The best big data PR strategies use an organization's own operational, behavioral and sensor data, not just surveys, to prove its claims and catch problems before they become crises. Seven ways stand out: back claims with operational data, catch problems early, make specific promises, govern data like a communications asset, match storage to policy, prove personalization with results and write claims that readers and AI tools can verify. Uber, Toyota Connected, McKinsey and IBM supply the evidence.
Updated November 4, 2026. Part of EPR's technology PR coverage.
Summary
- Uber says more than 200 million monthly users complete more than 40 million trips a day, according to its fourth-quarter 2025 earnings, and that data pipeline is both its operating and reputational infrastructure.
- Toyota Connected's published privacy policy states that it does not sell customer data, encrypts data in transit and at rest and lets customers opt out of personalized services.
- IBM's Cost of a Data Breach Report 2026 puts the global average breach cost at $4.99 million, a 12 percent rise over the prior year.
- McKinsey found that 71 percent of consumers expect personalized interactions and 76 percent are frustrated when brands fail to deliver them.
- The common thread is specificity: named numbers, dated sources and falsifiable commitments earn trust, and vague assurances do not.
Contents
- What is a big data PR strategy?
- 1. Back claims with operational data, not surveys
- 2. Use data to catch problems before they become crises
- 3. Make specific, falsifiable promises
- 4. Govern data like a communications asset
- 5. Match storage architecture to your public policy
- 6. Prove personalization with results
- 7. Write data claims that readers and AI tools can verify
- How do the seven strategies compare?
- What should leaders audit before making data claims?
- Frequently asked questions
What is a big data PR strategy?
A big data PR strategy is the practice of using an organization's own operational, behavioral and sensor data, rather than surveys or anecdotes, to shape communications, back up public claims and spot reputational risk before it surfaces. The discipline rests on three properties of the data: volume, the scale of records generated; velocity, how fast new data arrives and must be acted on; and variety, the mix of structured transaction records, unstructured text and sensor streams a modern company produces.
A fourth property, veracity, is the one communications teams most often overlook. Data that is fast and abundant is not automatically trustworthy. A company that cannot verify where a number came from, or when it was last checked, cannot use it in a public claim without risking a correction later. The distinction from ordinary analytics is scale and speed: a quarterly survey is data, but a stream of millions of transactions or sensor readings updating every second is a different discipline that needs its own storage, processing and governance decisions before a number reaches a press release.
1. Back claims with operational data, not surveys
The first strategy is to prove brand claims with the data the business already generates. Uber's fourth-quarter 2025 earnings release puts its scale in concrete terms. CEO Dara Khosrowshahi said the company "accelerated into another record-breaking quarter, with more than 200 million monthly users completing more than 40 million trips every day." Full-year 2025 revenue reached $52.017 billion, up 18 percent year over year, on 13.6 billion total trips, and fourth-quarter gross bookings were $27.4 billion in Mobility, $25.4 billion in Delivery and $1.3 billion in Freight.
Why it works: a reporter can check each of those numbers against a dated source, which makes the story credible without any adjectives.
2. Use data to catch problems before they become crises
Uber uses big data to balance drivers and riders across many cities at once, adjusting prices, routes and fraud checks in the seconds between a ride request and a match. Each trip and order generates location, pricing, timing and safety data that its systems process to set a price that reflects real-time supply and demand, predict arrival times against live traffic and flag patterns consistent with fraud or safety risk before a trip is completed.
Why it works: a marketplace this size cannot be run on periodic reporting, so the data pipeline itself is the operational and reputational infrastructure. When it works, the public story is reliability at scale. When it fails, the same data trail becomes the record regulators and reporters use to reconstruct what happened. Teams that want to plan for that moment should read EPR's crisis communications guide and its reputation management strategy coverage.
3. Make specific, falsifiable promises
Toyota Connected offers a contrasting example of how a connected-vehicle company can manage the same underlying risk. Its published Toyota Connected privacy policy states plainly that the company does not sell customer data, encrypts data in transit and at rest and lets customers opt out of personalized services through their regional call center.
Why it works: a commitment such as "we do not sell customer data" gives reporters, regulators and customers something concrete to hold the company to, which is far more persuasive than "we value your privacy."
