Methodology
How the DCF Evidence Score works
The DCF Evidence Score (0–100) measures how much verifiable public evidence a firm has for its data and AI work: official partner-directory listings, dedicated service pages, public case studies, industry pages, and published rates or reviews. Every point traces to a public URL. It does not measure delivery quality, client satisfaction, or price, and a lower score can simply mean a firm publishes less.
Edited by Peter Korpak·Evidence checked ·How we score
What goes into the score?
Five components, each capped. The table is generated from the same constants the scoring code uses, so it always matches the published scores.
| Component | Max | How points are earned | Evidence source |
|---|---|---|---|
| Platform credentials | 30 | Snowflake: the best tier badge on the firm's Snowflake partner-directory page (Elite 12, Premier 9, Select 6, listed without a badge 4). Databricks: listed as a Consulting & SI partner in Databricks' public partner directory (8; the directory shows no tier for consulting partners, so none is scored). AWS, Microsoft, and Google Cloud: a documented partnership claim, sourced to the provider's directory, the firm's site, or a press release (AWS Premier 4, AWS Advanced 3, Google Cloud Premier 4, a Microsoft Solutions Partner designation or specialization 3, any other documented partnership 2). Capped at 30. | Snowflake and Databricks partner directories; AWS, Microsoft, and Google Cloud partner pages or firm announcements |
| Capability pages | 25 | 3 points for each of the 9 tracked capabilities that has a dedicated service page on the firm's own website. A single overview page that lists several services earns credit for at most 2 of them. Capped at 25. | The firm's own website |
| Public case studies | 25 | Up to 3 public data or AI case studies count, 6 points each, plus 2 for each that names the client and 1 if at least one is dated 2024 or later. Capped at 25. | The firm's own website, or a platform vendor's customer-story page |
| Industry pages | 10 | 2.5 points for each dedicated industry practice page, up to 4 pages. | The firm's own website |
| Rate and review transparency | 10 | Clutch reviews (20 or more: 5; 5–19: 3; 1–4: 1) plus public rates: data-role rates on a GSA schedule or rates the firm publishes itself (5), or, failing that, an hourly band on the firm's Clutch profile (2). Capped at 10. | GSA CALC+ ceiling rates; Clutch profiles where accessible; the firm’s own pricing pages |
| Total | 100 | Firms are ranked by total score; ties are ordered alphabetically. | |
Worked example: current #1
- Platform credentials: 30/30 — Snowflake Elite (12); Databricks consulting partner (8); AWS partnership (4); Microsoft partnership (3); Google Cloud partnership (4)
- Capability pages: 21/25 — 7 of 9 tracked capabilities have a dedicated page on the firm's own site (3 each)
- Public case studies: 22/25 — 3 public data/AI case studies counted (max 3), 2 naming the client
- Industry pages: 10/10 — 4 industry practice pages counted (max 4)
- Rate and review transparency: 7/10 — 26 Clutch reviews (5); Hourly rate band on Clutch (2)
What counts as a dedicated capability page?
A page on the firm's own site that explicitly offers the capability. A generic "digital transformation" page does not count. The nine tracked capabilities and the wording that qualifies:
- Data strategy (directory page): data strategy, data & analytics strategy, data/AI roadmap, or data maturity assessment.
- Data architecture (directory page): data architecture, modern data platform design, data mesh, data fabric, or lakehouse architecture.
- Data engineering (directory page): data engineering, pipelines, data integration, ETL/ELT, or streaming.
- Analytics & BI (directory page): analytics, business intelligence, dashboards, or reporting.
- Data science & ML (directory page): data science, machine learning, MLOps, or predictive analytics.
- Generative AI (directory page): generative AI, LLM applications, AI agents, or RAG implementation.
- Data governance (directory page): data governance, data quality, master data management, data management, or data privacy.
- Platform migration (directory page): data platform migration or modernization (warehouse to cloud, Snowflake, Databricks, or Fabric migrations).
- Data products: data products, data-as-a-product, or a data mesh operating model.
Where does the evidence come from?
- Snowflake: Snowflake Partner Network directory (https://www.snowflake.com/en/why-snowflake/partners/all-partners/), tier = the partner-tier tag the directory shows for each listing; 1,535 listings read from the directory's public filter feed.
- Databricks: Databricks partner directory data feed https://www.databricks.com/en-partners-assets/data/partner/all/en.json behind https://www.databricks.com/partners/partner-directory (C&SI = Consulting & System Integrator; public feed exposes no Elite/Select tier for C&SI partners, only a 'Platinum' flag on 3 partners).
- Firm websites: service pages, case studies, industry pages, headquarters, founding year, and headcount as each firm states them, with the URL recorded for every fact.
- Rates: GSA CALC+ ceiling rates (U.S. General Services Administration, buy.gsa.gov/pricing) and, where accessible, Clutch profile rate bands. See rates and pricing.
The current evidence was checked on October 23, 2026. Of the 50 firms tracked, 31 appear in Snowflake's partner directory (19 with the Elite badge), 36 are listed as Databricks consulting partners, and 7 hold GSA schedule rates for data roles.
Which firms are not ranked?
Platform vendors' own services teams (Databricks Forward Deployed Engineering and Palantir) appear in the directory for comparison but are not ranked, because their partner credentials and service pages are not comparable with independent consultancies'.
What are the limits of this method?
- It rewards firms that publish evidence. A capable firm with a thin website scores lower than its work may deserve.
- Absence from a public directory feed is not proof that a firm has no relationship with that vendor.
- Case-study outcomes are reported by the firms themselves; DCF Research has not audited them.
- GSA ceiling rates are maximum prices for U.S. federal buyers, not commercial quotes.
- The data is a dated snapshot. Partner tiers and service pages change between checks.
Can a firm pay to change its score?
No. Scores are computed only from public evidence. DCF Research may sell clearly labeled Featured placements; a featured firm keeps exactly the score and rank its evidence earns, and featured slots are always marked. No firm pays for placement today, and DCF Research receives no referral fees. Read the editorial policy.
How do I report an error or submit evidence?
Email hello@dataconsultingfirms.com with a link to a public source. We correct verified errors and re-score the firm; submissions without a public source are not used. Partner directories and firm pages are re-checked at least quarterly; the next scheduled check is January 2027.
Changelog
- October 23, 2026: replaced the earlier hand-assigned 0–10 scores with the evidence-based score above; verified every firm against official partner directories; removed claims that could not be sourced; updated names to current brands (BCG X, which absorbed BCG Gamma; Databricks Forward Deployed Engineering, formerly Databricks Professional Services); added phData, Avanade, Xebia, and Lovelytics; removed Cloudreach, SADA, and GetInData, whose brands have been absorbed by their acquirers (GetInData is now part of Xebia), and DevsData, which operates mainly as a recruitment and staffing firm; limited capability credit to two capabilities per page, so one overview page listing many services no longer counts as several dedicated practice pages.