Generate now!

Synthetic data for banking

It’s time to put your data to work for seamless online banking, personalized services, improved business efficiency, fraud detection, and full, automated data privacy-compliance
Download the banking ebook
15%
Improvement in AI performance
$10M+
Annual savings on data provisioning​
#1
Mobile banking app in Europe

Why do banks and financial service providers need synthetic data?

Cumbersome legacy systems, tightening regulations, growing security concerns, the inability to move, share and scale data to drive data-centricity and innovation are familiar problems in the financial services sector. The winners will be able to create a competitive advantage by solving those challenges with synthetic data.

How can synthetic data help?

Banks need to leverage flexible, discardable, and most of all privacy-compliant synthetic data products to serve a variety of internal business functions. Synthetic data is not personal data and is free to use, share and hold.

Synthetic Data for Banking Resources

948,170,179,
Ebook
REPORT: Executives must utilize synthetic data to meet their AI and analytics challenges

Enterprises find themselves amid opposing forces. Be data-driven or fall behind, comply with regulations or face fines, go private-by-design or lose increasingly privacy-conscious customers, scale AI or be overtaken by those who can. Synthetic data offers a versatile digital transformation toolkit that lets you tackle these challenges. This report by Harvard Business Review Analytic Services in association with MOSTLY AI provides actionable insights on:

  • Training AI using synthetic data
  • Accelerating product development & testing
  • Enabling faster, privacy- compliant data sharing
  • The business impact of synthetic data
  • How to start utilizing synthetic data
Read More
Ebook
How innovators outperform competitors in banking

The winning strategy in banking is to accelerate digital transformation. Synthetic data plays a vital role in this process. Synthetic data enables the development of customer-centric products, mitigates data leak risks, and improves the performance of AI and machine learning models.

This ebook includes the best practices you need to maximize your synthetic data opportunities such as:

  • democratization of data access
  • streamlining privacy compliance
  • minimizing data retention
  • developing customer-centric products
  • and many more.
Read More
Podcast
Privacy blindspots in banking with Amir Tabakovic, Mobey Forum

Alexandra: Welcome to The Data Democratization Podcast.  I’m Alexandra Ebert, Chief Trust Officer at MOSTLY AI, the category-leading synthetic data company.  And here with me, I have my co-host, Jeffrey Dobin, who is a privacy tech expert from Duality.  Hi, Jeff.

Jeff: Hey, Alexandra.  I hear that you found a really interesting person to interview.  So, who is this person with all these data stories?  I hear he’s known as the “Data Whisperer”.  Really curious to learn more.

Alexandra:

Jeff: Sounds like he has an interesting perspective and a great background.  So, what types of people would find this episode particularly valuable?

Yes, the Data Whisperer, also known as Scott Taylor.  Scott has been working with enterprises for decades and advised them ever since on strategy and data management best practices.

Alexandra: Well, Scott has some really insightful thoughts around making data management visible and how to get buy-in from executives for your data project.  So, I would say that if you are in any way involved in making data projects happen, then Scott’s advice for crafting good data stories will really come in handy for you.

Jeff: Well, I’m all ears.  Looking forward to hearing this episode, so let’s jump right into it with the Data Whisperer.

Alexandra: Hi, Scott.  It’s so great to have you on the show today.  Where does this podcast find you?

Scott: Hi, Alexandra, good to see you.  I am in Connecticut, in a beautiful place called Black Rock, on a lovely harbor.  It’s a gorgeous day today, so nice to be here.

Alexandra: That sure sounds super nice.  So, can you introduce yourself to our listeners?  What’s your background?  And how did you end up becoming the Data Whisperer? 

Scott: How did I end up becoming the Data Whisperer?  Good question.  So, I help calm data down.  That’s what that moniker means.  We’ve got to tame data.  We’ve got to structure data.  We’ve got to manage data.  So, the Data Whisperer, à la Horse Whisperer, Dog Whisperer, that seemed to be a decent moniker, but I’m 25-plus years in the data management business, always from the strategic business side, and worked at a number of iconic data companies, like Dun & Bradstreet and Nielsen, WPP and Kantar.  And I’ve always assisted the business or the IT side of enterprises in helping explain the value of proper data management at the strategic level.

Alexandra: That sounds super exciting and as if you have vast experience about data in today’s world.  One thing that would be of particular interest to me today, if you look into discussions, you hear hot topics like data science, AI, advanced analytics.  But if I understood correctly, you were advocating more for the data management side and really tackling this first.  Why is that so important?

Scott: All those other things that you mentioned, AI, ML, data science, that’s where data ends up.  I focus on where data begins, where that data journey starts, and if you don’t start the right way, you’re not going to end up in the right place.  That’s pretty simple.

Alexandra: Makes sense.

Scott: Straightforward way to think about it.  I look at the data space kind of bifurcated into two big buckets.  There’s the business intelligence side and there’s the data management side.  Data management in my view is determining the truth in data.  Business intelligence and those other activities are deriving meaning.  People want truth and meaning from data, but it’s not chicken or egg here.  It is egg and omelette.  You’ve got to start with the truth.  Obviously, I’m on the truth team, 

Alexandra: Yeah, I wanted to reference that.

