Updated November 2026
Methodology
How TechStackCost.com sources per-employee SaaS spend benchmarks, per-engineer ranges, hidden-cost categories, and cloud reference rates. Every number on the site should be re-derivable from the cited public source using the formulas shown here.
Sources
Each cost reference on the site cites the originating source: research publication for per-employee benchmarks and waste rates, vendor public pricing page for tool list prices, FinOps Foundation for category framing, and BLS for the labour reference when converting per-engineer tooling cost into total-cost-of-engineer math. Aggregator restatements are not used as primary citations.
| Source | Refresh cadence | What we take from it |
|---|---|---|
| Zylo SaaS Management Index | Annual + quarterly updates | Per-employee SaaS spend benchmark (the $9,455/employee median from the 2026 index, on the homepage and SaaS-spend-benchmarks page), application count (305 per company), license utilization (36 percent of licenses unused), and shadow-IT prevalence (business units control 81 percent of SaaS spend, 21 new applications a month). Methodology is published with each report. |
| Deloitte technology-budget research | Periodic | IT-spend-as-percent-of-revenue benchmark: 5.49 percent of revenue in 2022 from the 2023 Global Technology Leadership Study, with Deloitte's earlier 2016, 2018 and 2020 points, and the sector spread from the 2016-2017 Global CIO Survey. |
| Stack Overflow Developer Survey | Annual | Developer tool adoption rates by language, framework, and category. Used to weight the developer-tooling category coverage so the most-adopted tools (VS Code, JetBrains, GitHub, Slack, Jira) lead each table rather than the most-expensive. |
| JetBrains State of Developer Ecosystem | Annual | Complements Stack Overflow on IDE share, language usage trends, and adoption of AI code assistants. Cross-checked against survey-of-developers reports for category framing on the developer-tools page. |
| FinOps Foundation publications | Continuous | FinOps lifecycle framing (visibility, optimisation, forecasting, continuous improvement) and cloud cost governance principles. Authoritative for FinOps category framing. |
| Vantage Cloud Cost Reports | Quarterly | Cross-vendor cloud rate aggregation for AWS, GCP, Azure spot and on-demand pricing trends. Used as a cross-check on vendor public pricing pages on the cloud-costs page; not the primary source. |
| BLS Occupational Employment and Wage Statistics | Annual | Software engineer base wage anchors (15-1252 Software Developers, 15-1253 Software Quality Assurance, 15-1244 Network and Computer Systems Administrators) used as the labour reference when converting per-engineer tooling cost into total-cost-of-engineer framing. We do not republish wages on-page; we cite BLS as the methodology anchor. |
| AWS public pricing pages | Monthly | EC2 on-demand instance rates, EBS storage, S3 storage, data transfer (egress) pricing. US-East-1 list price is the reference; spot, savings plans, and reserved capacity are out of scope to keep cross-vendor comparison apples-to-apples. |
| Google Cloud public pricing pages | Monthly | Compute Engine on-demand rates, Cloud Storage, BigQuery on-demand pricing, network egress. Used on the cloud-costs and data-platform-cost pages. |
| Microsoft Azure public pricing pages | Monthly | Virtual Machines on-demand rates, Blob Storage, Azure SQL, bandwidth pricing. Used on the cloud-costs page. |
| Datadog, Snowflake, GitHub, Linear, Notion, Figma, Slack public pricing pages | Monthly | Per-tool published list prices and per-seat tiers. Updates are pulled from each vendor's own pricing page; aggregator restatements are not used. Specific brand-pricing deep-dives are deliberately not built as standalone pages on this site to avoid the trust ceiling that comes with brand-named cost pages. |
In scope
- ●Vendor-published list prices on each vendor's own public pricing page.
- ●Per-employee SaaS spend benchmarks from Zylo's SaaS Management Index, with the citation visible on every page that uses the number.
- ●Per-engineer stage bands derived from the same research applied to engineering-organisation headcount.
- ●Hidden-cost categories (training, maintenance, lock-in, technical debt) explained with worked examples at a stated loaded engineering cost, so a team can rerun the arithmetic with its own numbers.
- ●The SaaS waste anchor (36 percent of licenses unused, Zylo 2026), cited inline.
- ●Cloud reference rates at US-East-1 list price; deployment-model TCO comparisons using the same on-demand baseline.
- ●FinOps framing (lifecycle phases, governance principles) cross-referenced against the FinOps Foundation.
Out of scope
- ●Enterprise-negotiated pricing. Volume commitments, contract minimums, custom MSAs, and EDP-class discounts are deliberately excluded.
- ●Spot, savings plan, and reserved-capacity discounts on AWS, GCP, Azure. Cloud rate comparisons use on-demand list price to keep cross-vendor math consistent.
