Ad Tech TCO Calculator Methodology: Costs, Formulas, Sources

Total cost of ownership for ad tech is seven lines, not one: platform fees, data fees, cloud, licenses, staff, migration and exit. The calculator prices each line for a rented DSP or SSP, a white-label platform and an in-house stack, from your annual spend and peak bid rate, using dated public benchmarks that you can override.

This page exists so that anyone can check the arithmetic. Every default has a source and a date or is labelled an editable assumption, and the downloadable model contains the same formulas as the calculator.

The matrix is deliberately symmetrical: every model has every line, even where the default is zero, so that a vendor's "no fees" and a builder's "no licenses" are both visible as zeros rather than omissions.

Inputs: SPEND (annual media spend), Q_peak (peak QPS, for the in-house cloud line),
        and the defaults below (all editable).

# Platform fees
PLATFORM_rented     = SPEND × dsp_fee_pct
PLATFORM_whitelabel = max(SPEND × wl_fee_pct, wl_minimum)
PLATFORM_inhouse    = 0

# Data and add-on fees (identical across models by default)
DATA_model          = SPEND × data_fee_pct

# Supply-side fees (shown for completeness; paid by the seller side)
SSP_model           = SPEND × ssp_fee_pct

# Cloud or hosting (in-house only by default)
CLOUD_inhouse       = compute + egress from the bidder capacity planner at Q_peak
                      (compute = instances × price_per_hour × 730 × 12; egress = tiered(GB_OUT))

# Licences
LICENCE_inhouse     = licence_annual     (0 for open source)
LICENCE_whitelabel  = wl_licence_annual  (0 unless quoted)

# Staff
STAFF_model         = FTE_model × wage × loading
                      where FTE_rented = traders only, FTE_whitelabel = traders + ops,
                      FTE_inhouse = traders + ops + engineering

# Migration (year 1 only, amortised over the horizon in the per-year view)
MIGRATION_model     = person_months_model × wage × loading / 12 + parallel_run_months × PLATFORM_old / 12

# Exit (provision; shown separately, not added to the annual run rate unless the user ticks it)
EXIT_model          = export_person_months × wage × loading / 12 + notice_months × PLATFORM_model / 12

# Totals
TCO_model_year      = PLATFORM + DATA + SSP + CLOUD + LICENCE + STAFF + MIGRATION / horizon_years
TCO_model_pct       = TCO_model_year / SPEND
TCO_model_3yr       = 3 × (PLATFORM + DATA + SSP + CLOUD + LICENCE + STAFF) + MIGRATION (+ EXIT if ticked)
BREAKEVEN_SPEND     = the SPEND at which TCO_inhouse_year = TCO_rented_year (solved numerically)

Use the interactive model above or the downloadable workbook to compare the operating models. The calculation includes the parallel-run cost entered for migration. Hosting is an editable annual budget; use the bidder capacity planner to estimate it from measured load.

The original draft’s static worked-example totals did not include all parallel-run costs described in its formulas. Those totals and the derived break-even claim are withheld. Compare the calculated results with your own contract, staffing plan and transition schedule. Data and supply-side fees remain visible in each model so the comparison uses the same cost boundary.

• The DSP fee default is derived from one company's filings (The Trade Desk) and is a ratio of revenue to gross spend, where gross spend includes the fee. It is not a published take rate, and other DSPs price differently; Criteo, for instance, describes "a negotiated SSP fee and sometimes a technology fee" for retailers and "a negotiated DSP fee" for brands (Criteo 10-K, FY2025). Use your own contract rate if you have one.
• The ISBA/PwC percentages are from a 2020 UK study of 15 advertisers; the 2022 follow-up is cited only for the unattributable share.
• The ANA figures mix three metrics: "reaching the consumer" (2023 study), "reaching publishers" (quarterly benchmark) and TrueAdSpend (2026). They are presented as cross-checks on the transaction-cost line, not as inputs.
• White-label fee, license, FTE counts, migration effort and exit provisions are editable assumptions; no vendor publishes them. The calculator labels them in the interface.
• The wage series is for software developers and is applied to traders and operations staff as well, which overstates those roles in some markets and understates them in others. EU users should replace it with a national series.
• Cloud prices are US-region list prices; reserved or committed pricing is lower, and EU regions are typically higher.
• Peak QPS per $1M of spend is a convenience ratio with no public source; replace it with a measurement.
• The model excludes media cost itself, agency fees, creative and the opportunity cost of data leaving the company, which managed vs on-prem: the right model discusses qualitatively.

The XLSX contains an Inputs sheet and a Model sheet. Inputs carry source links or editable-assumption labels. Results use native formulas for recurring costs, migration, an optional exit provision and the rented-versus-owned break-even calculation. Hosting is an editable annual budget informed by the bidder capacity planner.

Retain the dated workbook with your evaluation. You can import it into Google Sheets. This release does not provide a public archive route or a native Excel What-If Data Table.

