070 - Big data in finance with Bismark Adomako

March 02, 2021

Daniel speaks with Pluralsight author Bismark Adomako about how big data is impacting the financial industry, best practices for creating a big data strategy at your company, and more.

If you enjoy this episode, please consider leaving a review on Apple Podcasts or wherever you listen.

Please send any questions or comments to podcast@pluralsight.com.


Daniel Blaser:

Hello, and welcome to All Hands On Tech. I'm Daniel Blaser. Today I had the privilege of speaking with Bismark Adomako, a Pluralsight author and big data expert. We talk all about how big data has impacted the finance industry and mistakes that lots of organizations make when trying to adopt a big data strategy.

Well, thanks so much for chatting with me today Bismark. I'm really excited to talk all about big data and how it's transforming the finance industry. And I'd love if you could talk just a little bit about what the definition of big data even is. I feel like it's a term we hear a lot but what does that actually mean?

Bismark Adomako:

Okay, so big data is basically a term associated with complex and large datasets, right. A relational database cannot handle big data and that's why special tools and methods are used to perform operations on a vast collection of data. Now, big data enables companies to understand their business better and helps them derive meaningful information from the unstructured and raw data that they've collected. Now, big data also allows companies to take better business decision backed by data. Okay. Now, when we talk about big data you're basically talking about the five V's of big data. And that is volume, velocity, variety, veracity, and then value, right. I'll take all of them in explanation.

The first one, volume, is basically about the representation, the amount of data that is growing at a high rate, right. And here we are talking about terabytes and petabytes of data, right. Velocity is about the rate at which the data grows. Social media contributes a major role to the velocity of growing data. Now, when talk about the variety it refers to the different data types. Various data formats like test audios, videos, et cetera. And then the veracity refers to the uncertainty of the available data. Veracity arises due to the high volume of data that brings an incompleteness and an inconsistencies in your data. Here you are looking at the semi-structured form of data, right.

And the last thing which is very important is about the value, right. And the value returns to the turning data into something meaningful, the return on investment. Because if you take, or if you collect big data, you want to use it for something, that's the end goal. You are taking that data to process into something meaningful, right. So by turning asses big data into values, now businesses may generate revenue from it, right.

Daniel Blaser:

That's awesome. I love your use of the five V's. It makes it very memorable to focus on that. So now that we have an understanding of what big data is, we have those five V's in mind, can you give us a little bit of an idea about how companies across the financial industry are using big data right now to improve their processes?

Bismark Adomako:

I've worked in the finance company for almost five years now. When I talk about finance company I'm talking about the finance industry in the tech sector. So more like the fintech's and also the transition of finance sector, right. And what I've seen over time, what these financial companies are doing with big data in the first place and state, I mean the little, the few financial companies who are able to do it well, this what happens or these are the things that they do. Number one, first and foremost they are using it to expand their business intelligence, which is very key. You need to know more about your business in order to thrive. So now, when we talk about business intelligence it's mainly a technology driven process for analyzing data, for the sake of people who may not know what business intelligence is, and it has existed for a long time but big data has enlarged the capabilities of business intelligence.

Now analysts can use data both to get an overview of the past and to look into the future as well. Here with big data companies can mine massive amount of information and including findings from outside their own data sources, right. In addition to empowering the different capture and storage of large amounts of data, big data enables business of all kinds, which the financial institution is also apart, to analyze the data that they have and to look into the future, right. So the first part is expanding business intelligence.

Now the second part is basically about enhancing user targeting. Now, from any sector, right, any company deals with customers and it's about the users that they deal with. We want to target them in a different way or in a specific way, right. And each user's need or preferences is different from the next user. Big data has allowed for some pretty monumental changes in the way businesses target their consumers. Now, big data makes it possible to analyze a user's data footprint and then use the information to create a more targeted, personalized advertisement campaign. Okay. So just for instance, if you go onto Google, you just search for hotels, the next thing you realize is, okay, when you log onto a different website they're showing you advertisement of hotels, right.

