamed volume: 67% → 98%
“Referral payouts grew faster than real users. David examined the volume behind the rewards: self referrals through second accounts, circular trades, wash volume. Integrity checks now run before any payout; gamed rewards were clawed back.”
$83k a year found in 7 days
Interchange margin on the worst channel: +0.4% → +2.4%
“Every channel cost the same per merchant, so we treated them as interchangeable. David split profitability down to the interchange line: some channels brought merchants who earned us a fraction. Budget now follows margin, not just CAC.”
$79k a year found in 7 days
Contract margin vs live input costs: 9% → 16%
“Volume was stable and contract margin kept eroding. Cristian compared resin, board and energy cost movements against the surcharges we actually passed through, and the lag was costing us every month. An indexed pass through fixed it.”
$78k a year of wasted spend cut in 7 days
High intent visitors paid for twice: ~1 in 3 → ~1 in 30
“Two teams, two search programs, one visitor bought twice. Leslie looked at branded and disease awareness search together for the first time: about one high intent visitor in three was paid for twice. Shared negatives and one plan ended it.”
$74k a year of wasted spend cut in 7 days
Share of all impressions hitting no convert inventory: 40% → 6%
“Forty percent of our impressions landed where buyers never convert. Cristian's delivery teardown found placement, dayparting and geo settings untouched since launch on a seven figure budget. We rebuilt them.”
$74k a year of wasted spend cut in 7 days
Courier bonuses paid into already full zones: 22% → 3%
“Incentive spend rose while service levels stayed flat. Alex matched bonus payouts against real time courier availability by zone: in dense areas we paid couriers who would have worked anyway. Bonuses now fire only where coverage is short.”
$72k a year found in 7 days
Portfolio margin across the product mix: 11% → 17%
“One flat commission across very different products steered brokers to the wrong mix. David set commission against each product's margin and loss ratio; thin lines were being sold near unprofitable. Commission now follows the margin band.”
$72k a year found in 7 days
Landed margin on the flat pack lines: −1% → +6%
“We held prices for two years and freight did not. Cristian rebuilt landed cost from current carrier invoices: oversized carton surcharges had erased the margin on every flat pack line. We repriced and redesigned the worst packaging.”
$71k a year of wasted spend cut in 7 days
Loyalty program margin vs its liability: −8% → +6%
“Loyalty was untouchable, so nobody had priced it. David set our accrued point liability against the incremental margin the program drove, and it was underwater. We restructured earn and burn without gutting the rewards members value.”
$71k a year of wasted spend cut in 7 days
Acquisition cost vs reimbursement per test: 122% → 78%
“Acquisition and reimbursement sat with different teams, so nobody compared them. Cristian reconciled cost per patient against reimbursement per test: several high volume tests lost money on every patient we marketed in. We capped those.”
$70k a year found in 6 days
Billable usage actually invoiced: 92% → 99%
“I expected a slide deck. Alex reconciled metered usage against invoices and found a whole category of overages recorded for analytics but never billed. A config fix, no price change, and customers expected to pay for it anyway.”
$69k a year of wasted spend cut in 7 days
Online margin once its returns are charged in: 2% → 9%
“Online looked as profitable as our stores, which felt wrong. David charged returns processing to the channel that generated each return, and online's margin fell from flattering to honest. We stopped scaling the wrong channel.”
$69k a year of wasted spend cut in 7 days
Co-op budget spent promoting rival brands: 28% → 3%
“Our co-op money was promoting our competitors. Cristian reviewed what distributors actually spent it on: campaigns featured rival brands next to ours, sometimes more prominently. Reimbursement now needs proof of brand presence.”
$68k a year found in 6 days
Full price sell through: 61% → 78%
“Our best customers had learned to wait for a code. Victoria showed that a large share of discounted orders would have happened at full price anyway. We moved to a tiered structure for new customers and bigger baskets, and volume held.”
$66k a year of wasted spend cut in 6 days
Paid clicks the site already won organically: ~1 in 3 → ~1 in 30
“We were bidding on terms we already ranked for. Alex paused paid search on those terms in a test and bookings barely moved. Roughly one paid click in three we would have won for free. That spend now goes where paid actually adds demand.”
