By Jordan Baker, Founder of Fair Public Relations and former UK national journalist
A fintech company can sometimes turn product or customer data into a strong media story, but having a large database is not enough. The company needs a legitimate right to use the information, a finding that matters to readers, a clear methodology, meaningful denominators, careful anonymisation or aggregation, and language that does not extrapolate one platform's behaviour to an entire economy.
The strongest data stories usually answer a question rather than simply advertise scale. "We processed 2 million transactions" is a company statistic. "Invoice payments to small suppliers in our sample took longer in December than in September, with the methodology and customer mix clearly explained" may be a story if the finding is robust and useful.
A fintech data story should never require journalists to trust the headline more than the evidence.
Why is "we have a lot of data" not automatically a journalist's story?
Data is raw material.
A story needs a finding, context and consequence.
Imagine a payments platform that has 10 million transaction records. That sounds impressive, but a journalist still needs to know:
what the records represent
which customers are included
which countries are included
what period is covered
whether one customer can generate thousands of records
whether the platform changed during the period
whether the customer mix changed
what has been excluded
whether the company is allowed to analyse and publish the information
whether the result is genuinely unusual
A huge dataset can produce a weak story if the denominator is unclear.
A smaller dataset can sometimes produce a useful specialist story if the sample is well defined, the limitation is honest and the finding matters to the publication's readers.
From a journalist's perspective, the useful question is not "How much data do you have?" It is "What can this data responsibly tell me?"
Which fintech datasets may contain media-story potential?
Different fintech businesses may have very different first-party datasets.
Payments
Potential subjects might include payment timing, transaction frequency, merchant categories, failed-payment patterns or business-payment behaviour.
Risks include revealing customer relationships, exposing sensitive operational patterns or suggesting that platform activity represents the wider payments market.
Budgeting tools
Potential subjects might include categories of spending, recurring-payment behaviour or changes within the app's user base.
Risks include highly sensitive personal financial data, selection bias and unsupported claims about national household finances.
Accounting automation
Potential subjects might include invoice processing, reconciliation workload, common errors or seasonal finance-team patterns.
Risks include confidential business information, customer identification and treating software usage as a measure of the whole economy.
Compliance technology
Potential subjects might include operational volumes, types of checks or workflow trends.
Risks include revealing security processes, suggesting regulatory conclusions the platform is not qualified to make or publishing information that could help bad actors.
Fraud-awareness products
Potential subjects might include categories of attempted scams or suspicious activity observed by the platform.
Risks are particularly important. The company should not publish operational details that weaken security, expose victims or imply that an attempted event was legally established fraud without the appropriate basis.
The point is not that each dataset should become a press release. It is that each one needs a different risk and methodology review.
Start with the reader, not the spreadsheet
A useful data story begins with an audience question.
Consumer question:
"Are users of this budgeting app changing how often they use buy-now-pay-later services?"
Business question:
"Are SME customers in this payments platform waiting longer for invoices to clear?"
Accountancy question:
"Which reconciliation problems create the most manual exceptions for finance teams using this software?"
Payments-industry question:
"Has the share of transactions using a particular payment route changed within this merchant sample?"
The reader determines what context matters.
A consumer journalist may need plain language, population limitations and a qualified expert who can explain why the finding matters without giving individual financial advice.
A B2B trade journalist may care more about workflow definitions, customer type, system architecture and whether the result reflects a specific vertical.
A good PR team should be able to explain why the same dataset may support a trade story but be too narrow for a national consumer headline.
What does "anonymised" actually mean?
This is an area where loose PR language can become dangerous.
Deleting a customer's name is not automatically anonymisation.
The UK's Information Commissioner's Office distinguishes anonymisation from pseudonymisation. Pseudonymised data can still be personal data if the information can be linked back to a person using additional information.
A fintech company should therefore avoid phrases such as "the data is anonymous because names were removed" unless the privacy assessment genuinely supports that conclusion.
Re-identification can happen through combinations of data.
