Edition history and the decisions made before scoring

This page records what was published and when, and the rules settled before any provider was scored, so that they could not be bent to produce a wanted result afterwards. Factual corrections are listed separately on the corrections page, and no claim has been corrected since publication.

Edition history

Global Edition v1.0 (2026-08-28)

This entry records version 1.0 of the Global Edition.

The Global Edition is published at /, covering 29 providers admitted under two published entry rules. The European Edition is published at /europe/, covering eight European-headquartered providers in more depth.

  • Twenty-nine providers admitted under two published entry rules: approximately 10,000+ delivery FTEs across three or more continents, or a Leader rating earned in any scoped edition of this publication at any size.
  • Tier thresholds published: Leader ≥ 70.0, Strong Performer ≥ 60.0, Specialist ≥ 45.0, Emerging below 45.0. They reproduce the tier each of the eight European providers is published under, and a build guard checks it on every deploy.
  • Calibration addendum published at /methodology/calibration/, stating how the anchors’ requirement that a capability be evidenced is read, with the worked example each rule rests on.
  • Five providers appear in both editions and hold identical scores in each: Teleperformance, Simply Contact, Foundever, Transcom and Konecta. One data sheet serves both; a build guard fails the deploy if a score ever differs between the two pages.
  • The Global Edition holds / as the primary report and the European Edition sits at /europe/. Both carry a banner pointing at the other, so a reader who lands on either knows the other exists.
  • /data/ spans 32 provider data sheets. Conectys, Mplus and Armatis are European-only and their sheets say so.
  • Two scoring tables are published: scores-2026-global.csv and scores-2026.csv.
  • A plain-text summary for AI assistants (llms.txt) describes both editions, both fields, the entry rules, the published tier bands and the calibration addendum.
  • A4.1 employee-side signal is scored under a published source-selection rule: where more than one employer profile covers a provider, the one with the largest review volume governs; two sources diverge materially at 0.5 or more, being the width of one rating band; and a material divergence costs one interpolated anchor step. Every A4.1 claim names its platform, profile, rating and review volume. The rule and its derivation are published in the scoring anchors appendix and restated in the calibration addendum.
  • Evidence ledgers published for all 29 providers: 331 claims in total, each carrying its grade and, where one exists, a link to the source it came from. Every data sheet counts its own claims by grade, and the rankings page carries the same counts for the whole field.

2026 Edition published (2026-08-28)

This entry records version 1.0 of the European Edition.

Publication of the 2026 Edition: eight European-headquartered providers scored against the Birchy Provider Evaluation Framework, with the full scoring anchors, the screening record and the corrections process published alongside.

  • Scoring anchors appendix published in full under CC BY 4.0 (v1.0).
  • Screening record published: 8 evaluated, exclusions listed with reasons, next-edition candidates named.
  • Corrections process opened: any evaluated provider may submit corrections at any time, with no registration or participation required.
  • Evidence grading (E1–E5) applied to every material claim.
  • A4.1 employee-side signal is scored under a published source-selection rule: where more than one employer profile covers a provider, the one with the largest review volume governs; two sources diverge materially at 0.5 or more, being the width of one rating band; and a material divergence costs one interpolated anchor step. Every A4.1 claim names its platform, profile, rating and review volume. The rule and its derivation are published in the scoring anchors appendix and restated in the calibration addendum.

Findings cluster v1.1 (2026-09-23)

This entry records version 1.1 of the Global Edition.

Six new standalone finding pages and two evidence-ranked guide pages published under /findings/ and /guides/, plus new sections added to the existing pricing and AI-oversight findings.

  • The 18 diligence questions from the flagship comparison expanded into their own page at /findings/rfp-questions/, each tied to the scoring anchor it closes a gap in.
  • Marketed vs. verified language counts published for all 32 providers at /findings/language-claims/.
  • Registered HQ vs. stated delivery footprint published for the European Edition’s 8 providers at /findings/nearshore-europe/.
  • Employee-count claims for all 29 Global Edition providers split by evidence quality (filed, corroborated, self-stated) at /findings/largest-providers/.
  • Closest-scoring-alternative pages published for Teleperformance and Concentrix, each with the five nearest providers on the same twelve criteria.
  • /guides/top-bpo-companies/ and /guides/top-call-center-outsourcing-companies/ publish the same 29-provider board read for two high-intent search queries.
  • The pricing finding extended with generic cost-driver and pricing-model context, cross-linked to the provider-specific evidence already published.

Site redesign v1.2 — 2026-11-13

This entry records version 1.2 of the publication.

The publication is redesigned, and two tools are published: /compare/ reads two or three providers side by side on the twelve criteria, and /shortlist/ applies the 14 published situation thresholds to all 32 providers. Five more closest-alternative pages and this page are published alongside. No score, rank or tier moved.

  • Comparison tool published at /compare/: two or three providers side by side on their twelve criterion scores, their delivery-proof and capability-fit split, and their largest gap, read from the same data sheets.
  • Shortlist tool published at /shortlist/: the 14 situations and the published threshold each one reads, applied to all 32 providers — who meets it, who does not, and on which sub-criterion.
  • Closest-scoring-alternative pages published for TELUS Digital, Simply Contact, Sutherland, Foundever and HGS, joining the Teleperformance and Concentrix pages from v1.1 — seven in all.
  • This changelog is a page again, at /changelog/, carrying the edition history and the pre-scoring decisions together.
  • 743 copy edits applied across the publication. They change wording, not findings: no score, axis total, rank or tier changed, and build guards check every published score against the scoring tables on each deploy.
  • Simply Contact’s data sheet stated 23 evidence claims while its own ledger held 24. The sheet now states 24, and the build refuses to publish any sheet whose stated count disagrees with the ledger beneath it.
  • The Global Edition scoring table publishes the source and the evidence grade behind every headcount figure again, alongside the entry basis each provider was admitted under.
  • The modification date each page reports to search engines, and the sitemap’s lastmod, are both read from this log rather than from the clock the site was built on.

Pre-scoring decisions

A1 enterprise-equivalence rule

Decided before scoring.

Client-reference validation within a major analyst evaluation (Everest PEAK, ISG, NelsonHall; all of which conduct reference checks as part of assessment) anchors A1.1 and A1.2 at 3. Reviews remain required for 4–5. Adopted before any full scoring and logged, because changing it after scores existed would look like tuning.

Lowest corroborated band wins

Decided before scoring.

Where a provider’s headcount or scale conflicts across sources, the lowest corroborated band is used and the conflict is logged.

Marketed language counts are company-stated

Decided before scoring.

A language counts as verified native delivery only with corroborating hiring data or client evidence naming the language. Marketed totals are claims and are graded E3.

Certificate version-currency rule

Decided before scoring.

Certificates against superseded standard versions past their transition deadline score as expired.

A3.3 removed as a standalone sub-criterion

Calibration decision 1.

COPC/ISO 18295 becomes supporting evidence within B3.1; operations certification is rare below enterprise scale in a mid-market-scoped publication. A3 points redistributed: A3.1→6, A3.2→4, A3.4→2.

A3.1 five-anchor wording

Calibration decision 2.

Amended to “ISO 27001 plus a second verified attestation (SOC 2 Type II or ISO 27701)”.