$64k a year of wasted spend cut in 7 days
Detailing spend that moved no prescribing: 28% → 4%
“Reach and call counts looked healthy. Leslie matched detailing and samples to prescribing change per HCP, and 28% of the effort landed on physicians it never moved. Field time now follows responsiveness, not coverage targets.”
$64k a year of wasted spend cut in 6 days
Budget lost to the two teams overbidding: 22% → 3%
“Our HCP and patient agencies were bidding against each other for the same impressions. Cristian examined both programs together for the first time and merged audience planning. Reach held, and the internal premium was gone.”
$64k a year found in 6 days
Dealer margin on the engineered range: 14% → 20%
“One dealer discount across the whole catalog felt simple. Cristian set it against margin at product level: commodity fittings got a generous cut while our engineered range subsidized them. Discount tiers by margin band fixed it.”
$63k a year of wasted spend cut in 6 days
Supply incentives hitting a real shortage: 35% → 94%
“Our onboarding bonus solved a shortage we no longer had. Alex checked every incentive against current supply by category, and we were still paying to recruit providers in saturated ones. We kept bonuses only where supply is genuinely short.”
$63k a year found in 6 days
Active users actually being billed: 93% → 99%
“Expansion lagged what we saw inside accounts. Victoria reconciled active users against billed seats: many customers used more seats than they paid for. Automated true ups now capture it, and nobody pushed back.”
$62k a year found in 6 days
Qualified inbound calls that became bookings: 2 in 3 → 9 in 10
“Marketing hit every target and revenue stalled. Leslie followed our qualified calls into intake and about a third were dropped at busy times or never followed up. Fixing staffing and call handling converted demand we had already paid for.”
$61k a year found in 6 days
Net margin on hero ASINs, fees & returns in: −2% → +6%
“ACoS looked efficient while our hero ASINs lost money. Victoria rebuilt contribution per ASIN with FBA fees, storage, referral fees and returns loaded in. We repriced and repacked the affected lines.”
$61k a year found in 6 days
Support cost vs fee on small accounts: 140% → 85%
“Small accounts used as much planner time as large ones for a fraction of the fee. David modeled cost to serve by tier: below a certain balance the support ran at a loss. Human planning now sits with the tiers whose fee covers it.”
+26% media efficiency in 7 days
Media weight landing in real demand windows: 45% → 88%
“Our flighting calendar was older than our customers' habits. Cristian compared spend timing against when the category actually converts: our heaviest weight landed in weak windows. Same budget, real demand windows, efficiency up 26%.”
$59k a year found in 6 days
Contribution margin on the three bestsellers: −2% → +7%
“The more our bestsellers sold, the flatter profit looked. Victoria rebuilt landed cost from the supplier invoice up, and our three top SKUs were priced below cost after freight and packaging. We repriced in stages and volume held.”
$58k a year found in 6 days
Effective take rate after the old waivers: 9.5% → 12.5%
“Our effective take rate sat below our stated one for years. Alex traced it cohort by cohort to launch era fee waivers that had never been switched off. We retired them on matured cohorts and kept relief only for new participants.”
$58k a year found in 6 days
Margin on big ticket orders, delivery in: −1% → +6%
“Free delivery and assembly were eating our biggest orders. David rebuilt the true cost with two person crews, vehicle time and failed deliveries, and some lines were near break even. Tiered service pricing on the largest items fixed it.”
$58k a year of wasted spend cut in 6 days
Media spend outside eligible patient windows: 24% → 2%
“Reach looked healthy, but a quarter of our media reached people who could not act on it. Leslie compared delivery against the eligible population and its windows: 24% of spend fell outside them. Flighting now follows eligibility.”
$57k a year found in 6 days
Annual discount vs the churn it prevents: 1.8× over → right sized
“We were discounting loyalty we already had. Victoria compared the annual discount against the churn it actually prevented: a large share of members taking it would have stayed anyway. We right sized it and kept the deep offer for real churn risk.”
$57k a year found in 6 days
Margin on the smallest orders: −3% → +6%
“Our smallest orders cost more to pick, pack and ship than they earned. Cristian rebuilt handling cost per order and set a minimum that covers it. A small order charge and consolidation incentives did the rest.”