Now, in the financial sector today data landscape means that most of our interaction with technology, our Google searches, our tweets, our Facebook likes and comments generated information that can be used to inform the type of [inaudible 00:05:58] that financial companies can give to their people or to their customers, right. Now, companies, for example, utilize dual targeting to learn about the location or businesses a customer has listed recently. Now, if your historical location indicates that you made several visits to car dealership, businesses can, which is one of the project that I did work on in one of the companies I consulted with, we use this information to target users, giving them discounts on car loans or loans. Yeah, for instance. Right. So those are some of the things that we do, enhancing user targeting.

The other thing will be to improve customer service, which is very key. Like you need to improve your customer base and your customer service, the way you relate to your customers. And when I talk about customer service, now the businesses or the companies normally refer to this as customer experience, right. And it stems from when the customer walks in or even sees your brand to the next step that they try to work with you, right. So implementing technology that uses big data means businesses can address customer service concerns in a timely manner through the use of things like chatboards, artificial intelligence systems, work accustom to interaction with users through virtual needs, right. So imagine that we able to get help, quick help and without necessarily having to speak to a contact center agent, right. Getting this feedback on the go whenever you want it, right.

Now, the other point is increasing efficiency and reducing cost. Now I've done work over here which is very important to talk about, right. So big data is also used to improve operational efficiency. Now, it enables business to gain insights by analyzing the various data sources when it comes to manufacturing, for example, a little bit away from the financial sector, big data allows companies to analyze these things like production, feedback, customer feedback, and product returns, to determine the quality of production and overall profitability. Predictive analytics can also be used to increase production efficiency when businesses reduce outage by anticipating future demands. Okay. So let's start with the increasing efficiency and reducing costs.

And next thing I would like to talk about, last two things actually, an influencing customer behavior and also protecting against operational risk. Let me elaborate more on protecting against operational risk. Operational risks refers to potential loss related to human dependencies and error. The first thing that we're looking for, or this may include things like fraud, computer hacking, reactions to catastrophes' and failure to adhere to internal policies. Big data can help. When it comes to cyber security, for instance, big data enables analyst to examine, observe, and also detect irregularities within a network. This reduces the time it takes to detect them and then resolve such issues. Big data utilizes two fraud detection techniques to protect against potential risk, statistical techniques and also artificial intelligence. From matching algorithms, to detecting anomalies, to using machine learning to automatically identify characteristics of fraud, bring data has done a lot to help business in all industries, of which the financial industry's a part of. Right.

So these are the main things that I would like to talk about when it comes to how companies across their financial industries that have consulted for, have, or are using big data to solve their problems.

Daniel Blaser:

Yeah, you touched on a lot of great things there and definitely a lot of what you mentioned it sounds great. It sounds like there are a lot of benefits using big data within the finance industry. But I guess my question is, is big data always better data? Or what are some of the downsides of some of these approaches you talked about?

Bismark Adomako:

Okay, so, well this may be a little bit debatable, right, but in my view, big data or the bigger the data, the bigger the potential oversight. Now, I use the word carefully, potential oversight, here. Currently, most companies do not have structures in place for collecting these amount of data, right. So the few who do, I've seen do it well, right. And those that are coming up sometimes do get this oversight of the real type of data they should get. Now, if you do not have a precise understanding of free data sets you need to analyze you won't know where to start, which is very important. Where to start is very important. And it may actually be better to examine the issue on a smaller scale if you don't know where to start, right. So going for big data may not always be a case or big data may not always be that it is better data. It depends on where you're starting from, what you have and what you want. So if you don't have that in place, big data may not mean better data.

You don't need a data leak, for instance, to understand the sales of umbrellas and why they're spiking during the raining season. But no data leak will ever help you to understand this unless you tell it that, okay, it is raining and this is the reason why it's raining and these are the consequences of raining. That is people are basically going to order for new, more umbrellas, right. Now, if data misses assumption or contains errors, you are in trouble. You need an intelligent model to sift through the data and work out what is generally useful. Quantity does not always equal quality.

