Estimated vs Measured Data: How Data Quality Grades Affect ESG Scores

Two portfolios can submit the same total energy consumption figure and receive noticeably different treatment for it. The number is identical. What differs is where it came from and increasingly reporting frameworks are built to notice that difference explicitly. Data quality grades are the mechanism behind this: a formal tier attached to each reported figure marking it as measured estimated or derived from a proxy and that tier can matter as much to a final score as the figure itself. A portfolio that reports confidently but cannot say which of its numbers were actually read off a meter is not reporting from a position of strength. It just has not been asked the question yet.

This guide explains what separates a measured figure from an estimated one in practice why frameworks grade data quality at all and where this quietly costs portfolios points they never realised were at stake.

Not sure how much of your reported data would actually grade as measured? Sustainify AI helps real estate teams understand and improve their data quality grades across every framework they report against.

Why frameworks grade data quality at all

A carbon figure on its own tells a reviewer almost nothing about how much to trust it. Two identical numbers can have entirely different provenance one from a calibrated meter read at the point of consumption the other from a floor area benchmark applied because nobody could get better data for that building. Data quality grading exists to make that difference visible rather than hiding it inside a total that looks the same either way. GRESB scoring rewards portfolios that can demonstrate a higher proportion of measured verifiable data precisely because a score built on genuinely measured figures means something a score built on assumption does not.

What separates a measured figure from an estimated one

A measured figure comes directly from a meter reading an invoice or another primary source recorded at or close to the point of actual consumption. It can be traced back to that source without needing to explain or defend an assumption along the way. An estimated figure by contrast is derived: a benchmark applied to a building without sub metering an extrapolation from a partial year of data or a proxy used because the real source was never captured. Neither category is inherently wrong to use. What matters is that the difference between them is recorded honestly rather than both being presented with the same apparent certainty once they land in a final report.

The distinction gets more granular still in practice since some frameworks recognise an intermediate tier: data that has been calculated from measured inputs but through a method involving assumptions such as tenant carbon allocation applied to a whole building meter reading. This sits between fully measured and fully estimated and it needs its own honest label rather than being rounded up to one extreme or the other.

Where portfolios lose points without realising it

The most common way portfolios lose ground on data quality has nothing to do with poor underlying performance. It comes from figures that were genuinely estimated being submitted without that fact being flagged anywhere. A reviewer scoring the submission has no way to know the difference and depending on how the framework handles unlabelled data an unflagged estimate can be treated more harshly than a flagged one precisely because its provenance cannot be verified either way.

The second common loss comes from inconsistency across a portfolio. A handful of buildings with excellent sub metering pull the average up while a larger number relying on benchmark estimates pull it down and if the split between the two is not clearly documented the entire submission can read as less reliable than the well measured half of the portfolio actually deserves. A governed carbon calculation process that tracks this split explicitly building by building protects the portfolio's strongest data from being dragged down by association with its weakest.

Wondering whether your submission is being scored on the strength of its weakest data? See how Sustainify AI tracks data quality grades building by building so your best data gets credit for what it actually is.

What a higher grade actually requires

Moving a figure from estimated to measured is rarely about better maths. It is almost always about better infrastructure and better documentation. Sub metering coverage is the single biggest lever since it converts an entire category of estimated tenant consumption into directly measured data. Where sub metering genuinely is not feasible the next best step is documenting the estimation methodology thoroughly enough that a reviewer can see exactly how the figure was derived which does not make it measured but does make it a defensible clearly labelled estimate rather than an unexplained guess.

This work connects directly to calculation lineage. A figure cannot honestly be graded as measured unless there is a clear path back to the source reading that produced it and a portfolio that has not built that traceability into its process will struggle to prove a grade even where the underlying data genuinely was measured.

Building toward better grades without waiting for a mandate

Portfolios do not need a formal requirement to start improving their data quality grading. Beginning with an honest internal audit building by building of what proportion of current figures could genuinely be labelled measured versus estimated usually surfaces the gap clearly enough to justify prioritising metering investment where it matters most. Understanding how a governed data process works helps a portfolio see exactly where in its pipeline the estimated figures enter which is usually a smaller more addressable list than teams expect once they actually look for it. This discipline strengthens every framework a portfolio reports against from SECR to CSRD and supports stronger audit readiness as scrutiny increases across the board.

A test for your own data quality grade

Take your portfolio's most recent submission and ask three questions. Could you list right now which specific buildings contributed measured data and which contributed estimates. Is that distinction visible anywhere in the submission itself or only in your own head. If a reviewer asked you to defend the estimated figures specifically could you explain the method behind each one without having to go and find out first. If any of those answers is no the grade attached to your submission is being decided for you not by you.

Teams working through this for the first time often find it useful to review practical data quality guides and sector specific sustainability insights and to compare approaches with peers through a partner programme where relevant. If you are weighing up tools to support this reviewing pricing and learning more about the team behind the platform is a sensible next step before your next submission cycle.

Ready to know exactly what grade your data would earn before a reviewer decides for you? Talk to Sustainify AI about improving your data quality grades across every framework your real estate portfolio reports against.

Frequently Asked Questions

What is a data quality grade in ESG reporting? 

It is a formal tier attached to a reported figure typically marking it as measured estimated or derived which tells a reviewer how much confidence to place in that specific number.

Why do measured and estimated figures need to be distinguished at all? 

Because two identical numbers can have entirely different reliability depending on their source and blending them without distinction misrepresents the overall confidence a reviewer should place in the submission.

How does GRESB treat estimated versus measured data? 

GRESB scoring rewards portfolios able to demonstrate a higher proportion of measured verifiable data since figures with clear provenance carry more weight than those based on assumption.

What is the most common way portfolios lose points on data quality? 

Submitting estimated figures without flagging them as estimates which leaves a reviewer unable to assess their reliability and can result in harsher treatment than a properly labelled estimate would receive.

Does sub metering improve a portfolio's data quality grade? 

Yes significantly. Sub metering converts estimated tenant consumption into directly measured data which is usually the single biggest lever available for improving a portfolio's overall grade.

Can an estimate ever be considered defensible rather than a weakness? 

Yes when the estimation methodology is documented clearly enough that a reviewer can see exactly how the figure was derived even though it remains labelled as an estimate rather than measured data.

How does tenant carbon allocation affect data quality grading? 

Figures calculated by applying tenant carbon allocation to a whole building meter reading often sit in an intermediate tier calculated from measured inputs but through an assumption based method which needs its own honest label.

Why does inconsistency across a portfolio hurt data quality scoring? 

If well measured buildings and poorly measured ones are not clearly distinguished the entire submission can appear less reliable than its strongest data actually deserves.

Does data quality grading matter for frameworks beyond GRESB? 

Yes. The same discipline supports stronger outcomes under SECR and CSRD and it is increasingly relevant to audit and assurance review as scrutiny of underlying data sources increases.

How should a portfolio start improving its data quality grade? 

Start with an honest building by building audit of which figures are genuinely measured versus estimated then prioritise metering investment where the gap is largest. You can explore how a governed reporting process works or get in touch to discuss your portfolio specifically.