$57k a year found in 6 days
Margin on configured upsell orders: −3% → +8%
“Configured orders always carried thinner margins than manual quotes. Cristian priced real configurations down to installed cost: default discounts were selling premium finishes at a loss on every order. A margin floor in the tool ended it.”
$54k a year found in 6 days
Fulfillment margin on click and collect orders: −2% → +5%
“Click and collect was never free. David rebuilt the cost of a collection order, the picking, staging and handover labor hiding in store payroll, and on many order types it matched the delivery it replaced. We repriced those.”
$54k a year found in 7 days
Retainers priced below their delivery cost: 6 of 34 → 0 of 34
“Our blended margin said every account was fine. Leslie broke it apart account by account: six of thirty four retainers were priced below delivery cost, funded by the profitable ones. We repriced or exited them, our own advice.”
$53k a year of wasted spend cut in 7 days
Incremental ROAS: original campaign vs replacement channels: 0.2× → 2.1×
“Our agency's hero campaign was the biggest loser. Victoria ran a geo holdout and sales barely moved when it was off. Two years of attributed revenue, almost none of it caused by the ad. The budget now goes to channels that pass that test.”
$52k a year found in 6 days
Free account cost vs the plan revenue it earns: 130% → 78%
“Free felt cheap until Alex modeled cost to serve. Support tickets, infrastructure and engineering time made a free account cost more than our entry plan earned. We reset free tier limits and repriced the entry plan.”
$52k a year found in 5 days
Involuntary churn from failed payments: 4.1% → 1.6%
“A slice of our churn was never churn. Victoria split voluntary from involuntary and found expired cards getting one token retry before we wrote the subscriber off. Smarter retries and a card updater cut involuntary churn from 4.1% to 1.6%.”
$52k a year found in 6 days
Margin on the fixed fee engagements: 9% → 17%
“Busy, revenue holding, and profit per partner hour falling. Leslie reconciled logged hours against every fixed fee package: complexity had pushed many below cost. We repriced and rescoped them, with a scope trigger so it stays fixed.”
$49k a year found in 6 days
Stock marked down before demand softened: ~40% → ~8%
“Our markdown calendar was cutting prices on lines still selling at full price. David compared markdown timing against each SKU's real demand curve. Discounts now trigger on live sell through instead of a date, and clearance still clears.”
$49k a year found in 6 days
Margin on the thin vet channel lines: −2% → +5%
“One discount for the whole range looked competitive. Leslie checked it against margin by product line, and on our thinner lines it pushed them below cost. Discount tiers now follow margin band, and the channel kept its competitive pricing.”
$48k a year found in 5 days
Catalog margin after the dead SKU tail: 11% → 16%
“A broad catalog felt like an asset. Cristian ranked every SKU by velocity against the cost it carried, and a long tail of dead products was dragging the whole range's margin. We retired them and set a velocity threshold to stay listed.”
$47k a year found in 5 days
Orders shipped free below break even: 1 in 3 → ~1 in 20
“One order in three shipped free at a loss. Our threshold was set years ago and never revisited, and the cost sat in logistics where nobody looked. Victoria reset it, and average order value went up, not down.”
$47k a year of wasted spend cut in 6 days
Referral fees paid on self sourced deals: ~24% → ~2%
“Partner fees grew faster than partner pipeline. Alex reconciled partner attributed deals against their true origin: many were sourced by our own team with a partner tagged late. Fees now require documented sourcing.”
$47k a year found in 6 days
Would be pauses lost to a full cancel: ~4 in 10 → ~1 in 10
“Pause was buried and cancel was one tap away. Victoria mapped the pause, skip and cancel journeys and about four in ten would be pauses ended as full cancellations. We rebuilt the flow and offer pause first inside the cancel journey now.”
$47k a year found in 6 days
Delivery hours actually billed to clients: ~80% → 98%
“Every project ran thinner than its day rate. Leslie reconciled real effort against billing and found travel, kickoff and early onboarding absorbed as goodwill. Those hours are written into every scope now.”
$44k a year of wasted spend cut in 5 days
Commissions paid on sales you already owned: ~29% → ~2%
“Affiliate revenue kept climbing and margin didn't. Victoria tested incrementality and coupon partners were collecting commission on customers who arrived through branded search and direct. We now pay for demand partners actually create.”