A location, employer, transaction size, date and rare product category may identify somebody even without a name.
The company also needs to consider its customer agreements, privacy notices, contractual commitments and applicable law.
PR is not the function that decides whether data can legally be disclosed. The data owner and appropriate privacy/legal professionals should make that determination before a journalist sees the material.
What should a proper methodology note include?
A journalist-readable methodology should answer the questions most likely to alter the interpretation.
Source: Where did the data originate?
Eligibility: Which records were included?
Period: What dates were analysed?
Geography: Which countries or regions are represented?
Customer mix: Consumers, micro-businesses, enterprise accounts or a mixture?
Denominator: Users, accounts, transactions, invoices or merchants?
Exclusions: Which records were removed and why?
Missing data: Are some fields incomplete?
Duplicates: Can one event appear more than once?
Weighting: Were high-volume customers allowed to dominate the result?
Comparison: What exactly is being compared with what?
Uncertainty: How sensitive is the conclusion to sample size or customer mix?
Conflicts: Does the company benefit commercially from a particular interpretation?
A strong methodology note does not need to be academic. It needs to make the result inspectable.
What does an illustrative fintech data analysis look like?
ILLUSTRATIVE ONLY. The following numbers are fictional and are not Fair PR client data.
Imagine a B2B invoice-processing platform analysing how long invoices in its customer sample take to move from "issued" to "marked paid".
September:
18,200 invoices
Median time: 12 days
Customer mix: 60% professional services, 25% retail suppliers, 15% other
December:
21,500 invoices
Median time: 15 days
Customer mix: 38% professional services, 47% retail suppliers, 15% other
A tempting headline is:
"Businesses are taking 25% longer to pay invoices."
That headline is too broad.
The platform only observed its own customers.
The customer mix changed significantly.
Retail suppliers may have different payment patterns from professional-services firms.
Seasonality may matter.
The product may also have acquired larger customers between the two periods.
A more defensible headline might be:
"Invoice payment times rose in December within our platform sample, but a shift towards retail suppliers means the change cannot be attributed to the wider economy."
That is less dramatic. It is more credible.
A journalist may still find the pattern useful, especially if the company can segment the data further and explain the limitation.
What is the difference between an observed trend and a causal claim?
This distinction is fundamental.
Observation:
"Users in our app reduced discretionary spending in January compared with December."
Causal claim:
"High interest rates caused consumers to reduce discretionary spending."
The first statement describes the sample.
The second explains why it happened.
The company's platform data may not be able to prove the cause.
Other explanations might include:
Christmas seasonality
changes in the app's user base
a marketing campaign attracting different customers
changes in categorisation
product redesign
income timing
geographic mix
economic conditions unrelated to the claimed cause
A useful expert can discuss plausible context, but the story should distinguish observation from explanation.
Correlation is not causation, and proprietary platform data is not automatically a representative economic survey.
Which fintech data stories fit different subcategories?
| Fintech type | Possible data source | Defensible calculation | Reader question | Important caveat |
|---|---|---|---|---|
| Budgeting tool | Aggregated app category usage | Share of active users using a feature over a defined period | Are users changing how they budget? | Users are self-selected and may not represent all households |
| Payments platform | Merchant transaction records | Median transaction or settlement timing within eligible records | Are payment patterns changing for this sample? | Merchant mix, payment type and geography can distort comparisons |
| Finance-operations software | Invoice/reconciliation records | Exception rate or median processing time | Where do finance teams face manual work? | Software behaviour does not measure the whole finance profession |
| Accounting automation | Workflow events | Share of invoices requiring manual intervention | Which tasks remain difficult to automate? | Product configuration and customer maturity may affect results |
| Fraud-awareness product | Categories of attempted events | Proportion of flagged events by defined category | What kinds of suspicious activity are users encountering? | Do not reveal security-sensitive methods or label unproven events as criminal fraud |
The table is a story-development tool, not permission to publish any dataset.
