ESG in private equity: why buy-and-build breaks portfolio ESG data

ESG in private equity now runs on portfolio data: GHG emissions, renewable energy, diversity, accidents, hiring and engagement, collected from every company once a year. The frameworks are settled. What breaks is collection, and it breaks fastest in buy-and-build platforms, where each bolt-on arrives with its own sites and systems and nobody who owns the numbers.

What does ESG in private equity actually involve in 2026?

For most general partners, ESG has settled into a predictable annual cycle. Limited partners ask for a standard set of metrics. The fund's ESG lead sends a template to each portfolio company. Somebody at each company, usually a finance manager with a day job, fills it in. The answers get rolled up into an LP report and, for many firms, a submission to the ESG Data Convergence Initiative.

The EDCI is the closest thing private markets have to a common language. Its own site lists more than 400 general and limited partners as members, covering around $40 trillion of assets, with about 10,000 portfolio companies in its benchmark. Participating GPs share data collection templates with their portfolio companies and must complete 2026 data collection by 30 April 2027. The metric categories cover GHG emissions, decarbonisation, renewable energy, diversity, work-related accidents, net new hires, employee engagement and cybersecurity, and the summary of changes for 2027 adds a metric on artificial intelligence.

None of that is hard to understand. A first-time deal team can read the guidance in an afternoon. The difficulty sits somewhere else entirely.

Why portfolio ESG data breaks in buy-and-build

Recently we spoke to the ESG lead at a fast-growing UK professional services firm. In one financial year it had acquired five businesses and grown headcount by roughly half. She was the whole ESG function. Her description of the job will be familiar to anyone running ESG for a platform company: the strategy work was fine, the data work was swallowing her.

Every acquisition brings the same set of problems:

• New offices with energy bills paid by a landlord, a parent company or nobody anyone can find.
• A different finance system, so spend-based Scope 3 has to be rebuilt from a new chart of accounts.
• HR data held in another payroll provider, which makes headcount, turnover and diversity figures inconsistent for the first year.
• A carbon baseline that may need recalculating.

That last point catches people out. The GHG Protocol Corporate Standard expects companies to recalculate base year emissions when structural changes such as acquisitions have a significant impact, and leaves each company to define what significant means. Deloitte's guidance on the standard points to the 10 per cent cumulative threshold used by the California Climate Action Registry as one example. Do five bolt-ons in a year and you may well cross it.

So the platform company ends up with a footprint that changes shape every quarter, data owners who joined last month, and an LP template due in spring. A fund with ten platforms has this problem ten times over.

The questionnaire is the easy part

We have written before that in ESG due diligence the questionnaire is the easy part, and the same holds after the deal closes. Most GPs we speak to have a perfectly good template. Some have two, one for LPs and one for EDCI. What they lack is a reliable way to get honest numbers back from people who do not see ESG as their job.

Here is where we disagree with a lot of the market. The usual answer has been a portal: give every portfolio company a login, assign a contact, and wait. Portals work for the first year, when everyone is keen. By year two the contact has left, the login has expired and the fund ESG lead is back to emailing spreadsheets. We have seen assigned-portal models stall even at organisations with strong ESG teams, because the person holding the data is rarely the person who was given the login.

What works better is to go to the data owner directly, ask for one specific thing, accept it in whatever form they have it, and keep asking until it arrives. A PDF electricity bill is a perfectly good answer to the question "what did the Leeds office use in March". So is a forwarded email from the landlord. The system should do the reading and the follow-up, and the human should review the result.

What a workable portfolio ESG process looks like

If you run ESG for a fund or a platform company, this is the process we would build, whether or not you use software to do it.

• At acquisition: map the bolt-on's sites, suppliers, finance system and payroll provider within the first 100 days, and name a data owner for each metric. Common mistake: waiting until the annual cycle to find out who pays the gas bill.
• Collection: request each metric from its owner directly, accept documents as evidence, and send reminders automatically. Common mistake: sending one long template to one person per company.
• Cleaning: convert units, apply emission factors, flag gaps and mark every estimate as an estimate. Common mistake: silently filling gaps with last year's figure.
• Baseline: agree a recalculation policy and significance threshold before the first bolt-on, then apply it consistently. Common mistake: rebuilding the base year differently each time.
• Reporting: produce LP, EDCI and company-level outputs from the same data set, with a source behind every number. Common mistake: keeping separate spreadsheets for each audience.

The checklist matters more than the tool. That said, the collection and cleaning rows are where the hours go, and they are the rows most worth automating. Our guide to ESG data collection software covers what to look for, and our comparison of carbon accounting software sets out the questions to ask about estimates and audit trails.

Carbon is where the effort concentrates

Of all the metrics in a typical LP template, GHG emissions take the most work by a distance. Diversity and headcount come from one HR system. Emissions need energy data from every site, travel data from expense tools and, if you report Scope 3, supplier spend mapped to emission factors. Many portfolio companies above the size thresholds also owe SECR disclosures in their own annual reports, so the same numbers get collected twice if nobody joins the processes up.

Our advice to platform companies is to make the carbon data collection the backbone and hang everything else from it. Get one clean, sourced set of energy and travel data per entity, and the LP template, the EDCI submission and the statutory report all become exports. When we see one person carrying ESG for a whole group, this is usually the change that gives them their weeks back. We wrote more about that situation in ESG data management when one person owns it.

Where AI helps, and where it does not

General AI assistants are good at reading a guidance document or drafting an LP narrative. They cannot email the finance manager at a newly acquired subsidiary, wait three days and email again. They do not keep an evidence trail an auditor would accept. That gap is why we built Ella as an agent: it runs the collection, chases the people, reads the invoices and keeps a source on every number, so a fund or platform ESG lead reviews data instead of hunting for it. If that sounds like your portfolio, our ESG software for private equity page shows how it works across a group.

ESG in private equity will keep asking for more metrics. AI use is the newest, it will not be the last, and every addition lands on the same overstretched finance teams. Funds that fix collection now will barely notice the next metric. The rest will hire another consultant every spring.