- ●Regional surcharges beyond US-East-1 / generic SaaS. AWS pricing varies 10-20 percent by region; the comparison uses US-East-1 unless flagged.
- ●Brand-named pricing pages (a dedicated /aws-pricing, /datadog-pricing, or /snowflake-pricing page). These attract brand-trust-ceiling competition the build cannot win and risk Lanham Act exposure that has bitten sister sites in the security cluster.
- ●A published benchmark for the hidden-cost multiplier. We have not found a primary source that measures one, so the hidden-costs page shows the categories and the arithmetic, and teams should substitute their own loaded engineering cost and maintenance share.
- ●Lead-generation forms, gated content, or affiliate links. The site is a reference, not a funnel.
Calculation framework
Sourced from Zylo SaaS Management Index ($9,455/employee/yr median in the 2026 index). The 2025 index reported a $4,830 average, a different measure, so the two are not shown as a trend. We have not found a free primary source for per-employee spend by industry or company size, so the site does not publish one.
Stage bands derived from the typical tool stack at each company stage (seed through enterprise) applied to engineering-organisation headcount. Each band has a worked example showing component breakdown: cloud allocation per engineer, IDE and CI/CD per seat, monitoring share, data platform allocation, security tooling, collaboration tools. The example on the cost-per-engineer page shows the math so the band is verifiable.
Costs that never appear on a vendor invoice: training and onboarding, maintenance burden, vendor lock-in exit cost, and technical-debt drag. None has a published industry benchmark we can cite, so the honest method is arithmetic over your own inputs: engineering hours spent on each category multiplied by your loaded engineering cost, added to visible spend.
The SaaS waste anchor is Zylo's 2026 finding that organizations leave 36 percent of their SaaS licenses unused. It measures unused licenses, not money a given audit will recover; what an audit reclaims depends on contract terms and renewal dates.
AWS EC2 on-demand US-East-1 is the reference instance pricing on the cloud-costs page. Google Cloud Compute Engine on-demand US-Central-1 and Azure Virtual Machines on-demand East-US-2 are presented at matched instance sizes. Egress pricing is shown for each cloud at the standard internet-egress rate, with regional and inter-region rates flagged where they differ materially. Spot, savings plans, and committed-use discounts are explicitly out of scope.
The seed, Series A, Series B, growth, and enterprise stage bands on the by-company-size page use a typical engineering-team size at each stage (1-5, 5-20, 20-80, 80-300, 300+ engineers), apply category-specific tool list prices at that headcount, and roll up to a monthly total plus per-engineer cost. The example in each card shows the actual line items so the band is verifiable.
Refresh cadence
Vendor list prices and benchmark citations are re-checked against their source when a page is reviewed. The “Prices verified” label records the last full price check and moves only when that check is done again. Each page's “Updated” date and the dateModified field in its Article JSON-LD are set at build time from the last substantive change to that page in the site's source history, so neither can be moved by hand.
Out-of-cycle refreshes trigger on:
- ●Annual benchmark publication from a primary source (Zylo, Deloitte) that shifts the anchor number.
- ●Vendor public pricing-page change that moves a list price by more than 10 percent on a referenced tool.
- ●Major cloud rate change at AWS, GCP, or Azure on a reference instance type or egress tier.
- ●New billing model from a referenced vendor (per-seat to usage-based migration, or vice versa) that changes how the cost is calculated.
- ●Substantive correction reported via the corrections inbox that updates a published figure.
Refreshes that move a per-employee benchmark or vendor list price by less than 5 percent batch into the next monthly pass. Refreshes that introduce a new billing model, shift an anchor benchmark by more than 10 percent, or correct a substantive error ship as soon as the change is confirmed against the originating source.
Limitations
Calculator and benchmark outputs are estimates. Production tech stack cost depends on enterprise agreements, EDP/MSA committed-use programs, regional surcharges, reserved-capacity commitments, and per-workspace tier negotiations that are out of scope here. Always verify with the vendor before purchasing or making a budget commitment.
Hidden costs have no published benchmark we can cite. Treat any multiplier on visible spend as a planning assumption, and replace it with your team's own loaded engineering rate and measured maintenance burden.
Per-employee SaaS benchmarks from different research firms use different methods (invoiced spend versus estimated spend including shadow IT; median versus average; different customer samples), so figures from different reports or different editions are not directly comparable. The site names the report, edition and measure next to each figure.
Corrections process
Spotted a stale price, a missing tier, a research source we have not cited, or a methodology divergence we have not flagged? Email [email protected] with the page URL and the source you would like cited. Substantive corrections (benchmark shifts, vendor pricing changes, new billing models) are typically actioned within five business days. Non-substantive corrections (typos, link rot, structural edits) batch into the next monthly pass.
See also the about page for the site's editorial position, disclosures, and full coverage map.