How should financial-benefit or risk-reduction claims be handled?
Claims such as "our users save £500 a year", "customers cut fraud by 60%" or "this tool reduces investment risk" require more than an interesting chart.
The company should ask:
what exactly was measured?
compared with what?
was the comparison controlled?
did customers self-select?
is the figure an average, median or best case?
what proportion achieved the claimed result?
does the claim apply in every market?
could the statement be a regulated financial promotion?
has the wording been reviewed by the appropriate internal or professional owner?
In the UK, financial promotions within the FCA's remit must be fair, clear and not misleading. The specific rule depends on the product and communication.
A data press release does not magically convert a financial-performance claim into a neutral fact.
Nor does independent journalism mean the company can send unsupported source material and transfer responsibility to the reporter.
What should a journalist-readable evidence pack contain?
The evidence pack should be separate from the pitch itself.
Include:
Headline finding: One sentence stating exactly what was observed.
Dataset: What records were analysed.
Dates: The period covered.
Geography: Where the data came from.
Population: Who or what was included.
Denominator: Users, invoices, transactions, businesses or another unit.
Method: How the calculation was performed.
Exclusions: What was removed.
Limitations: What the data cannot prove.
Chart or summary: A simple table or chart whose numbers match the methodology.
Expert comment: A credible interpretation that does not overreach.
Data owner: A named person able to answer methodology questions.
Approval trail: Internal confirmation that the material can be released.
Do not include raw personally identifiable information.
Do not include security-sensitive records.
Do not send a spreadsheet of customer transactions to prove that the headline is true.
How can fintech storytelling avoid fearmongering?
Financial stories can become dramatic very quickly.
Weak fictional headline:
"Everyone is drowning in debt as spending crisis explodes."
Stronger fictional headline:
"Among active users of our budgeting app, the share recording a credit-card payment rose from X% to Y% between two defined periods."
The second line still needs context, but it tells the reader what the data actually represents.
Another weak headline:
"Fraud up 300% in Britain."
Stronger:
"Flagged account-takeover attempts tripled within a small sample of our UK business customers, from 12 to 36 events, so the result should not be treated as a national fraud estimate."
Sometimes the limitation is part of the story.
Credible brands do not need to hide uncertainty to be interesting.
When is customer data the wrong PR route?
Not every fintech company has a publishable dataset.
Customer data may be the wrong route if:
the sample is too small
the users are too atypical
the data is not collected consistently
privacy rights are unclear
customer contracts restrict use
one major customer dominates the sample
the company cannot explain the methodology
the finding is commercially sensitive
the story would reveal security methods
the desired headline depends on causation the data cannot prove
In those situations, expert commentary may be more honest.
A product lead or qualified specialist can sometimes explain a real market issue using public evidence without turning customers into a dataset.
A genuine business milestone may also be more appropriate if there is actual company news.
How should fintech charts and percentages avoid misleading readers?
A chart can distort a sound dataset if the presentation removes the context.
If a percentage rises from 1% to 3%, the relative increase is 200%, but the absolute change is two percentage points. Both calculations can be mathematically correct while creating very different impressions. A journalist should be able to see the underlying numbers.
Likewise, a graph with a truncated vertical axis can make a small movement look dramatic. A chart showing transaction volume can also be misleading if one enterprise customer suddenly contributes most of the activity.
For every chart, label:
the unit being measured
the denominator
the time period
the geography
whether the figure is a count, mean, median, rate or percentage
any material exclusions
whether the customer mix changed between periods
If the press release says "late payments doubled", the evidence pack should show exactly what doubled. Ten late payments becoming twenty is different from 10,000 becoming 20,000.
The same principle applies to percentages of users. If only a subset answered a survey question, use the number who answered that question as the relevant denominator rather than the entire platform population unless the methodology genuinely supports otherwise.
A simple chart that exposes its assumptions is more credible than a dramatic graphic that hides them.
What does Fair PR's adjacent finance evidence prove, and what does it not?
