Eldes Mattiuzzo is the founder of Bidu, a Brazilian financial services startup. A financial services veteran, Mattiuzzo founded Bidu 18 months ago to enable consumers to compare prices on a range of products, including insurance products and credit cards online. In this Q&A, he talks about the Big Data strategy at Bidu, his plans for the future and the challenges and opportunities for his market.
Why is Big Data important for your business?
Eldes Mattiuzzo: Gathering customer information is something that was part of my learning process when I worked in the credit card world. This approach of using data makes a lot of sense and it is very useful for the analysis process and for actually granting credit. We used past customer data also to know the likelihood of future trouble for some customers.
That was something I tried to apply at Bidu. We analyze the past of all the clients that are quoting with us and those who have purchased insurance through us, be it through the website or a phone call, then we try and identify the reasons, which are a combination of a massive amount of factors. We then build a model that involves variables that tries somehow attach a likelihood of purchase to this person.
We have a ranking system, so when the customer has just made a quote and is very likely they will buy, we call them straight away as the conversion chance is quite high. When it is the opposite, when there is hardly any likelihood that they will buy, we end up not calling them - an example would be a person who is quoting for a car that they are yet to buy. This model has worked very well and greatly improves the efficiency of the contacts that we make.
How has data analytics evolved in financial services in Brazil since the early noughties until now?
When I worked in the credit card industry around 2004 until about 2008, data analytics in Brazil was not as focused on "Big Data" per se as it is today. We had our databases, which were limited to what we had internally and that was often complicated. Today at Bidu, our data can be easily accessed by those who work with it - it is not like you need a programmer or an IT specialist to get that data. For example, our marketing team accesses that information all the time to do their job. That is a massive change form what it used to be.
Another big change is that we are not limited to the data that we collect ourselves - we have a myriad of data across all the social networks like Twitter, Facebook, Linkedin - so with a bit of extra resources we can extract the data that is also available on that world. And we use some partners here in Brazil who can also map the purchasing behavior of some of those customers. For example, a customer that buys items such as books, appliances or travel online is much more likely to buy insurance online than someone who has never used the internet. Of course, our own data is important, but we are not limited to it in order to make our conclusions around customer's buying behavior.
How was your journey around building your data analytics set up?
Nowadays , both Amazon and Google they have data tools that have an infinite storage capacity at a very low cost. So we started hosting our data with Amazon, because no matter how much data you store, the performance is always high with a maximum level of security. All the data that we collect from external sources and social networking tools that we compile is also stored on Amazon's servers and it is very simple and it's a very cost-effective option.
Given your reliance on Amazon and Google tools, what is your opinion about a possible government requirement that data be stored in Brazil?
I don't really believe that will happen. If it does, Brazil will take a massive backward step. It would be such a big setback for Brazil that small companies would be a lot less competitive. For that reason, I hardly think that a requirement like this will actually become law - I just hope that our leaders have the common sense to prevent it form happening.
When you started to put your Big Data strategy together, what were the main considerations in terms of tools and processes needed to make it all work?
We have a very big challenge here at Bidu which is the integration with the insurance companies, which is a very laborious process. Being able to get into the systems of the insurers in the appropriate and legal manner and present quotes in real time and according to the systems of the insurance companies is very complex. There were other companies that tried to do what we do, but because each quote request would take several minutes, customers just gave up. So our main concern was precisely this: capture data from all parties and ensure the processing works and the customer gets what it needs in a agile way.
We now work with nine different car insurance companies, each of which needs a certain set of data to present a quote. So if we were not organizing our processes in an intelligent way, our quote questionnaire would have more than 300 questions. So the practical challenge in technical terms was: how do I cover all the needs of all insurers with a minimum set of data? that was a very intellectually-intensive project. Another major challenge was convincing people of this quote comparison model: here in Brazil, this all very new, people don't understand the concept very well, so that's why our models to work out the likelihood of purchase are so important, to help us content people when the time is right.
What are your future plans around use of Big Data to bring in more customers at Bidu?
We still have a lot of work to do. With the amount of information we have, plus what we collect across social media tools we can offer insurance to people not just in a passive but active way. For example, reaching out to people who found us through Google. The next big challenge is this, finding prospective clients in a smart way: if the person has mentioned they are looking for insurance on Facebook, or asked on Twitter where they can find a good insurance plan. I want to be able to capture this information, use data to actively approach customers, as opposed to just passive. This is my biggest challenge at the moment.
How do you find information and exchange best practice around Big Data in Brazil, given the fact the market is relatively incipient here?
I think there is a lot of information and not a lot of knowledge about how to use data analytics. Here at Bidu, we do not think that we use even one percent of what we could use. There are several startups like mine where this is also a challenge. Of course, there are big case studies at large companies but for smaller businesses, everybody is more of less at the same stage of development.
We know there is massive potential in Big Data, but in Brazil there isn't a big success case that we can replicate - there is still some way to go to get to that stage.
Eldes Mattiuzzo, founder, Bidu
What is the biggest obstacle for business leaders trying to make the most out of their data in Brazil?
The cost of solutions is the smallest concern. The options available in the market today are very good and accessible. The challenge is how you capture external data - which data should be collected and how that information cant help you achieve your desired outcomes, I think this "how" is the biggest difficulty - how to make sense of everything that is available in emails, conversations and behaviors online and transform that into something valuable for the business. That said, I think there is a lot of opportunity for experts in statistics and maths to come up with solutions that can be very useful for businesses.
Do you have a question for Eldes Mattiuzzo about Big Data? Access our full education program on Big Data in Brazil and get access to a team of experts that can help you with best practice and specialist advice.