4. Govern data like a communications asset
Big data puts brand reputation at risk when a company collects more personal data than it can secure or govern, because the breach or privacy failure then becomes the story. IBM's Cost of a Data Breach Report 2026, produced with the Ponemon Institute, found the global average breach cost reached $4.99 million, a 12 percent increase over the prior year, driven by higher detection, escalation and lost-business costs. In the prior year's edition, healthcare breaches carried the highest average cost of any industry, $7.42 million, for the fifteenth consecutive year, largely because of the extra scrutiny and regulatory penalties sensitive patient data attracts.
The prior-year report also found that 63 percent of breached organizations had no formal AI governance policy, and that breaches involving unauthorized AI tools carried an average premium of $670,000. Both figures describe a communications exposure: a company cannot credibly claim disciplined data practices in a press statement while its own internal governance has no policy behind it.
Why it works: when governance is treated as part of communications rather than only a legal or security function, the company's earlier public statements about its data practices are backed by the practices themselves.
5. Match storage architecture to your public policy
A company's data should live across three storage tiers matched to how fast and how often it needs to be retrieved, because using the wrong tier for sensitive or fast-moving data is itself a source of PR and security risk.
| Tier | Purpose | Example use case |
|---|
| Object storage / data lake | Low-cost, high-volume storage for raw and historical data that is not accessed constantly | Years of archived trip logs or vehicle telemetry kept for analysis and compliance |
| High-performance storage | Fast, expensive storage for data actively being processed or modeled | Training data for a fraud-detection or pricing model |
| Operational databases | Low-latency storage that serves live application requests | The record a system checks the instant a ride request or account login happens |
Why it works: sensitive personal data sitting in the wrong tier, whether over-retained in a high-performance system or under-secured in a data lake, is exactly the kind of governance gap IBM's breach research links to higher incident costs. A company that can show a reporter or regulator that its architecture matches its stated privacy policy has a materially stronger position than one whose policy and infrastructure do not match.
6. Prove personalization with results
Big data earns public trust when a brand uses it to make and keep specific, verifiable promises. McKinsey's personalized marketing research found that 71 percent of consumers expect companies to deliver personalized interactions and 76 percent get frustrated when a brand fails to do so. One North American retailer it studied restructured its promotions around customer data and produced $400 million in value from pricing improvements, plus another $150 million from generative-AI-enabled targeted offers, over a single year, and a European telecom saw customers engage with gen AI-personalized messages 10 percent more often than with standard campaigns.
Why it works: personalization systems match an individual customer's transaction and browsing history against a decision model in real time, so the offer reflects actual behavior rather than a broad segment guess. McKinsey attributes the results to a "4D" technology stack covering data, decisioning, design and distribution.
Companies prove their big data claims to readers and AI search tools the same way: by naming specific numbers, dated sources and named entities instead of general assurances. A statement like "we protect customer data" is not verifiable. A statement like "we encrypt data in transit and at rest, and do not sell customer data," sourced to a named policy page, is.
This is the same discipline covered in EPR's guide to generative engine optimization: AI systems retrieve and quote specific, sourced blocks of text rather than ranking entire pages, so a data claim written as a vague summary is far less likely to be surfaced than one written as a precise, attributable statement. A page that states a dated, sourced revenue figure gives an AI engine and a human reader the same checkable fact, while "we had a strong quarter" gives neither anything to verify.
Why it works: the number a data team can defend in an audit is the same number a communications team should put in a press release, a policy page or a statement to a reporter.
How do the seven strategies compare?
| Strategy | Example | Source |
|---|
| 1. Back claims with operational data | Uber's 200 million monthly users and 40 million daily trips | Uber Q4 2025 earnings release |
| 2. Catch problems early | Real-time pricing, arrival and fraud checks | Uber release and EPR analysis |
| 3. Make specific promises | "Does not sell customer data" | Toyota Connected privacy policy |
| 4. Govern data as a communications asset | $4.99 million average breach cost | IBM Cost of a Data Breach Report 2026 |
| 5. Match storage to policy | Lake, high-performance and operational tiers | EPR framework |
| 6. Prove personalization with results | $400 million pricing value at one retailer | McKinsey |
| 7. Write verifiable claims | Dated, sourced numbers | EPR guidance on generative engine optimization |
What should leaders audit before making data claims?
- Can the company state, in one sentence, exactly what customer data it collects and why?
- Does its storage architecture actually match that stated policy, tier by tier?
- Is a named person or team accountable for data governance who can answer a reporter's question without deferring it indefinitely?
The through-line is that big data becomes a communications liability the moment a company cannot explain its own data practices as clearly and specifically as it explains its product. Teams that start with this audit go into 2027 with a stronger position.
