There’s a leak in your office: do you call a plumber to fix the pipe, or dedicate your team to full-time mop duty?
Fixing the issue is more efficient, less risky and considerably better for your employees. When it comes to bad data, though, most asset managers and fund servicers have had no choice but to grab the proverbial bucket.
Traditional systems are built so that reconciliation sits at the end of the chain – flagging the mess caused by bad data rather than stopping it at the source. By the time an error surfaces, it has already disrupted a process. Whether you’re waiting months for a new control, processing a batch of trades overnight, or even getting results in minutes, you’re still too late.
There is an alternative. One that switches reconciliation from a last resort into a first line of defence: catching errors before they touch your NAV calculations, your IBOR positions, your regulatory submissions, or your fund flow data. The result is higher straight-through processing (STP), operations teams freed from manual exception-chasing, and a data infrastructure that can scale.
Reconciliation as the bouncer, not the janitor
Improving efficiency is about finding errors before they have a chance to break anything. Every process – confirming a trade, reconciling custody positions against an IBOR, validating report inputs – depends on clean, accurate data to function. Data quality is what enables the process, not just what gets checked afterwards.
Considering the data requirements of a process before it runs is data-centricity: building proactive controls around the data each process needs, rather than reconciling after the fact.
How to build data-centric controls
Start by mapping the data requirements for each process: what data is needed, where it comes from, and what must happen to it before it’s usable. Then build controls that verify accuracy proactively.
A straightforward example: a weekly check with trading counterparties to confirm that Standard Settlement Instructions (SSIs) are still current – even when no trades are active. Any errors get flagged and fixed before they matter. The next time you trade, the SSIs are accurate, STP is higher and one of the most common causes of settlement failure is gone before it arrives.
Data-centric controls also cut false breaks – the kind generated when a date is formatted differently across two datasets, or when a minor field mismatch triggers an exception that wastes an analyst’s morning. Operations teams stop chasing low-value noise and focus on genuine errors.
You’re probably wondering whether this is just moving some reconciliation processes up the chain. Can’t your existing technology handle it?
Three reasons it can’t.
The need for rapid time-to-value:
In a reactive reconciliation environment, any new or updated control requires a business requirements document, a slot in the IT development queue, a build cycle and a testing phase. All before a single transaction has been checked. For custodians, fund administrators and transfer agents operating under tight turnaround windows, that pipeline kills the point of a proactive model. Operations teams need to build and update controls themselves, fast, without a coding dependency.
The need for accurate matching:
Older reconciliation algorithms match on key fields. If the data in those fields doesn’t align exactly, the system flags it as a break. For fund operations processing thousands of trades daily – OTC confirmations against risk booking systems, fund flow data for transfer agency, position data across multiple custodians – that rigidity generates false exceptions at scale. The only way to grow a key-field matching system is to add headcount. That is expensive and defeats the purpose.
The regulatory clock is running:
The pressure to move reconciliation upstream is now operational and regulatory.
T+1 settlement is already live in North America. T+1 settlement is going live in Europe and the UK in October 2027. What we’ve learned from last year’s transition in the US is that overnight batch reconciliation just isn’t fast enough. Errors that currently surface in a two-day settlement window will instead produce failed trades, penalty costs and reputational exposure.
For funds operating under frameworks like UCITS and AIFMD – where inaccurate data feeding portfolio management systems, risk engines or regulatory reports can be expensive and slow to unwind – the case for proactive data controls is a regulatory argument as much as an operational one.
Build a control framework that actually puts you in control
Data errors will always happen. Even if your own operations are clean, you cannot control how external counterparties format and govern their data.
What you can control is the impact of bad data. A data-centric operating model lets reconciliation do its real job: stopping errors at the door rather than cleaning them up from the floor. STP rates rise. Operational risk falls. Headcount stops scaling in step with volume, because the work generating that headcount – manual exception processing, false break resolution, emergency corrections ahead of reporting deadlines – is no longer being manufactured by the system itself.
The firms that invest in this model now will stay ready. The ones that don’t will find out the hard way why the plumber was always the better call.
