Thursday, November 19, 2026
HomeFintechThe a part of the analytics stack we stopped questioning — FinTech...

The a part of the analytics stack we stopped questioning — FinTech Weekly


Onur Alp Soner is the co-founder and CEO of Countly.


FinTech strikes quick. Information is in all places, readability isn’t.

FinTech Weekly delivers the important thing tales and occasions in a single place.

Click on Right here to Subscribe to FinTech Weekly’s Publication

Learn by executives at JP Morgan, Coinbase, BlackRock, Klarna and extra.


When an information breach makes the information, it’s often framed as an exception – a misconfiguration, an neglected permission, a human error that would have occurred to anybody. The dialogue typically stops there, as if the incident itself had been the trigger. In actuality, breaches are extra typically alerts than failures. They expose dependencies that turned too central and too opaque lengthy earlier than something went improper. By the point knowledge is leaked, the danger has often been constructing quietly for years.

For a very long time, analytics sat in a protected psychological class. It was seen as observational, one thing that watched the system quite than formed it. In contrast to funds, id, or core infrastructure, analytics was not often handled as a layer that would materially have an effect on outcomes. 

In fintech, particularly, analytics now influences how techniques evolve and the way choices are made, shaping product behaviour, threat controls, and even automation. But the infrastructure behind it’s nonetheless typically exterior, working on third-party platforms exterior the organisation’s direct management.

That is the invisible dependency we stopped questioning.


Why “no PII” stopped being a adequate definition of security

When groups justify outsourcing analytics, the argument often centres on private knowledge. Occasions are anonymised. No names or emails are collected. With out PII, the danger is assumed to be low.

Whereas that logic held when analytics was primarily about counting customers and classes, it breaks down as soon as analytics begins capturing how techniques behave.

Fashionable occasion knowledge does way over describe particular person customers. It exposes the interior construction. Characteristic names, inner URLs, experiment variants, error states, timing patterns, and backend responses reveal how a product is designed and the way choices movement via it. None of this straight identifies an individual, but collectively it might reconstruct giant elements of a corporation’s inner logic.

That is the place the mosaic impact turns into related in observe. Particular person occasions seem innocent in isolation. Aggregated over time, throughout options and flows, they reveal how a product actually works. In fintech, this has actual penalties. Even anonymised occasions can trace at approval thresholds, threat scoring guidelines, or escalation paths. The sensitivity of analytics knowledge at this time lies much less in who it tracks and extra in what it reveals.


The bounds of “We deal with safety for you.”

Analytics distributors excel at scale, efficiency, and pace of integration. These strengths matter. What they don’t optimise for is long-term security, regulatory defensibility, or an organisation’s capability to clarify its personal structure beneath scrutiny.

When distributors say they “deal with safety,” they often imply the complexity is hidden. You may’t see how knowledge is mixed, retained, or what secondary alerts are derived. Invisibility is bought as simplicity, however management is changed with belief. Requirements like SOC2 validate controls, not architectural decisions. A system will be totally licensed and nonetheless focus delicate analytics knowledge in ways in which can be tough to justify beneath scrutiny.

That trade-off could also be acceptable elsewhere. For analytics that form choices, it creates structural threat by changing verifiable security with hidden techniques and assumed belief.

Monetary ledgers already function beneath this logic: traceability, auditability, and possession are non-negotiable. Analytics now shapes choices simply as consequential, however it has not but been handled with the identical self-discipline.

How structural threat accumulates in analytics techniques

Most analytics incidents don’t stem from a single unhealthy alternative. They emerge steadily, as techniques tackle duties they had been by no means designed to carry.

Groups add extra occasions, then extra context, then extra metadata. Characteristic flags, experiment IDs, inner error codes, backend states, and person classifications slowly discover their manner into occasion streams. Over time, analytics turns into an in depth mirror of how the product really works. At that time, it stops being a passive reporting layer and turns into a type of institutional reminiscence.

When knowledge is uncovered, what leaks isn’t simply uncooked numbers. It’s construction: how options are rolled out, how choices are staged, how providers work together, and the way edge circumstances are dealt with. Latest incidents have proven this clearly,  with logs as soon as thought of innocent revealing inner routing logic, experiment configurations, admin paths, and behavioural patterns that ought to by no means have left organisational management.

AI doesn’t introduce this threat, however it amplifies it. Behavioral analytics more and more feeds automated resolution techniques, which means structural publicity can affect mannequin habits, bias, and resolution logic. A single incident can have an effect on not simply transparency, however how techniques act going ahead.

In fintech, the impression is amplified additional. Analytics knowledge typically sits near techniques that assess belief, detect fraud, or automate approvals. Even when analytics doesn’t make choices itself, it more and more shapes the techniques that do.


Comfort as an alternative to scrutiny

For groups beneath stress to maneuver quick, polished dashboards, fast integrations, and immediate insights are exhausting to withstand. Over time, although, comfort tends to exchange scrutiny. Few organisations map their analytics knowledge flows intimately, assess how tough it will be to exit a platform, or account for a way a lot institutional information has successfully been outsourced. That is not often a deliberate alternative. It’s the results of treating analytics as tooling quite than infrastructure.

This isn’t an argument in opposition to third-party providers typically. The truth is, some layers are well-suited to being rented, particularly when failure is contained, and exit is easy. The excellence that issues is whether or not a system shapes outcomes.

To place it plainly, any system that influences entry, belief, eligibility, or core person expertise needs to be seen, auditable, and totally understood by the organisation that depends on it. Programs which can be simple to exchange and don’t encode institutional logic can safely stay exterior the establishment.

A easy take a look at clarifies the boundary: if this method disappeared tomorrow, would you continue to have the ability to clarify how your product behaves and why choices are made the best way they’re?

The broader accountability query

Fintech techniques more and more perform as public-facing infrastructure. They form who can open accounts, entry credit score, or take part within the economic system. That actuality shifts the duty mannequin. Architectural choices are now not purely inner technical decisions; they carry societal penalties.

When important layers reminiscent of cloud platforms, analytics techniques, or AI fashions are concentrated in a small variety of opaque techniques, failures and unexplained choices can ripple far past a single firm. Invisible dependencies do greater than improve safety threat. They weaken accountability.

Finally, if a system can’t be seen, it can’t be ruled. And techniques that can’t be ruled shouldn’t be trusted with choices that materially have an effect on folks’s lives. Analytics stopped being purely observational a while in the past. Our structure, requirements, and assumptions have but to catch up.


Concerning the creator

Onur Alp Soner is the co-founder and CEO of Countly, a digital analytics and in-app engagement platform. A technologist and self-starter, he bootstrapped Countly from the bottom as much as give firms extra management over how they perceive and work together with their customers. Beneath his management, Countly has grown right into a trusted platform for enterprises worldwide that wish to innovate rapidly whereas retaining person privateness on the centre of their development methods.
 

RELATED ARTICLES

Most Popular

Recent Comments