TD Ameritrade’s big data push 1 yr. later: Benefits coming from all corners

Data quality is improving, personalization capabilities are emerging, and the pace of innovation is on the rise

1 2 Page 2
Page 2 of 2

That would take three to six months to do.  And only when the data was in the warehouse could they determine, “Yeah, this is good stuff, this is what I wanted,” or, “This is not what I thought it was.” 

So now, before anyone makes a request of the data warehouse, they must first have created a virtual view which takes days as opposed to months to do and they must analyze that data as it stands and determine whether that is actually the data they want.

If it does prove to be data you want, then you have to go back and submit a request in the traditional manner?

You’ve got a choice at that point.  The virtualization software has the ability to cache the data so you can add it to an existing data warehouse record, or you can say, “I’m absolutely convinced this is something we want to have perfected through the extract, transform and load process and we’re going to do that through the classical route.”  So you can either append or build it into the plumbing. 

That has saved a lot of time because you know how it goes.  There are always myriad requests to the data warehouse, and those folks get swamped and then you have to prioritize the requests and some get pushed to the bottom. Now we have a tool that lets people mine for nuggets, lets them prove the nuggets are there. 

Is it self-service?

Yeah.  It’s pretty cool.  There’s a bit of hand-holding, but over time it’s going to be more and more self-service as their skills get stronger. 

Have all these new capabilities led to some creative new thinking?

Very much so. The analytics teams are in each of the lines of business, and now they’re getting their hands on the data with these improved tools and starting to see these pictures emerge, so they’re coming forward with a lot of the ideas.  So it has absolutely increased the rate of innovation.

We’ve also integrated ourselves tightly with the technology innovation center in the organization, which was just getting started when I last spoke to you.  They’ve done a wonderful job of getting a strong team together, so now when anyone comes forward with a new idea we have this innovation center we can go to and they can very quickly put together platforms and bring in software and try stuff out, which has been a big step forward.  We used to put new ideas in a queue and eventually they would become projects and it would take six to nine months to do something. 

And that swings us back around to client. In the first six months I was here I did a detailed survey of all the business needs and opportunities.  Client data was one of the big things, but to address that we needed a lot of these other capabilities in place, so we decided to just muddle through and address it at a later stage.  Well, it is now that later stage and people are saying the biggest pain point they have is householding.  How do we do householding?

If Derek is married to Denise and they have a daughter, Joni, how do I connect them?  Right now all we know is Derek has some accounts and Denise has some accounts and Joni has one account.  We’ve got work-arounds but they’re pretty clunky.  For example, typically you would figure out that Derek and Denise are attached if they have the same address.  But in high-rise apartments in Manhattan thousands of people have the same address.  So we run into real constraints in some of these work-arounds and we need a lot more sophistication to be able to understand who probably constitutes a family without a lot of manual effort or going out to third-party information providers.

What we’ve also recognized is that we don’t just want to know that Derek and Denise have these accounts.  We want to know Denise and Derek are joint owners of that account, and that Denise has power of attorney on her dad’s account and has influence there. And then there are institutional advisors working with households, and we need to understand their different roles to really understand the full extent of the household. 

With our work-arounds, all we know about are the primary ownership relationships.  We have no easy way of viewing these joint relationships or power of attorney or any of the other relationships.  So when it comes to figuring out if, say, this household fits into our private client segment, how am I going to know that if I don’t have a clear understanding of the big picture?

That is what’s driving the need for a client master, where we can show that Denise is a primary account holder, she’s a joint account holder and she’s a power of attorney, and all three of those roles reflect under her as a client.  That’s become the big topic of urgency in the corridors.  What we’re busy doing now is collecting all the user stories around this.  In other words, how would we use this?  So we’re breaking it down into user stories which we then prioritize. This leads nicely into an agile development and implementation plan.

But to do this successfully you have to have a metadata repository, you have to have data governance, you have to have data quality tools, etc.  We’ve put all of that in place already. If we didn't have these already we’d be looking at a very long time to value.  It’s an exciting time for us because we feel like we’re in a perfect situation to very rapidly get value from implementing a client master.

 It’s most likely going to be a registry first, but over the course of a couple of years it will become the golden record that everyone will use.

And it sounds like you will realize interim value as you go.

We’re looking to get real benefit from this within six-months.  That’s why we’re using an agile approach and breaking it down into very small pieces.  Each piece is going to deliver some value.  Instead of releasing it into all of our client base, for example, we might just take a certain geographic region or we might classify a particular product or a particular set of securities and release that piece first and then build on and on and on.

Has upper management’s expectations of what you’re doing changed at all with time?

They understand it better.  So many of these things take time, and some of them are more visible than others, so there was a bit of a leap of faith on their part given the extent of the complexities of what we were dealing with.  I think they understand that much better now.  The great thing is they’re still highly supportive of the whole initiative.  They see it as being critical to the company’s growth. 

The other great thing about the company is, a lot of my peers in the industry spend most of their time trying to prepare for the regulatory waves that are hitting the finance industry.  They spend all their time just trying to get into compliance and very little time on innovation.  We’ve been able to balance it out about 50%/50%, and I am really happy about that.  When we started this journey three and a half years ago we weren’t really sure how it was going to pan out.  You never are.  As you well know, the devil is in the details. So I think everyone is feeling good about where we are now and how we are positioned to continue to deliver real value.

 

This story, "TD Ameritrade’s big data push 1 yr. later: Benefits coming from all corners" was originally published by Network World.

Related:

Copyright © 2016 IDG Communications, Inc.

1 2 Page 2
Page 2 of 2
7 secrets of successful remote IT teams