Daniel Blaser:

Yeah. I think that's really important. Like you said, I mean it makes sense, but you take something like big data, which is definitely a trend, it's definitely something that a lot of people are paying attention to, and I think it's important to emphasize, like you said, just jumping on the trend is not enough. You have to be able to have a good way to analyze that data and to make sense of it. Otherwise it's just a, it just can be kind of a mess. And that actually brings me to my next question. As we were speaking about the finance industry, specifically, is financial data just too messy? Does it provide it's own set of challenges in relation to big data because of the nature of finance data?

Bismark Adomako:

Yeah. For finance data, right, due to the fact that it has always been a traditional way of keeping data, like in formatted or in schema type tables, right. Now, but even before that let's face it, a lot of enterprise data is in a bit of a state. Now, companies have, or haven't actually needed to put in a lot of work at this scale previously but their whole raft of tools to help clean the data that you are taking up. Now when you take a data management and data governance softwares it also helps in capturing and cleaning the required information so that only the useful stuff stays in the frame. With analysis and visualization tools becoming more and more user friendly, more people than ever, today, are beginning to get more insight from the data that they have.

There is lot of bad data out there, not just the financial sector, but there are many ways to combat it. Right. And it all stands from you getting the right resources. When I talk about resources I'm talking about the people actually to do this cleaning for you, to create the data pipelines to move it from one place to the other. And also the data architect to going forward, prescribing the right way or the right type of data to be collected. Bad data is everywhere, right. Now you knowing exactly what you want helps you get the right data that you want.

Daniel Blaser:

So I guess my other question, you talked about all the advantages of big data within the financial industry. If someone listening to this works in the finance industry and maybe their company hasn't embraced big data the way that they wish they had, have they kind of missed the boat? Are they kind of now playing catch in the world of big data?

Bismark Adomako:

Too many companies assume that as everyone is talking about big data, everyone is taking advantage of big data. Now, that now produced a survey some years ago showing that only 13% of large companies, right, even the large companies themselves in the survey had actually deployed a big data solution. This may have increased by now but many more are planning investment in the area of big data. But you will be far from late to the party. Now the party is just starting. What I will advise is that you start small, prove this concept with an easy early win and then start to build your big data team, right, for which is very important. So start from somewhere. If you don't start you wouldn't know the right things that you need to get, right. You may be reading a lot of things online but when it get to doing the actual work it's something different.

Daniel Blaser:

What are some recommendations that you would give to kind of cultivate big data adoption in a company?

Bismark Adomako:

Well, so when it comes to recommendation, right, I would say, first of all, commit an initial effort to customer-centric outcomes, right, which is very important. Your focus should be on the customer, right. How is our solution or our proposal going to help the customer in the long run, right. If you put the financials at the back of a head you, trust me, you will lose out, right. But if the whole project or the whole idea of big data is about helping the customer, that is when you realize a lot of profits, right. Now, the next thing to do is to start with an existing data to achieve a near term result.

Now, I've seen people or companies trying to do big data, right, and they themselves, the company themselves, they take the data that they have, it's not that much to write home about. So what they try to do is contacting the telecommunication industries, contacting outside sources to seek for data that they're trying to join with the data they have in-house in order to create something. That may be an overkill from the beginning, right. If you don't have structures in place to collect data and then also house data in a way that makes it easier for you to report on it, it become very difficult for you to do, right. So the second thing that you would like to do is to start with an existing data to achieve a near time result. If you're able to say that, okay this is the data I have in-house, it's how I'm using this data to help my customers. Then that's it. The focus, remember the focus is on your customers. Okay. So when you're able to help your customers, you're able to grow forward, right?

And the next thing is to define a big data strategy with a business-centric blueprint. It's very key. Designing or defining a data strategy is very key. If you don't have a data strategy, as a company it becomes very difficult for you because if it's only one person's idea of pioneering or spearheading this data journey, it becomes very difficult when the rest of the company are not on the same period too. People tend to sabotage what you're doing because they may think that you're taking their job from them because once you gather data it allows for automation. And once you automate a lot of these processes people feel threatened, right. So define a big data strategy. How are we going to do this? How is it going to affect our initial operation? And all that, right.

