Method
What does Salesforce technical debt cost? The formula we use.
VanguardOps prices every finding with the same formula: hours to fix times a labor rate gives the one-time cost, and a yearly carrying cost by type of problem gives the cost of leaving it alone. Every number in that formula is on this page, and you can run it yourself below.
By VanguardOps, written by a Salesforce architect with 13+ years in enterprise orgs. Updated November 11, 2026.
Technical debt cost calculator
Add the problems you already know about. The calculator applies the formula the product uses. It starts with three example findings; edit or remove them.
Finding 1
$3,024 to fix (17.3 hrs) · $3,024 a year · $9,072 over 3 years
Finding 2
$1,231 to fix (13 hrs) · $9,180 a year · $27,540 over 3 years
Finding 3
$669 to fix (7 hrs) · $2,112 a year · $6,336 over 3 years
- 3-year cost if nothing changes
- $42,948
- One-time cost to fix
- $4,924
- Hours of work
- 37.3
Directional estimates for prioritizing and budgeting, not a quote. This runs in your browser; nothing you enter is sent to VanguardOps.
What is the formula?
Each finding gets three figures.
- One-time fix cost = hours to fix × hourly rate × scope.
- Yearly carrying cost = the yearly cost for that type of problem × scope.
- 3-year cost = yearly carrying cost × 3.
How many hours does a fix take?
The model uses three sizes of fix, chosen by the finding’s severity.
| Severity | Hours | What that covers |
|---|---|---|
| Low | Hours: 1.5 | What that covers: Delete something unused, add a description, simple cleanup |
| Medium | Hours: 4 | What that covers: Needs investigation and a safe way to make the change |
| High or critical | Hours: 12 | What that covers: Has dependencies and needs testing |
What hourly rates does it use?
US blended rates for the role that would do the work. The role is picked from what the finding is about: Apex and code go to a developer; flows, Process Builder, integrations and sharing go to an architect; fields, permission sets and validation rules go to an admin.
| Role | Rate |
|---|---|
| Admin | $95 an hour |
| Developer | $145 an hour |
| Architect | $175 an hour |
What does it cost to leave a problem alone?
Each type of problem has a yearly carrying cost: the upkeep it adds plus a risk-weighted allowance for what it could break. These are conservative on purpose.
| Type of problem | Per year | Over 3 years | Example |
|---|---|---|---|
| Security risk | Per year: $8,500 | Over 3 years: $25,500 | Example: A profile with Modify All Data handed to many users |
| Compliance gap | Per year: $7,800 | Over 3 years: $23,400 | Example: A control an auditor would expect that isn't there |
| Data integrity | Per year: $5,200 | Over 3 years: $15,600 | Example: Apex triggers with no test coverage |
| Integration health | Per year: $4,100 | Over 3 years: $12,300 | Example: A named credential on basic auth, an expired certificate |
| Performance | Per year: $3,400 | Over 3 years: $10,200 | Example: An org limit close to its ceiling |
| Retired or deprecated feature | Per year: $2,100 | Over 3 years: $6,300 | Example: Process Builder processes and workflow rules still running |
| Unused metadata | Per year: $1,200 | Over 3 years: $3,600 | Example: Permission sets nobody holds |
| Missing documentation | Per year: $600 | Over 3 years: $1,800 | Example: Custom fields with no description |
One exception: unused paid licenses. Those are priced at what the seats cost, using list prices of $175 a month for a Salesforce seat and $25 a month for a Salesforce Platform seat. Contract prices are often lower, and the finding says so.
How does the number of affected items change the cost?
A finding that covers 40 permission sets costs more than one that covers a single permission set, but not 40 times more. The baseline covers the first item. Each extra item adds 4% of the baseline, and the multiplier stops at 6.
- 1 item: ×1
- 10 items: ×1.36
- 26 items: ×2
- 126 items or more: ×6
Where does the scan price differently?
A few findings don’t fit the tables above neatly, and the calculator on this page won’t match them to the dollar:
- Agentforce readiness and integration findings are graded “important” or “minor” rather than high, medium or low. Both are priced as medium: 4 hours.
- Agentforce readiness findings are carried at the unused-metadata rate, $1,200 a year.
- Integration and readiness findings list at most 25 affected items, so their scope multiplier tops out just under ×2.
- Rounding can differ by a dollar between the scan and this calculator.
A worked example
An org still has 12 Process Builder processes running. That is a retired feature, high severity, and an architect’s job.
- Scope: 1 + 0.04 × 11 = 1.44
- Fix cost: 12 hours × $175 × 1.44 = $3,024
- Yearly carrying cost: $2,100 × 1.44 = $3,024
- 3-year cost: $3,024 × 3 = $9,072
What the formula doesn’t know
- Your rates. If your team costs more or less than the US blended figures, the fix cost moves with it.
- Your business. A broken integration costs a retailer in November more than it costs most companies in a year. The model uses one figure per type of problem.
- How findings overlap. Two findings on the same object are often one piece of work. The model prices them separately.
- Stacked automation. Where several automations fire on the same object and event, the report flags the risk but leaves it out of the dollar totals.
Who decides the type and severity?
Two sources. Rule-based checks, such as integration health and org limits, set them by fixed rules. For the rest, the AI reviews the org’s metadata, writes the finding and labels its type and severity. The formula then prices every finding the same way. So the AI can affect a dollar figure by how it labels a problem, but it never writes the figure itself. The 0 to 100 health score is different: the AI sets that.
Common questions
- How much does Salesforce technical debt cost?
- It depends on what the debt is. In the VanguardOps cost model, one finding costs between $600 and $8,500 a year to leave alone, depending on its type, and between about $140 and $2,100 to fix, depending on severity and who does the work, before scaling for how many items it covers. A demo org with 15 findings came to an estimated $125,064 over three years and $17,021 to fix.
- Are these figures a quote?
- No. They are directional estimates built from US blended labor rates and conservative carrying costs. They are meant for deciding what to fix first and for budget conversations. Your actual cost depends on your org, your team and your rates.
- Does the AI decide the dollar amounts?
- No. The dollar amounts come from the fixed formula on this page. The AI reviews the org's metadata and writes findings, each with a type and a severity. The formula then prices each finding the same way every time.
- Why three years?
- Technical debt is a running cost, not a one-time one. Three years is long enough to show what leaving a problem alone adds up to and short enough to sit inside a normal planning cycle. The 3-year figure is the yearly carrying cost times three.
- What is a carrying cost?
- The yearly cost of leaving a problem in place: the extra maintenance it causes plus a risk-weighted allowance for the incident it could lead to. A security risk carries $8,500 a year in the model; a missing field description carries $600.
Questions about the method? Email chris@vanguardops.ai or read how scans handle your data on the Security page.