Every recoverable amount in a Fortis Scan ships with the method behind it. No "we would recover", no "we would lift", no promised numbers. This page describes the framework we apply on a full Fortis Scan engagement: four estimation phrasings, eight audit categories, the public data sources we cite, and the limits we admit to. If a specific finding cannot be traced back to a formula on this page, the finding ships with a "no projection" label instead of an invented recoverable amount.
The framework was formalized in 2026 and is refined across active engagements. Scope on any given scan depends on which tools the store has connected and the read only access we receive.
Every finding in a Fortis Scan ships with one of these four labels. The label is in the report next to the recoverable amount. If a finding does not fit one of them, the figure is dropped and the finding ships with a "no projection" label instead.
The finding is estimated against a published industry benchmark. The benchmark source is cited in the report. The benchmark is applied to the store's own observed traffic, AOV, or order count over the last 12 months.
Example. When an ESP publishes a benchmark for the vertical, we cite the band by name and the publication date. The store's reported attributed revenue is compared against the conservative low end of the band, applied to the 12 month revenue baseline from Shopify. The benchmark source is named in the finding so the figure can be re checked.
The figure is computed entirely from the store's own numbers. No external benchmark is involved. The formula is stated in plain terms inside the report.
Example. Weighted average shipping margin gap per order in the under $80 bracket, multiplied by the number of orders in that bracket over the trailing 90 days, annualized to 12 months.
The gap between the store's observed rate and a published median for the vertical, applied to the relevant 12 month base. Used most often on checkout completion, abandoned cart open rates, and product page bounce.
Example. Baymard Institute documents that mobile checkout completion falls sharply as the number of steps rises beyond what is standard for the vertical. The store ran seven steps on mobile. The completion rate gap against the Baymard documented baseline is applied to the 12 month mobile session volume. The exact Baymard figure used is cited inside the finding so the calculation can be re checked.
We report the discrepancy and stop. No recoverable amount. Used when the data cannot defensibly support a projection, when the sample is too small, or when the cause is structural and not measurable in revenue terms.
Example. Privacy policy page has not been updated since the GDPR deadline. We flag it, document the risk, and recommend a fix. We do not attach a recovery figure because the figure would be made up.
A Fortis Scan covers eight categories in scope. What we look at inside each category depends on the tools the store has connected and the read only access we receive. The list below describes the scope and the dominant phrasing used when a recoverable amount is defensible.
The public checkout flow on desktop and mobile, step count, friction points, payment options, gated promotions, and post add to cart behavior. Funnel data comes from GA4 when connected. Reference points come from documented checkout UX research.
Dominant phrasing. Industry deltas, applied to 12 month session volume when GA4 is connected.
The ESP account inventory: flows installed, flow status, list health signals, sign up form configuration, consent capture, attributed revenue share. The depth depends on the read only access we receive on the account.
Dominant phrasing. Baseline benchmarks when the ESP publishes public benchmarks for the vertical. Otherwise store's own data.
Meta Ads and Google Ads account structure when connected: campaign objectives, audience overlap, creative fatigue signals, budget pacing, attribution windows. ROAS reported in the ad platform is cross referenced with the store's actual revenue in the same window when the data is available.
Dominant phrasing. Store's own data, projected against current monthly ad spend.
The pixel and the server side conversion feed on the public site, plus deeper checks inside Meta Events Manager and GTM when we have access. We document duplicate events, missing keys, and wrong currency codes when they can be reproduced.
Dominant phrasing. Store's own data, projected against current ad spend when the overstatement is reproducible.
Public on page signals: structured data, canonical setup, indexation status of collection and product pages, Core Web Vitals on the page speed report. Search Console data is included when access is granted. Findings here usually ship with no projection because organic lift is rarely defensible as a 12 month recoverable amount on small stores.
Dominant phrasing. Observed gap with no projection on most items.
The public mobile experience on a real mid range device: landing pages, product page, cart, checkout. Friction points are documented step by step. Session recording tools like Microsoft Clarity are included when the store has them installed and grants access.
Dominant phrasing. Industry deltas, applied to 12 month mobile session volume when GA4 is connected.
Shipping rates as configured in Shopify, compared against the store's actual outbound costs when fulfillment data is shared. The loss zone is modeled by destination and cart total bracket. When fulfillment costs are not shared, the finding is qualitative only.
Dominant phrasing. Store's own data when fulfillment receipts are shared, otherwise no projection.
The public legal pages, VAT and OSS configuration when EU is in scope, cookie consent, returns and refunds policy. Most findings here ship with no projection because compliance risk is not the same as a revenue projection.
Dominant phrasing. Observed gap with no projection.
Every benchmark we use comes from a public, named source. We do not cite "industry average" without naming the source. The list below is what gets used when the data is available on a given engagement.
Every number in this report is specific to one store's data and one store's traffic. The brackets we work in tend to hold across a vertical at a similar scale. The exact recoverable amount on your store will not match the exact recoverable amount on another store's report. That is by design. The point is to put numbers your CFO can verify, not numbers that look identical across pitch decks.
Take any finding in a Fortis Scan and check it. The path is short.
That is the whole audit. There is no proprietary engine, no black box. The work is the work.
The Fortis Scan at $1,500 runs 200 plus checks across the eight categories above, with the same methodology applied to every finding. Full PDF report in 48 to 72 hours after read-only access is granted.
See pricing See the methodology in a sample report