The model is dated October 17, 2026. Shared links carry the model version and input values; a version mismatch produces a notice. Download filenames include the model date. Keep that file with the assumptions and commercial evidence used in your decision.

Review sources and defaults at the next scheduled review. Record formula changes, source changes and changed assumptions in the page changelog. Historical source drafts and model definitions remain in repository history; a public version archive is not provided in this release.

What percentage of ad spend goes to ad tech fees? The ANA found transaction costs of 29% of each DSP dollar in its 2023 study and 29.8% in its 2024 benchmark (ANA, 17 June 2024; ExchangeWire, 16 September 2026). ISBA and PwC found 7% agency, 8% DSP, 10% demand-side technology, 8% SSP and 1% supply-side technology in 2020 (Marketing Dive, 7 May 2020). Fees are the part of the gap that does not go to media productivity losses.

How much does a DSP charge? The Trade Desk's revenue was 21.6% of gross spend in 2025 by our calculation from its 10-K (The Trade Desk, FY2025); gross spend includes data and services, so the pure media fee is lower. Contract rates vary with spend commitments and are not published. The calculator uses 21.6% as a ceiling and 8% (ISBA 2020) as a floor preset.

What does it cost to run a DSP in-house? Use the calculator or workbook with your own license, hosting, staffing, integration and migration inputs. The model includes recurring costs and transition assumptions within a common cost boundary. Its defaults are editable planning assumptions, not market quotes or a performance benchmark.

Is white-label cheaper than renting a DSP? The result depends on contracted fees, staffing and operating scope. Compare the calculated totals under the same cost boundary, including migration and parallel operation. Change the assumptions before drawing a conclusion; there is no universal ranking across providers or budgets.

How do you calculate ad tech TCO? Price seven categories for each model (platform fees, data fees, cloud, licenses, staff, migration, exit) from annual spend and peak bid rate, sum them per year, express the total as a share of spend, and solve for the spend level at which two models cost the same. The formulas above are the complete set.

Use the in-house programmatic guide to compare the related operating models, definitions and evidence before making this decision.

1. The Trade Desk, Form 10-K FY2025. https://www.sec.gov/Archives/edgar/data/1671933/000167193326000014/ttd-20251231.htm
2. Marketing Dive, "15% of programmatic ad spend disappears in the supply chain, study says" (7 May 2020). https://www.marketingdive.com/news/15-of-programmatic-ad-spend-disappears-in-the-supply-chain-study-says/577525/
3. ExchangeWire, "The Ad Tech Toll Road" (16 September 2026). https://www.exchangewire.com/blog/2026/09/16/the-ad-tech-toll-road/
4. ANA, "ANA, TAG TrustNet Launch Programmatic Transparency Benchmark" (17 June 2024). https://www.ana.net/content/show/id/pr-2024-06-tag
5. ANA, Q3 2025 Programmatic Transparency Benchmark release (5 November 2025). https://www.ana.net/content/show/id/pr-2025-11-transparency
6. Fiducia, "ANA Programmatic Transparency Benchmark Q2 2026: working media efficiency reaches a record 45.1%" (6 August 2026). https://www.fiducia.eco/post/ana-programmatic-transparency-benchmark-q2-2026-working-media-efficiency-reaches-a-record-45
7. Criteo S.A., Form 10-K FY2025. https://www.sec.gov/Archives/edgar/data/1576427/000157642726000014/crto-20251231.htm
8. Vantage, EC2 c7g.xlarge (updated 17 October 2026). https://instances.vantage.sh/aws/ec2/c7g.xlarge
9. CloudZero, "AWS egress costs" (updated 1 May 2026). https://www.cloudzero.com/blog/aws-egress-costs
10. AWS, EC2 On-Demand Pricing (accessed 17 October 2026). https://aws.amazon.com/ec2/pricing/on-demand/
11. Google Cloud, VPC pricing announcement (effective 1 February 2024). https://cloud.google.com/vpc/pricing-announce
12. US Bureau of Labor Statistics, Software Developers (May 2025; modified 27 August 2026). https://www.bls.gov/ooh/computer-and-information-technology/software-developers.htm
13. AWS, Guidance for Deploying a Prebid Server on AWS (accessed 17 October 2026). https://docs.aws.amazon.com/solutions/latest/prebid-server-deployment-on-aws/configurations-for-elastic.html

• 18 October 2026: removed stale static FAQ totals and cost rankings. Use the implemented model’s results and disclosed assumptions.

• 18 October 2026: aligned the title and heading and applied recorded source-URL corrections where needed. Factual research cutoff remains 17 October 2026.

• 17 October 2026: methodology published. The executable model and workbooks use a dated implementation version; inputs and formulas are available for inspection. Hosting is an editable annual budget, informed by measured capacity rather than a hidden automatic link.

• 1.x, before 17 October 2026. Original /tco calculator without a published methodology; results generated under 1.x are flagged in shared links.