PowerGuardian Insights
Research methodology and privacy standard
PowerGuardian only publishes research it can defend line by line. This page sets out where our data comes from, what we will not do with it, and the checks a figure must pass before anyone outside PowerGuardian sees it.
Three kinds of figure
- Measured value — a figure counted directly from records, published with its sample size and sample definition.
- Modelled estimate — a figure produced by a model with stated assumptions. It is an estimate, and it is labelled as one.
- PowerGuardian calculation — arithmetic applied to a published external source, such as the regulator's capped rates. Reproducible from the published inputs.
We never present a modelled or calculated figure as an official statistic, and we never describe a PowerGuardian sample as representative of UK households.
Privacy safeguards
- Research using data from people who use PowerGuardian uses only records where the person has actively opted in. Consent is never pre-ticked, applies from the point it is given, and can be withdrawn at any time in My PowerGuardian.
- No individual, household or address is ever identified, described or made identifiable by combining figures.
- A published figure requires a minimum of 30 qualifying records. Segments below that are withheld and shown as withheld — they are never merged into an "other" category to inflate a number.
- Percentages are never published without their sample size.
- Geography is published no finer than: United Kingdom, Great Britain, UK nation, Region, Postcode area. Full postcodes are never published.
- Records are excluded where consent is absent, the entry is incomplete, implausible or a duplicate, or where a test entry is identified.
See also our privacy notice and data deletion process.
What data we hold, and whether it is publishable
This is the honest position, not an aspiration. Most of our own activity data is not yet large enough to publish anything from.
| Source | Status | What is missing |
|---|---|---|
| Ofgem price cap — canonical recordThe capped unit rates, standing charges and benchmark consumption for the current cap period, with source, publication date and verification date. | Publishable now | Only the current period is verified. No verified prior-period standing charges are held, so change-over-time analysis is not yet possible. |
| Ofgem price cap — change logRecorded changes to the headline cap figure over time. | No qualifying records yet | Too few historical entries to support a period-over-period release. Needs verified back-filling from Ofgem's published decisions. |
| Supplier performance scoresPowerGuardian's scored assessment of suppliers with per-claim evidence records. | Held, not yet verified | Scores are published on site, but a publicity release would need each contributing input re-verified against its dated source for the release period. |
| Calculator and tool resultsSaved results from PowerGuardian calculators, such as boiler and quotation screenings, where a visitor chose to save them. | No qualifying records yet | No qualifying saved records. Nothing can be published until the sample clears the 30-record threshold and consent is recorded. |
| My PowerGuardian household profilesSelf-reported household characteristics and energy details from signed-in users. | Sample too small to publish | Far below the 30-record threshold, and only records with explicit research consent are eligible. |
| Energy journey progressWhich money-saving actions signed-in users have marked as completed. | Sample too small to publish | Below threshold. Also needs a documented definition of a 'completed action' before any share can be published. |
| Search demand for PowerGuardian pagesAggregate search queries and impressions for PowerGuardian pages. | Held, not yet verified | Usable as a signal of public interest, but it measures PowerGuardian's own visibility, not household behaviour. Any release must say so. |
| Briefing subscribersSubscription records for the Monday UK Energy Savings Briefing. | Sample too small to publish | Not a research dataset. Subscriber counts will not be published as a household finding. |
Repeatable research programmes
Standing Charge Burden Index
How much of a capped household energy bill is fixed daily cost before any energy is used, and how does that share change with consumption?
Measures: Annual standing charges for electricity, gas and dual fuel, and the share of the annual bill they represent at three usage levels. · Cadence: Each time the price cap changes
Published view the release
Price Cap Period Comparison
How have capped unit rates and standing charges changed between cap periods?
Measures: Period-over-period change in unit rates, standing charges and the resulting annual bill. · Cadence: Quarterly, on each cap announcement
Not published: Verified prior-period unit rates and standing charges are not yet held in the canonical record.
Regional Fixed Cost Gap
How much does the fixed element of a bill differ between UK regions?
Measures: Regional standing charges and the annual gap between the highest and lowest region. · Cadence: Each cap period
Not published: Regional standing charges on site are unverified placeholders. They must be sourced and verification-dated before publication.
Supplier Practice Tracker
How do suppliers compare on the practices that cost households money?
Measures: Scored supplier assessments with per-claim dated evidence. · Cadence: Half-yearly
Not published: Each contributing input needs re-verification against its dated source for the release period.
Household Action Gap
Which money-saving actions do households complete, and which do they leave undone?
Measures: Share of consenting users completing each action category, with sample sizes. · Cadence: Quarterly, once the sample clears the threshold
Not published: Consenting sample is far below the 30-record publication threshold.
Pre-publication checklist
Every item is confirmed by a person before a release can be published or sent to a journalist. Nothing about prices, tariffs, dates or supplier performance is shared automatically.
- Source verified against the original publication
- Sample definition verified
- Sample size verified
- Duplicate and invalid records removed
- Calculation independently checked
- Collection and publication dates checked
- Geographic labels checked
- Privacy suppression threshold passed
- Limitations written
- Headline supported by the underlying result
- Chart agrees with the underlying data
- Supporting file archived
- Reviewer approval recorded
Verification, versions and corrections
- Each figure is traced to the metric and dataset version that produced it. If the dataset changes, every claim written against the old version is flagged for re-checking before reuse.
- Each release carries a dataset version and a review date. A review date means the evidence was re-checked, not that the page was edited.
- Material corrections are dated and explained on the release and listed in the corrections log. Corrected releases keep their URL so anyone who cited them can see what changed.
- Where a release is superseded by a newer period, the older release stays available and points to the newer one.
Reuse terms
PowerGuardian research figures and charts may be reproduced, including commercially, provided PowerGuardian is credited as the source, the figure is not altered, and any stated limitation is not removed. Each release includes the exact attribution and citation wording to use.