In this article, we’re going to look at the problem with silos. In the financial world, silos are everywhere, so let’s have a look at how they originated, why some silos are necessary, and what the main issues are when they’re not necessary.
A silo can be a process, system, department, database or anything that operates in isolation; and when operating in a silo, it can be difficult or impossible to communicate and relate to other entities outside of the silo.
Indeed, the term ‘silo mentality’ arose due to teams working in isolation from other teams and the breakdown in communication between them.
Silos in the banking world
A good example of how this can affect everyday applications can be seen in banking. It’s common knowledge that silos in banking can be the enemy of good customer service, and a pertinent post on Chris Skinner’s blog discusses “the hand-off between customers with a deposit account, to their mortgage account, their cards, their loans, their savings, their investments and so on” and that they are all “segregated, separated and difficult”.
This practice goes way back to the 1960s and the consultants McKinsey; when they recommended such processes to Citibank – mainly to satisfy the wishes of big corporations that didn’t like the idea of dealing with a branch manager whom also dealt with the man in the street. This marked the beginning of the management of products.
Fast forward some years and the world’s premier bank became part bank and part insurance company, with an ever-increasing range of products to deal with. But because different types of clients still all used the common services of the bank (including the branch network and back office processing), so the product line organisation had to be imposed on top of the existing structure. Bankers now had multiple bosses in the form of product managers, and were suddenly working in a matrix organisation.
As well as the organisational complexity, the bank now had problems with mismatched computer systems and conflicting organisational manuals. When other banks saw the short-term profits generated, they followed suit and the silo problem throughout banking and beyond was born!
In the book The Silo Effect, Gillian Tett from the U.S. Financial Times said that information silos and poor risk management have cost global banks billions. Going on to point out that “risk managers didn’t know what was happening in some parts of the bank, and sometimes didn’t even know those parts existed. The traders who did understand had little interest in telling because they were making so much money, often because the bank’s accounting practices had mis-categorized financial instruments in a way that concealed their risk”.
What else creates silos?
There are many other barriers that create silos – think geographically; including country borders, languages and cultural differences; which can all lead to compartmentalising. So too can technical barriers, data barriers, industry silos (businesses working in isolation with no collaboration), and many others.
Why create silos?
We should remember that silos may sometimes be required or desirable. This could be due to regulatory issues where crucial customer data, processes or general intelligence needs to sit apart. Or where silos can create genuine focus and expertise, rather than a solution more akin to a jack of all trades and master of none.
There are myriad other reasons why technology and data silos may have been created, including the recent explosion of technical capability, new data insight and the birth of Fintech and RegTech – giving businesses far more choice. This capability includes increased and more efficient use of digital identity data, including social, device, IP and mobile. Add to this the growth and acceptance of automated document capture and Biometrics for on-boarding, and the improvements in the way traditional KYC data is matched and used.
To help today’s businesses manage identity and detect fraud, as well as comply with legislation, there are many, many options to consider. Legacy technology makes it hard to create the central view, so many ‘pile’ capability on top of other checks. Add to this a demand for cross border growth, and FinTech type businesses that may now have outgrown their original tech stack; and you can soon see the scale of the potential problem.
The main issues in data and technology
The financial industry continues to take a siloed approach to providing financial services, with financial institutions having to manage several partnerships to provide the full spectrum of services to keep their account holders satisfied. Data and technology silos produce many issues, but let’s concentrate on the main ones:
A lack of accuracy in making a risk decision, or missing a compliance or fraud risk can obviously have serious implications. In the banking world, let’s suppose that our customer has made applications and each time hits a soft risk flag – it’s a warning but it’s not enough to stop the application. Let’s suppose she has been flagged three separate times, but because it is across separate silos, there is no discernible link and therefore she passes.
If all three insights were in one single place, the application would be stopped. This is a basic example, but can apply to the lack of sharing from application data to transactional monitoring (imagine being able to immediately segment monitoring strategy based on insights gathered at application?). For fraud and money laundering, this lack of shared information is certainly dangerous.
Multiple decision points also lead to wastage and increased false positives as there is no data insight to counter the initial flag. In essence, fraudsters will prevail due to lack of insight being shared across the checking process, or indeed more widely across industries and borders.
Spiralling costs of IT
As we’ve seen, a lack of ability to quickly share data and insight across applications means incorrect decisions are made. It can also mean that it’s necessary to create manual processes to move the data and insight. However, silos require IT and development time and cost, and when information is held in established systems, updating those systems can be slow, expensive and sometimes impossible.
The constraints of legacy systems, over-reliance on techniques such as machine learning; escalating IT maintenance costs, and delays from managing multiple API’s can all rapidly increase business costs; and sticking to those systems means that there is no practical way to reduce those costs. When market situations change, updates such as plugins can take an age to implement, and this inability to react and change quickly is yet another danger of relying on silos. As they said in the wild west – it’s the quick and the dead in this game!
Lack of Agility
It’s clear that silos create both cost and a reliance on IT, but perhaps the bigger risk is the lack of agility that this leads to. Why can’t the owner of fraud or compliance make a rule change to adapt to a dynamic fraud threat? Why can’t I integrate this new reference file or dataset that I can prove will reduce my risk? It is because the current model of being reliant on suppliers and IT departments to integrate and change systems is outdated and leads to increased risk on a business.
Many companies tolerate the status quo because they can’t justify the case to move suppliers. This is another symptom of a business that is set up in silos; typically, the cost of internal resource to make the integration and configuration change outweighs the value of doing it – especially when compared to the value of the other projects in the development stack.
All in all, we can see that silos can be a huge challenge in any industry, and perhaps this is especially true in finance. Watch out for our part two of this article which looks at solutions that help fix the silos issue.
If you wish to explore this topic with us, please get in contact with the team on +44 (0) 20 3468 2719 or send an email to firstname.lastname@example.org.