$44k a year found in 7 days
Take rate actually kept vs the 15% contracted: 11.1% → 14.0%
“Finance couldn't explain why we kept less than our contracted take rate. Alex traced every dollar from gross booking to net revenue and found it in refunds nobody owned. Clear rules and one accountable owner closed the gap.”
$44k a year found in 5 days
Job margin once callbacks are counted: −1% → +5%
“Jobs kept landing below what we quoted. Leslie loaded callbacks and warranty visits back onto the jobs that caused them, and the quoting model had never counted them. Quotes now price the whole job, not the first visit.”
$41k annual budget redeployment identified in 5 days
Share of all paid spend that was incremental: 28% → 82%
“Our safest lines were the least incremental. Alex showed brand term search and retargeting were intercepting demand from sales calls and word of mouth, then invoicing for it. We cut both to a defensive minimum.”
$41k a year found in 6 days
Shipping margin on the monthly cohort: −4% → +8%
“A flat shipping tier felt fair. Victoria mapped shipping cost against cohort value and our most loyal subscribers were subsidizing the least loyal ones. Tiers now follow order frequency, so loyalty stops paying for churn.”
$41k a year found in 5 days
Promo discount covered by vendor funding: 78% → 96%
“The vendor pays for it, we assumed. David reconciled every cofunded promotion against the support we actually received, and the discounts routinely ran deeper than the funding. Renegotiated terms and promo guardrails closed it.”
$41k a year of wasted spend cut in 6 days
Trade show spend tied to a booked order: 6% → 90%
“Six figures a year on trade shows because we always had. Cristian traced the spend to booked orders and found no link at all; in our specifier led category the events reached the wrong buyers. The budget now follows traceable pipeline.”
$39k a year of wasted spend cut in 5 days
Search cost per click across the clinics: $2.40 → $1.78
“Our cost per click climbed with no new competitor in sight. Leslie found the competitor was us: clinics bidding against each other on the same keywords in the same auctions. One geo partitioned account structure ended it.”
$36k a year of wasted spend cut in 7 days
Content spend with a traceable path to revenue: 9% → 92%
“Eighteen months of content, not one sourced opportunity. Cristian traced the program forward through the funnel and could not connect it to pipeline. We kept the assets sales actually used and moved the rest of the budget to demand capture.”
Questions before you grant access
The ones security reviews ask first. If yours is not here, ask the person accountable for the answer.
Do you need admin access to anything?
No. Viewer or analyst roles, granted from your side, limited to the accounts in scope. We never hold passwords and never make changes in your accounts.
Where is our data processed?
Audit files are maintained in MarginFix’s local working environment. Anonymized audit information is processed through the Claude API to assist analysis. Client data is not used for model training. A senior auditor reviews and approves every finding. We do not claim a processing region we have not verified; the provider terms govern where the API runs.
Is our data used to train AI models?
No. Client data is not used for model training. Anonymized audit information is processed through the Claude API to assist analysis, and a senior auditor reviews and approves every finding.
What happens to the data after the readout?
Access is revoked at the readout. Working data is deleted within 30 days of close; where backups exist they are purged within 90 days. The retention schedule is on the privacy page.
Can we use our own NDA?
Yes. Send yours or ask for ours; either is signed within one business day, before any account is shared.
Do you ever see customer level personal data?
It is never requested. The audit works on aggregates, reports and exports; customer PII stays in your systems.
Security cleared? Let’s find your margin.
Fixed fee. 5-7 working days. $10k to $50k of findings, or your money back.
Fixed fee. No retainer. NDA first.
Go deeper
Read on
- Privacy PolicyWhat we collect, the retention schedule, your rights and every processor with its lawful basis.→
- Terms and ConditionsThe engagement, the findings guarantee, corrections, cancellation and which document governs.→
- How it worksThe week, day by day: access, analysis, readout and the report.→
- FAQThe guarantee, data access, the timeline and what an audit does not cover.→
- ContactSecurity questions, DPA and NDA requests. Reply within one business day.→
- Case studies74 anonymized audits across 9 industries.→