About CodeSmellCost.com
CodeSmellCost.com is the cost-attached translation of Martin Fowler's 22-smell catalog into language engineering leaders can take to a CFO. Every smell on the site gets a dollar range, a detection recipe, and a refactoring pattern. The calculator behind the home page is the full math, open and reproducible, with CSV export.
The site is published by Digital Signet. It is independent, vendor-neutral, and not a lead-generation funnel for a tooling vendor or consultancy.
Fowler's Refactoring (Addison-Wesley, 1999; 2nd ed. 2018, with the JavaScript examples) defines the vocabulary every engineering team uses for code quality. The book describes 22 smells, gives each one a refactoring pattern, and explains the mechanism. What it does not do is attach a dollar figure. The book pre-dates the CFO conversation.
The vendor glossaries (SonarSource's SonarQube docs, JetBrains' refactoring catalog, refactoring.guru, CodeAnt) describe smells in plain prose but never quote a number. The peer-reviewed empirical literature (Tufano et al. ICSE 2015, Bird et al. FSE 2011, Khomh et al. 2012, Palomba et al. 2018) reports correlations and effect sizes but is not aimed at engineering leaders making spending decisions. The Stripe Developer Coefficient Report (2018) gives the headline macro number ($85 billion per year in lost developer productivity globally) but does not break it down by smell.
The gap is the translation layer. CodeSmellCost.com is that layer: Fowler's catalog, dollar-attached, cited to peer-reviewed research, organised for the executive committee.
Oliver Wakefield-Smith runs Digital Signet, an independent consultancy that builds AI systems and operates a portfolio of editorial reference sites in the engineering, FinOps, and tech-debt space. CodeSmellCost.com is one of those sites. Oliver writes the analysis, runs the cost model, and signs the corrections.
Sister sites in the same editorial family:
- featurebloat.com - The product-management mirror: same argument, from the PM's side of the table.
- techdebtcalculator.com - Calculator-first sister site on the broader 'technical debt' surface.
- incidentcost.com - Production incident cost modelling. Cited in our refactoring-ROI risk math.
- codedebtcost.com - Synonym-adjacent peer on the technical-debt head term.
- reworkcost.com - The broader rework-economics view (engineering plus manufacturing plus construction).
CodeSmellCost.com is independent. It is not a reseller of SonarQube, CodeScene, Sourcegraph, GitHub, or any other tool reviewed on the site. It is not a consulting lead-generation funnel disguised as content; the consulting CTA on the home page is honest, but the site stands on its own as a reference whether anyone hires Digital Signet or not.
The editorial position rests on three sources of authority. First, Fowler's Refactoring as the canonical vocabulary. Second, the peer-reviewed empirical software engineering literature (Khomh 2012, Palomba 2018, Bird 2011) for the link between smells and fault-proneness. Third, named industry research with public methodology (Stripe Developer Coefficient; Tornhill and Borg's Code Red study of 39 proprietary codebases) for the productivity-cost translation.
The site deliberately does not chase vendor-pricing SEO. There is no /sonarqube-pricing page, no /jetbrains-pricing page. The /tools page compares the eight tools that matter for smell detection in 2026 but does not rank them with affiliate-driven scores. Where pricing changes (the tools page is monthly verified), the change ships with the next refresh pass.
Source pattern
Research findings cite the paper or report they come from: Fowler's Refactoring 2nd edition, Robert C. Martin's Clean Code, peer-reviewed empirical studies (Tufano ICSE 2015, Bird FSE 2011, Khomh EMSE 2012, Palomba EMSE 2018), or industry research (Stripe Developer Coefficient; Tornhill and Borg, Code Red, 2022). Dollar bands are the output of our own cost model, and the cost-model page shows its inputs.
Dated changes
Each page's Updated date comes from the site's version history at build, so it moves only when that page changes. The Data verified badge is separate: it records the last full re-check of numbers, cited sources and tool pricing, and moves only when that re-check is done.
Conservative cost ranges
Dollar bands on the smell pages assume a team of 8 engineers at fully-loaded cost and are model estimates. Ranges, never point estimates. An engineering leader presenting a range is more credible than one presenting a single number.
Cost ranges per smell are the output of one model of our own, with three components: developer time lost to the smell, extra defects (the direction from Khomh 2012 and Palomba 2018 on fault-proneness), and, for God Class and Inappropriate Intimacy, slower onboarding. Each smell takes the inputs of its severity tier, at a team of 8 and $180K fully-loaded cost per engineer. The inputs are our assumptions, listed on the cost model page. The team-level calculator is a separate tool with four components: PR review drag (SmartBear's study of a Cisco team on review-session limits), incident cost, development-time drag (Tornhill and Borg 2022 on issue resolution time) and onboarding drag.
For the full source list, the per-component formulas and the limitations: see the methodology page. Or run the model with your own numbers on the calculator.
Found a number that looks wrong, a citation that does not check out, or a smell missing from the catalog? Email [email protected]. Corrections turn around in 5 business days. The corrected version ships with the next monthly refresh pass and rolls the LAST_VERIFIED_DATE forward.
For the two-week code-debt audit consulting engagement (we run the model with your real incident history, PR review data, and onboarding records, and deliver the memo your leadership team will sign), the same address. Honest report whether the recommendation is “refactor now” or “this is defensible.”
Disclosures
CodeSmellCost.com is an independent editorial site published by Digital Signet.
This site is not affiliated with Martin Fowler, Pearson, Addison-Wesley, O'Reilly, SonarSource, CodeScene, Sourcegraph, GitHub, JetBrains, or any other publisher, author, or vendor mentioned. Quotes are attributed to their original sources. Trademarks remain the property of their respective owners.