And the next thing to do is to build an analytics capability based on business priorities, right? So what, as an organization, what is the analytical capabilities that you need? Do not go in looking for data scientists, data architects, when all your company needs is basically a simple, descriptive analysis of what they have currently, right. So start from there, define a solid analytic capability based on what the business needs. Remember the keyword here is based. Based on what the the business need. If you're able to do that, you are able to add onto this once you've built your case that this is what we've done with the little that we have. You'll be able to move forward from there, right.

And next thing or the last thing, basically, is to create a business case based on the measurable outcomes, right. After you've done all these things, right, after you've built the analytical capabilities and achieved some momentum, you are able, or to help you to create a business case based on the measurable outcomes or the output of what you did, right. You are able to tell that to management, to pitch that idea to the whole company and then tell them that this is the journey that we want to [inaudible 00:19:54], we want to embark on. And it will also help them appreciate the whole strategies that you've put in place. Right. So these are the recommendations that I personally will give to any company, whether big or small, starting to form any big data team of sort.

Daniel Blaser:

Yeah, I mean, a couple things that stuck out to me from what you just said is the importance in creating that plan and having a lot of people sign onto it. So it's not just one or two people or one team trying to force something, you've got to make sure that you have buy-in across the organization. That definitely, yeah that makes sense. And then the other thing that I like that you said is just reinforcing how this strategy will connect to business needs. It's not something you're doing because it's just cool or because maybe it will help down the, in the future, it's something that actually can serve business needs right now. As companies are maybe going through this process, they're drying up their big data roadmap and they're making all these plans, like you mentioned, what are some common mistakes that tend to be made?

Bismark Adomako:

Well, you see the very common mistake that is available or I've seen happen is that companies that are trying to implement any big data solution or big data strategy, first of all, those that fail they don't have a big data roadmap or a strategy in place, right, or even if they do it is poorly defined, right. And I think that is basically a recipe for disaster. If you don't have a roadmap, well there's this saying that says, "If you don't have any destination, anywhere is actually a destination," right. So it's like you're in a car or you're in a train, you're going somewhere, now any train stop that you've stopped is actually your destination because you actually don't know where you're going to. Right. But if you have that strategy in place, I think it's something I should have talked about, right.

So, and in adopting a strategy you should try to look at how to educate to explore, to engage and to execute. There are basically four E's over here. So we said this another thing I'm putting a clause, four E's. Educate, that's building a base of knowledge of big data also to explore defining the business case and then the roadmap. Then to engage is to embrace in the big data roadmap or strategies that you've put in place here. To engage you bring onboard the customers and also the employees and then other people or stakeholders in your business, right. The last part is to execute, now implementing the big data at scale, right. Here you've tested and then you have a lot of things that you can actually boast about or you can actually showcase to your end users, then you execute. Now, at each adaption stage the most significant obstacle to big data effort reported by the banking and then also the financial markets that I've worked on is the gap between the need and an ability to articulate measurable business value.

Okay. Now, executives must understand that the potential or they have to realize business values from big data strategies. Now they also have to pilot and also do the implementations themselves, right. The executives must be involved in the whole thing, right. You can't not just put this in the hands of just one particular team, right. And also the big data team or big data strategy should be a part of a higher strategy, the strategic team of the company, right, because it's there to inform and to advise, right.

Now, organizations must be vigilant and articulate values that they need to use. Now, if you forecasted this values based on detailed analysis when needed also tie the pilots result with possible outcomes, right. And also for executives to commit to the investment in time, the money, and then the human resources needed for this journey, right. Now, when this fails, all these things that I mentioned fail, basically your implementation of big data is very likely to also fail if you don't have the things in place. And I also, I'll mention it again. First, educate, second, explore, third, engage and lastly you have to execute your plan.

Daniel Blaser:

Expanding beyond finance for a second. What advice would you give to anyone that wants to master big data?

