
Original Research
Data-led PR assets
A data-led PR asset gives a journalist enough information to check a finding, describe it accurately and judge its limits. A striking chart without the underlying …
A data-led PR asset gives a journalist enough information to check a finding, describe it accurately and judge its limits. A striking chart without the underlying values, time period or method is difficult to use responsibly. Build verification material alongside the headline, not after a question arrives.
Separate the finding from the package
Write one sentence stating what the data shows, with the population, measure and period.
Prepare the supporting material a journalist needs: the values behind the headline, a chart only if it clarifies the pattern, a short methodology note and a source-data file that can be released. Keep internal calculations and personal records out of the public pack.
The ABS provides examples of identifying the exact publication, table and reference period when citing its data. These are practical models for a press pack even when the underlying study was produced by a company.
When citing ABS material, include an access date as well as the reference period and publication details. The ABS notes that its content is continuously updated, so a citation should identify the version or time period relevant to the finding rather than leaving readers to guess which release was used.
For ABS webpages whose URLs contain a reference period, the ABS recommends using that period-specific page rather than a latest-release page. This helps keep the citation aligned with the data underpinning the asset.
ABS Data Citation Best Practices vs. Common Pitfalls
- Common Pitfall
- Cite latest-release page without specifying reference period
- Best Practice
- Include access date and reference period in citation
- Common Pitfall
- Omit access date, leading to ambiguity about data version
Make every number traceable
Keep a working record from raw input to final figure. The companion guide on writing a methodology note covers the calculation choices a public study needs to document.
Do not turn an association into a causal claim. A survey of customers who answered voluntarily does not represent every Australian customer unless the design supports that inference. A comparison between two periods also needs context about product mix, measurement changes and seasonality. Put the important limit beside the finding, not buried in an appendix.
Start with the audience’s question
A dataset does not determine the most useful PR angle on its own. Isentia recommends understanding the audiences and conversations connected to a brand, then shaping stories around the topics and channels those audiences follow.
That can mean looking at website visitors, customer data and social media conversations to identify audience groups and interests. Existing customers are already engaging with the brand, while social conversations can reveal communities and topics beyond that customer group.
Keyword research can also help identify the queries people use in search engines and connect an asset to current topics in an industry. Treat those keywords as a way to understand audience interest, not as evidence that the data proves a particular conclusion.
Isentia gives an example from a Pulsar panel search of Reddit communities interested in public relations and communications: it found around 1,500 unique authors, with male Redditors at 6.3% and female users at 5.8%. This is an example of a defined audience insight, not a figure about all Australians or all Reddit users.
Audience Insights from Reddit Community Analysis
- 1,500Total Unique Authors Identified
- 6.3%Male Redditors (%)
- 5.8%Female Redditors (%)
Publish only what can be released safely
Before offering source data, check privacy obligations. The OAIC publishes guidance on de-identification and the Privacy Act, and the companion guide covers how to make source data usable without exposing private information.
For a real press pack, ask a second reviewer to check the released material and make corrections versioned and visible. The asset is ready when a journalist can understand the result without accepting the brand's headline on trust.
Pre-Publication Safety Review for Data-Led PR Assets
- De-identify personal information per OAIC guidanceYes
- Second reviewer checks released materialYes
- Versioned corrections visible in release packYes
- Source data file available with anonymised identifiersYes
In this guide
- Creating a table journalists can verifyA journalist should be able to read a table and reproduce the claim in the pitch. Provide the values, definitions and source context in a format that survives …
- Preparing a chart with a visible source and dateA chart should make one supported pattern easier to see. It must also tell a journalist what was measured, when, where and from which data version. Without those …
- Writing a methodology note for a public study: a guideA methodology note tells a reader how a finding was produced and where its inference stops. It should be short enough to scan but complete enough that another …
- Making source data usable without exposing private informationSource data helps journalists verify a finding, but a raw file may contain personal, commercial or third-party information that should not be released. Decide what …



