This MTF Research Report examines how often a frozen sample of 100 public management vacancies from eight 2026 Fortune 100 employers explicitly mentioned six families of data and analytics capability. It is a new question applied to an existing corpus captured on 10 August 2026. It is not a new September vacancy wave or a market-wide estimate.

Report number: MTF-RR-2026-09-11-01
Publication date: 11 September 2026
Author: MTF Institute Editorial Team
Reviewer: Igor Dmitriev
DOI: 10.5281/zenodo.22699797

Abstract

Seventy-six of the 100 vacancies contained at least one explicit data or analytics signal under the study rules. Analysis and insight appeared in 59 vacancies, measurement and reporting in 39, named analytical tools in 17, forecasting and modelling in 15, decision use in 11, and data quality or governance in five. Signal breadth was uneven: 24 vacancies contained none of the six families, 33 contained one, and 43 contained two or more. These results describe language in the frozen pages. They do not show the full capability required in practice or the prevalence of these requirements across the U.S. management labour market.

Research question

How did the frozen 100-vacancy corpus express data and analytics capability, and what evidence should a management learner build to demonstrate that capability responsibly?

Scope and sample

The evidence base contains 100 unique public U.S. management vacancies captured on 10 August 2026 from Amazon, Apple, Alphabet (Google), UnitedHealth Group, CVS Health, McKesson, Microsoft and JPMorgan Chase. Employer counts are unequal: Apple 18, Alphabet 15, Amazon 13, UnitedHealth 13, McKesson 12, Microsoft 12, JPMorgan Chase 12 and CVS Health five.

The sample was originally assembled for an earlier management-skills study. This report preserves all 100 records and asks a distinct question. Reusing the frozen corpus permits comparison across research waves without claiming that the vacancies remained open or unchanged on 11 September.

Method

We defined six signal families before reporting the counts:

Signal family Qualifying language
Measurement and reporting explicit metrics, KPI, dashboard, measurement or reporting responsibility
Analysis and insight explicit analysis, analytical reasoning, insight generation or interpretation
Forecasting and modelling explicit forecasting, modelling, scenario or predictive work
Data quality and governance explicit quality, integrity, stewardship, governance or controlled-data responsibility
Named analytical tools explicit analytical software or languages such as Excel, SQL, Tableau, Power BI, Python or R
Decision use explicit use of data or analysis to make, recommend or improve a decision

Rules-based retrieval identified candidate passages. Every positive bounded excerpt was then reviewed. Generic website boilerplate, search widgets, broad references to software and unqualified uses of words such as “model” or “insight” were excluded. A family was counted once per vacancy even when it appeared multiple times.

An absent code means qualifying wording was not found in the captured page. It does not prove that the capability was unnecessary in practice.

Results

Signal prevalence in the 100 vacancies

Signal family Vacancies Share Employers represented
Analysis and insight 59 59% 8
Measurement and reporting 39 39% 8
Named analytical tools 17 17% 4
Forecasting and modelling 15 15% 7
Decision use 11 11% 6
Data quality and governance 5 5% 4
Any of the six families 76 76% 8

Analysis language was more common than tool naming. The sample therefore does not support a “one software tool unlocks management work” conclusion. It more often described the ability to interpret or measure than a specific technical stack.

Breadth of explicit signals

Number of signal families in a vacancy Vacancies
0 24
1 33
2 25
3 12
4 4
5 1
6 1

Forty-three vacancies contained at least two signal families. Only two contained five or six. The pattern suggests that employers often combine selected analytical expectations rather than publish a uniform “data-driven manager” specification.

Role-family diversity check

Broad title-derived role family Sample At least one signal Share
People / General Management 25 19 76.0%
Technology / Product / Data 24 19 79.2%
Other management 15 13 86.7%
Operations / Supply 15 12 80.0%
Finance / Risk / Legal 12 7 58.3%
Commercial / Marketing 9 6 66.7%

These broad families come from job titles and exist only to inspect diversity. Unequal cells and employer mix make ranking inappropriate. The lower observed share for Finance / Risk / Legal should not be read as evidence that analytics matters less in those occupations.

What the result means

Three conclusions are supportable within the sample.

First, analytical capability is not confined to titles containing “data.” At least one signal appeared in every broad role family and every sampled employer.

Second, interpretation and measurement appeared more often than named tools. A manager needs to connect a calculation to a question, boundary and decision. Tool fluency without that connection is incomplete evidence.

Third, explicit data quality and decision-use language was comparatively rare. That is a disclosure result, not permission to ignore either discipline. A credible analyst or manager should still document source quality, uncertainty, decision ownership and possible harm.

Practical application: build a six-artifact evidence portfolio

Students can translate the six signal families into six reviewable artifacts:

Artifact What it demonstrates Minimum control
Metric tree measurement and reporting numerator, denominator, population, period and owner
Analysis memo analysis and insight question, source, method, finding and alternative explanation
Forecast or scenario forecasting and modelling drivers, range, sensitivity and update date
Data-quality log quality and governance source, missingness, duplicates, transformations and access boundary
Reproducible workbook or notebook named tools readable inputs, calculations, checks and handoff instructions
Decision note decision use options, recommendation, authority, risk and review date

A four-week practice sequence

Week 1: choose one real but non-confidential management question. Define the population, period, decision owner and what evidence would change the decision.

Week 2: create the metric tree and data-quality log before producing a chart. Record missing observations and transformations rather than hiding them.

Week 3: build a base analysis plus one sensitivity or scenario. Separate a descriptive finding from a forecast and from a recommendation.

Week 4: write the decision note. Ask another person to reproduce one result from the workbook or notebook and record the discrepancy.

This sequence turns “data-driven” into inspectable work. Use synthetic, public or properly authorized data. Do not place employer, customer or personal information in a portfolio without permission.

Implications for educators and employers

Educators should assess whether learners can frame a decision, preserve data lineage, calculate a result, challenge it and communicate limits. A software demonstration alone does not test transfer.

Employers can improve vacancy clarity by naming the decision, evidence and accountability expected. “Data-driven” is less useful than “build a weekly capacity forecast, explain variance and recommend an approved staffing action.” Tool requirements should reflect the work rather than an inherited list.

Limitations

This purposive sample covers eight employers with unequal counts. Public pages may omit requirements and reuse templates. The pages were captured on one date and may now differ or be unavailable. Title-derived role families are broad. Rules-based retrieval and human review remain interpretive. The six families overlap conceptually, and their counts do not measure proficiency, hiring weight, applicant success, employee experience or national prevalence. No causal inference is made.

Reproducibility, rights and integrity

The archive contains this searchable PDF and a supporting workbook with all 100 observation IDs, employer, title, capture date, public source URL, six binary codes, bounded qualifying excerpts, signal breadth and aggregate tables. The workbook also records method and limitations. Full vacancy bodies are not redistributed.

AI assisted source triage, coding implementation, drafting and QA. AI is not an author or evidence source. The named reviewer verified sample size, uniqueness, definitions, positive excerpts, arithmetic, public claims and archival files under the standing owner-approved research workflow.

Learning pathway

MTF Institute's Executive Certificate in Digital Transformation covers data-driven organization design, business analytics, technology strategy and responsible transformation. Learners can use the six-artifact portfolio above to turn the concepts into decision evidence. It is professional, non-degree education and does not guarantee employment or promotion.

Sources and files