Fair PR does not currently have a live fintech customer-data case study that I can responsibly present as proof of this exact tactic.
Experity Wealth is a live adjacent finance example. It is a wealth-management business, not a fintech platform. Its three-month PR campaign recorded 17 published media placements across six publication titles by positioning founder Robert Cannon as a finance expert.
That case demonstrates the value of credible finance expertise.
It does not demonstrate a fintech data report, customer transaction analysis or proprietary research campaign.
That distinction is important because good case-study use should tell the reader what evidence actually exists, not stretch neighbouring work to fit the headline.
What is a responsible release sequence for a fintech data story?
A lean process can look like this:
1. Identify the internal data owner.
2. Define the editorial question before analysing everything.
3. Confirm whether the data may be used for this purpose.
4. Assess anonymisation, aggregation and re-identification risk.
5. Define the sample and denominator.
6. Audit calculations.
7. Check whether customer mix or product changes distort the comparison.
8. Identify any financial, security or regulated claim.
9. Obtain appropriate subject-matter, privacy or compliance review.
10. Write a methodology note before the press release.
11. Develop a restrained headline.
12. Identify the journalist audience.
13. Supply the finding and limitations together.
14. Keep an expert available for follow-up.
15. Correct errors quickly if a calculation or source detail changes.
This is slower than inventing a provocative statistic on Friday afternoon.
It is also far more likely to survive serious journalistic scrutiny.
What 12 questions should a fintech company answer before releasing customer data?
1. What exact reader question does the finding answer?
2. What is the denominator?
3. How large is the relevant sample?
4. Which dates and markets are included?
5. Has the customer mix changed?
6. Are the records anonymous, aggregated, pseudonymised or still identifiable?
7. Are we authorised to use the information this way?
8. What does the data not prove?
9. Are we making a causal claim from an observation?
10. Does any wording imply a financial benefit, return, security outcome or regulated status?
11. Can a named data owner explain the method to a journalist?
12. Would we be comfortable publishing the methodology beside the headline?
If question 12 produces discomfort, the headline is probably ahead of the evidence.
Frequently asked questions
Can we pitch anonymised customer transaction data to financial journalists?
Possibly, but "anonymised" needs to mean more than deleting names. The company should assess whether individuals or customers can still be identified from the remaining information, whether it has the right to use the data, and whether the sample can support the conclusion. Appropriate privacy, legal and data-owner review may be needed before publication.
What if our platform has very few customers?
A small dataset may still illustrate a narrow operational point, but it often cannot support broad market claims. Be explicit about the sample size and avoid presenting a handful of users as evidence about all consumers or businesses. If the data is too thin, use verified expert commentary, public evidence or wait until a more meaningful sample exists.
Can PR claim that app customers save money based on platform usage?
Only if the company has evidence that genuinely supports that claim and the wording has been appropriately reviewed. An observed change in user behaviour does not automatically prove the app caused a saving. The comparison, denominator, timeframe, customer selection and other explanations all matter, especially where the statement could affect financial decisions.
Will a fintech data story guarantee coverage or backlinks?
No. A strong methodology makes a story more credible, but independent editors still decide whether the finding is relevant enough to publish and whether to include a link. PR cannot guarantee a named publication, backlink, ranking, sign-up or commercial result.
Talk to Fair PR about Fintech PR
If your fintech business has genuinely useful data and can demonstrate how it was collected, checked and responsibly cleared for disclosure, Fair PR can assess the potential journalist story under Standard PR. If you have specialist knowledge but no safe publishable dataset, we can also discuss suitable expert-commentary opportunities. Tell us what you actually know, what evidence exists and who has authority to release it.
Written by Jordan Baker
Founder of Fair Public Relations and former UK national journalist
Jordan Baker is the founder of Fair Public Relations. He previously worked as a UK national journalist and now advises founders, experts and small businesses on earned media, journalist outreach and credible public positioning.
View Jordan Baker's media expert profile