Trust is a Long-term Investment 


“If you are not able to produce trust in the data, then it becomes a liability and no longer remains an asset,” said Kamran Ahmed – Vice President of Data Quality at The Saudi National Bank.

Ahmed will speak about using analytics to help improve the data quality to fuel user trust and data utilisation in a data governance panel along with other industry experts at the Velocity summit in Saudi Arabia.

Excerpts from the interview;

How do we gain end user trust in data?

Trust is not easily gained. But when it is, it’s a valuable achievement for a company.  Sharing data requires a lot of effort not only from the data team but also the end user. It requires a lot of structural effort.

Attempting data acquisition doesn’t stop at acquiring the data. We have to track the data lineage and identify how much it needs to be processed. Data may have different values, formats and may need corrections to ensure it is consistent for further processing.

Data teams need to do something called data trimming which is a process of removing or excluding extreme values, or outliers from a data set. It may also require independent reconciliations.

Now that the data is usable, the team needs to make sure it is up-to-date, relevant and contextual so it can be used.

Trust is basically the long term definition. It is on the data team to ensure that there is consistency in what we do with this customer data because it is what the consumer agreed to, you can’t change what that means.

The consumer also needs to have the right to call out if their information is wrong, and there needs to be a feedback loop that relays that message back to the teams that can make the correction.

It’s a sanity check to ensure that the data is reconciled. Data can have variations. From the time your company started collecting data to now, processes and circumstances may have changed. This is where a human needs to intervene. Machines can’t make such connections as yet.

A human being needs to understand the problem, investigate and take actions to course correct it.

Transparency is another element that needs to be addressed. This isn’t simply about transparency between consumer and brand though it is essential to show them value against their data.

It’s equally important to have transparency between members of the data team and the rest of the organisation.

There needs to be transparency in how data flows within the organisation. To some extent, this is about data democratisation. There needs to be clear communication about the way data is ingested, processed and where the data comes from – the source and frequency with which this data is collected and updated.

If you are not able to produce trust in the data, that data becomes a liability and no longer remains an asset.

How can leaders use data to fuel innovative product development in finance?

It’s not only finance, any relevant department within the organisation that owns or collects data can use it to fuel their efforts. For example, the central government has information about a person. It knows which cars, properties, the family, earnings,  insurance, legal, etc.

A bank has information about a user’s legal transactions, spending behaviour, credit rating information and pulls more information from central repositories. If you combine all of that data, it’s the whole package along with how this data integrates into various systems within the enterprise.

Twenty years ago, people went to a bank to withdraw money, then to the ATM and how transactions are happening online. This includes remittance which can happen online while sitting at home. This digital journey has opened up the amount of data companies have to understand their customers. For a digital bank, there is a KYC that establishes whether a customer is legitimate or not. Every time a customer logs in, the customer ID which is validated by the government ID is verified within seconds. Similarly, when a customer asks for a loan, their history is available to make a decision. This history includes past transactions with other banks.

Integration is a big component. Blockchain has played a big role, especially for loan contracts. The pandemic accelerated consumer onboarding to digital products. What was earlier taboo is now the new normal. This opens up an opportunity for brands, but it also adds more complexity. It forced enterprises to innovate because suddenly the industry had to decide that it was okay for customers to complete their KYC via mobile. For example, AI is an added layer on top of data to make it more meaningful. Biometric recognition adds a secure layer to the KYC so that customers can have ease of access from different devices. It’s been a long journey.

Banks and corporates need to reinvent their old products and services to mimic what customers would usually go to the bank for online. But they need to add checks in place to ensure risk doesn’t escalate. Innovation is needed to  make the products robust and scalable.

You need business intelligence for decision-making. A basic requirement here is labelling of all customer information so it is easy to access and use across departments. This also means having governance in place.  You need AI machine learning for predictive analytics which helps in credit ratings and defining the risk.

Metadata, data catalog and data dictionary – how do you think these are the best deliverable?

Just imagine a physical warehouse, and somebody has to go in and bring something small like a pen. How would somebody be able to find that a timeframe with a specific output?

The data warehouse operates in the same way. We can see the terabytes, number of tables or number of files. We are producing more data than ever before. This is because of IoT and cell phone usage. This data is valuable but just like oil it has to be refined.

Methods like metadata data, cataloguing, creating a data dictionary and a business glossary help document the data to transform it into an asset. Data governance and data stewardship plays a wider role. When you add industry knowledge to data – it changes the value your products can offer customers. For someone who works in the loans department, they need to know where to find the right contextual data to make the right decision for the business.

The main challenge is that data is documented initially but when it is operationalised and many cross-functional departments are using it or updating it – there is no control over the changes the data is going through. In other cases, the data isn’t updated, and circumstances have changed.

Lineage documentation can help in understanding what the source of the data is, how it has been transformed and when. Documentation makes it easy to know whether the data is trustworthy or dated.

How will data democratisation help shift the organisation to the next level?

Let’s assume a scenario where a bank or a big corporation has invested billions of dollars in setting up a good data warehouse. They have business intelligence machinery in place but it’s being used by a limited number of people. The organisation is safe, it has maintained privacy but it hasn’t tapped the value. It’s all strategy and no execution. How do we change that?

The ideal end user in this scenario is the customer-facing team. This end user needs cross-departmental information to make a difference. Data democratisation empowers the employee or team to connect the dots between delayed delivery and friction in the customer journey. Ultimately, he or she is able to link processes internally with consumer complaints and offer a fix.

Data sharing within the organisation makes employees feel important– like a stakeholder in the business. Leaders need to equip their workforce with the right tools if they want to see innovation from the frontline.

There is definitely a possibility that there will be misinterpretation or wrong conclusions. It is a steep learning curve. Data democratisation comes with a separate investment in education and training staff to know how to use the data to inject insights into their work and to follow compliance protocols when dealing with sensitive data.

Data democratisation is mandatory in the current environment. There is a shift from the old to a new school of thinking. Business leaders are concerned about compliance issues but this has to be tackled via ethics and education, which needs to be a part of the culture of an organisation.

Tell us about your session at Velocity.

Technology leaders will get an action plan on how to produce the result from their data warehouse. It will explain how to plan for master data management and reference data management. This is contextual information based on my subject speciality.

It will also cover regulatory reporting in financial analytics. It will touch upon how data strategies need to be created for financial organisations so it is standardised and consistent to support business objectives and adhere to compliance.

To hear Kamran speak, register for Velocity.

If you liked reading this, you might like our other stories
Tips For Becoming A Successful Data Analyst
How Do You Measure The Success Of Digital Transformation?