Bismark Adomako:

Well, okay. So this advice will basically go to both individuals and also to companies, right. So the individuals first. Now, if you want to master big data or any data related sector, do not go in all out with your hands open trying to do things like data science, machine learning, AI. You have to know the niche where you are, where you want to get to. So if it's the financial sector that you're going to, for instance, you have to know what are some of the things that are needed in the financial sector, right. Is it accounting, right? Well, if it is accounting what are some of the things I need to know about accounting? Okay. Where exactly, which field in the data, in the big data field am I going into? Am I looking at going into data science? Am I going into data engineering? Am I doing machine learning? Right. You should have data back of your head, but am I going to do data analytics? You have to know all these things before you go into it. Don't just go.

Because there currently there a lot of courses, there are a lot of information online about big data and if you don't take care you'll get yourself so soaked in all this information and the long run, you've really not made any impact in your pursuit to be a data expert, right. And also anyone at all who is trying to embrace or to go all in, in big data I'll advise that you should understand data engineering, right. Data engineering is very key. And when I talk about data engineering I'm not talking about database management, I'm talking about data engineering. The process of creating pipelines, the process of working with data, moving data, manipulating data and cleaning data in the right form, right.

I wouldn't say that be an expert in that field, but if you're an expert in that field it actually gives you a very big advantage, right. Because currently there are companies who are recruiting data scientists but then they realize that, I think this advice will also be to the companies, right, you realize that they basically have data scientists and then the data scientists are there waiting for data to work with, right. That is the issue I've also seen with a lot of companies, they are recruiting the so-called luxurious roles, right. And you ask them, "Okay we want to do prediction, we want to do this, but do you have your data in place?" Right. You should have that structure.

Now, I keep telling people that if you give me a budget to recruit for ten data aspects, right, I'll probably recruit eight data engineers and two data scientists, or eight engineers, one data scientist and one data analyst. Right. My idea is that when I have these data engineers in place, whatever they're doing, they know what they're doing, they are going to eventually or transition to become data analyst and data science moving forward because I want to have my data and I want to have it as quickly as possible in order for the data scientist to work on it. It reduces the burden of the data scientist and the data analyst and also keeps the business moving whilst you're reporting and also doing predictions for the business, which is very key.

So all the companies out there, please, when you are recruiting or you are looking out for data experts try as much as possible to concentrate for now on a lot more on the data engineers. Yeah. And just give some more respect to the data engineers because they are going to make the work of the other data sectors in your team very easy. Right. So I think that would be an advice to individuals trying to filter into data, big data, and also companies trying to be on that roadmap.

Daniel Blaser:

Is there any resources that you want to plug on the individual side, if you're trying to get better understanding of big data? I mean, obviously this is a Pluralsight podcast so maybe it's a softball question, but what have you found to be the most helpful resources to get that understanding of big data?

Bismark Adomako:

Okay. So the most helpful resources will be the person's enthusiasm and then the type of mistakes that a person makes, right. It's very key. I wouldn't point you to any specific course, right. And even if I will point you to a course, I definitely am going to point you to Pluralsight. There are a lot of courses on Pluralsight. I have added a course but it's a data science course. It's not really about big data, at the core of big data, right. When you talk about big data you are talking about the Hadoop, the Spark and all that, right. It's about data science on Azure machine learning, using Azure machine learning, right. You can take a look at it. The name is Bismark Adomako. So you can basically take a look at it. Now, the mistakes are very key. You have to get dirty with the data, right. You have to get dirty with the data. You build the experience as you move on and as you move on you get that experience every time you solve new issues.

Daniel Blaser:

Yeah, that's great. And I will include a link to your course, your Pluralsight course in the show notes to this episode. So they can take a look at that. But I love that you also mention, include that a person's curiosity and their drive to learn and work on their mistakes that that's a really important resource too. A lot of times we overlook those things. Yeah. Thanks so much for chatting with me today Bismark.

Bismark Adomako:

Okay, bye.

Daniel Blaser:

Thank you for listening to All Hands On Tech. To see show notes and more info visit pluralsight.com/podcast.