Stage 7: ESG data management, controls and tools · topic 26 of 47

Running an ESG data cycle with controls

An ESG data cycle gets numbers from many sites on time, checks them and keeps the evidence. You will build a data dictionary, write validation checks and keep a sign-off and restatement trail.

  • Core
  • Intermediate
  • about 18 hours

Part of the ESG analyst roadmap: from the basics to mid level

Written by the GreenHired editorial teamSources last checked 20 November 2026How we research roadmaps

Assurance providers will ask two things about a figure you report: where did this number come from, and who approved it? Design the cycle backwards from those questions. Build the dictionary first, then templates and owners, then the checks, then the evidence folder. Add automation only after a manual cycle works, so you know exactly what the automation has to reproduce; automation and AI picks up from there.

Site C types 41 into a kWh field

Site C normally reports about 38,000 kWh of electricity a month, and Sites A and B send plausible figures. In March the Site C owner types 41, meaning 41 MWh in a field that expects kWh. All figures are made up.

Three layers could catch the error, in order:

  1. At entry: a locked drop-down for the unit would have stopped it. This template lacked one, so the number got through.
  2. At the year-on-year check: the rule, with an invented limit, flags any month more than 30% away from the same month last year unless explained. The change is (41 minus 38,000) / 38,000 = minus 99.9%, so it fails.
  3. At tie-out: the twelve months should add up to the utility’s annual statement. The slip leaves the total short by 41,000 minus 41 = 40,959 kWh.

You ask the owner, who confirms 41 MWh. The corrected March value is 41,000 kWh, which is 3,000 / 38,000 = +7.9% on last year and passes. Log the exception: old value, new value, reason, who corrected it and when. Had the slip surfaced after sign-off, the same entry would go into the restatement log.

The check tab is easier to build with the lookup and reconciliation habits in spreadsheet modelling. The folder you keep is what an outside examiner will ask for, as described in assurance readiness.

Late data and changes in the company

Site B misses the deadline. The invented rulebook says to use the previous month’s value, flag it as estimated and have you approve it. You then report the share estimated: 1 site-month out of 36 (three sites, twelve months) is 1 / 36 = 2.8%.

When the company buys or sells a business, follow the base-year recalculation policy you wrote in advance, described in the GHG Protocol Corporate Standard in practice. A divestment that lowers the total is not an improvement, and the log should say so.

What you will learn

  • Build a metric inventory and data dictionary with owners, definitions, units and deadlines
  • Design collection templates and portals that cut manual rework
  • Handle consolidation, boundary changes and acquisitions, and know when a base year is recalculated
  • Write checks for range, year-on-year change and tie-out to source records, and log each exception
  • Document estimation rules and treat data quality as more important than method sophistication
  • Keep an evidence folder, sign-off trail and change log that an assurance provider can follow

The parts of this topic

Metric inventory and data dictionaryabout 3 hours

A single list of every metric you report, with a written definition, unit, boundary, owner and the frameworks that ask for it.

Start from the frameworks you must answer. EFRAG publishes the ESRS and a datapoint list (implementation guidance IG 3), and the BRSR or GRI lists give theirs. Merge them into one dictionary with columns for definition, formula, unit, boundary, owner, source system and frequency. An entry that gives a name but no formula lets two sites calculate the same metric differently, so no row is finished without one. Add a column for the frameworks that use each metric: a number is then collected once and used many times.

  • EFRAG Knowledge Hub EFRAG · Tool Interactive ESRS and the implementation guidance; a free Starter plan lets you explore, a Professional plan adds saving features

Owners, deadlines and collection templatesabout 3 hours

Each data point needs a named owner, a deadline and a template that makes the wrong answer hard to enter.

The GHG Protocol page offers calculation tools with step-by-step guidance and electronic worksheets, which are a good model for templates. Give every site one workbook with locked formulas, drop-down units and a notes column; free-form spreadsheets that arrive in twelve layouts mean someone has to re-key all twelve. Agree the calendar with the people who own the source data, such as facilities, HR, EHS and finance, and add a reminder date before the deadline.

Consolidation, acquisitions and base-year recalculationabout 3 hours

When the organisation changes shape, comparisons only stay fair if the boundary and the base year are adjusted by clear rules.

GHG Protocol's Appendix E explains that base year emissions need recalculating when structural changes such as acquisitions or divestments change the inventory boundary. Under the fixed base year approach, an acquired company's sources are counted in both the base year and the current years, and divested sources are excluded from both. A year-on-year drop that is really a divestment should be flagged, not reported as progress. The standard leaves the significance threshold for recalculating to the company, so write yours into the recalculation policy.

Validation checks: range, change and tie-outabout 3 hours

Automatic checks catch typing errors, unit mistakes and missing months before a human reviews the story.

Write checks as rules: a value inside a plausible range for the site type, a year-on-year change within a stated percentage unless explained, monthly values that sum to the annual figure, and totals that tie to a source record such as an invoice or meter report. In pandas the same rules are a few lines of grouping and comparison, which makes them repeatable across sites. Log every exception with a reason. A check sheet that has never failed proves little. Break one number on purpose and confirm a check fails.

Estimation rules and data qualityabout 2 hours

Gaps are normal; what matters is that estimates follow written rules and are flagged.

The GHG Protocol standard says data quality matters more than method sophistication and that no methodology can compensate for poor quality input data. An estimation rulebook states when to use the previous month, a similar site, spend-based data or an average, who approves it and how the estimated share is reported. The Scope 3 standard, with its 15 categories and data quality indicators, shows how data quality varies by method. Report the share of site-months that were estimated; a reader can then see how firm the total is.

  • GHG Protocol: Scope 3 Standard GHG Protocol · Standard 15 categories and data quality indicators in chapter 7. The Scope 3 Evaluator tool was discontinued in August 2023, so do not rely on it

Staging data and portalsabout 2 hours

Move from many emailed files to a repeatable import step so that each cycle starts from the same clean structure.

Power Query records every transformation as a refreshable step, so a new month's files can be loaded and cleaned the same way; Microsoft's documentation says it does not modify the source data and offers hundreds of transformations. A supplier or site portal can do the same upstream with forms. Cells that are hand-edited after import cannot be repeated by anyone. Whichever route you choose, keep raw files untouched and make the cleaned layer reproducible.

Evidence, sign-off and restatement logabout 2 hours

Keep proof that each number is right, who approved it and what changed after publication.

The IAASB implementation guide for ISSA 5000 follows an engagement from acceptance to reporting and lists common challenges; reading it shows what a practitioner will want from you. Build an evidence folder per metric with source files, calculation, check results and approver. Record every change after sign-off in a restatement log with the old value, the new value, the reason and the date. A number corrected quietly in next year's report, with no log, leaves readers unable to compare years.

Practice task

Build an ESG data collection pack for three invented sites: a data dictionary of ten metrics, a template with an owner and deadline for each, a checks tab with range, year-on-year and tie-out tests, an estimation rulebook, an evidence checklist and a change log. Inject one deliberate error and show which check catches it.

Skills you build

Data dictionaryData validation checksInternal controlsEvidence managementRestatement logging

Resources for this topic

Every source here is free. Each link opens the provider's own site in a new tab.

Back to the whole roadmap

Live jobs

Live ESG analyst jobs

Roles whose title matches this career path.

All ESG & Sustainability Reporting jobs

Spot a problem with this topic?

Tell us what is wrong, missing or unclear here. Your note is attached to this topic.

Notes are sent from this page with JavaScript on. Without it, write to us on the contact page.