Scott: Right.  Before you derive any meaning.  The meaning part is much sexier.  People like that better.  It’s very - that’s where all the exciting stuff’s happening, but you got to start somewhere, so that’s where I’m advocating.  Start correctly and your data journey will be better off.

Alexandra: Yeah.  I agree it’s the more sexy part, but how do we actually get truthful data?  What are the steps to achieve that?

Scott: You know, all around, I mean that’s part of the foundation of what master data, reference data, metadata, data management, data governance, data stewardship, all these foundational activities that help organizations create and manage and distribute that core content.  How you do it, there’s lots of different techniques.  I stick with the “why” part.  So, I’m much more of a “why” than the “how”.  And the “why” for me is that kind of data, master data as an example, the most important data any organization has.

Alexandra:  Yeah. 

Scott: People don’t often think about it that way.  But all the other data you have in an enterprise tends to be about what the master data is.  Customers, vendors, partners, prospects, product services, assets, employees, you know, relationships, and brands.  Those are the macro domains that I think people need to focus on.

Alexandra:  Scott, you’re also an author.  Your book is called “Telling Your Data Story”.  And in this book, you mentioned the three Vs: vocabulary, voice, and vision.  What are these three about?

Scott: Yeah, the book’s been going great.  “Telling Your Data Story: Data Storytelling for Data Management”, which we can talk about the differences there, if you like, in a moment.  But those three Vs are just an approach that I share with people to help them structure the narrative they need to create about why managing data’s so important for an organization.  So, obviously, a knowing wink to the three Vs of big data.  But my vocabulary, the words you use; voice, the way you talk about it; and vision, why it’s important.  And that simple framework helps guide people to organize their many thoughts and ideas around explaining the value of data management to the business side.  But you got to use the words of the business. You’ve got to harmonize to a common voice.  Harmony doesn’t mean everybody sings the same notes, but it does mean they sound good together.  Does your entire team know how to describe this to their business side, for instance?  And the vision, what’s the vision of data management?  The vision of data management should be to enable the strategic intentions of your enterprise.  Where is your company trying to go?  What kind of data does it need to get there?  And data management always has a role in that.  Even though they’re not always invited to the party, they need to try and get into that room one way or the other to show why the work that they’re doing is going to help the company to where it needs to go. 

Alexandra: Understood.  So, you say storytelling is really the most important aspect to get buy-in from the business side in achieving proper data management.

Scott: I think it’s part of it.  You got to reach out and connect with people on an emotional level.  If you just show them a bunch of bad data or, you know, your data quality scores, or if you’re - even if you’re in the business intelligence side, if you’re just showing charts and graphs, that doesn’t tell the story.  That’s just a whole bunch of numbers or stats or reports.  You’ve got to put it in some form of narrative to connect with especially the business side because people tend to get way too technical.  The reason I make this distinction about the “how” and the “why” is many data practitioners, no matter which part of the space they’re in, spend way too much time explaining how they are going to do it or how they do it.  And business people don’t care about that.  I don’t know a business executive who cares about the “how” until they truly understand the “why”, why this is important, the rationale for doing this at all.

Alexandra: That makes sense.  Do you have some type of success story to share with our listeners, where somebody you consulted was focusing too much on the “how” and then learned to frame the story around it and eventually achieve the goal that he was after?

Scott: I think almost all of them, they all start with, you know, showing me their data model.  “Here’s our data maturity score.  Here’s our, you know, API.”  All these other technical things.  I just stop and say, “Okay, you got five minutes with your CEO.  What approach would you take to explain it to them?  And if you show them your data model, you’re going to get kicked out of the office.”  But even just this week, I worked with an organization - because I work with enterprises to help them construct this.  And this organization was - it’s a global bakery organization named - at the simplest level, they make bread and pastries.  And they’re a huge company, and they were launching a new initiative to really introduce a whole bunch of enabling technologies that they’ve got for the field level, for the market level, all this different software.  But the core of it was the need for all this high-quality data.  It seemed a really straightforward approach to take the idea of baking and bread-making, and the end result of that being a wonderful sandwich.  And talk about your goals you’ve got in the business are this sandwich.  You want to leverage data better.  You want to be more data-driven.  You want digital transformation.  You have all these terms that they had internally.  That’s the end goal.  That’s what everybody sees, the sandwich.  Now we all know you can’t make a sandwich without bread.

Alexandra: True.

Scott: To bake the bread and all that manufacturing, those are your enabling technologies.  But before you can even make bread, you’ve got to have the proper ingredients.  Flour, water, salt, you know, whatever, yeast.  Those have to be measured correctly.  They have to be stewarded.  They have to be governed.  The recipe has to be accurate.  So, very quickly, in a very simple analogy, it was, if you want a great sandwich, you need bread.  If you can’t bake bread unless you got good flour.  Simple.  You know, that’s how I would describe on a - you know, that’s not even a conceptual level.  That’s almost a symbolic level.  But those things stick with people.

Alexandra: Uh-hm, sure.

Scott: One of the bigger pieces of advice I give to folks is use the imagery that already exists in your business.  I wouldn’t use that bread analogy with, you know, a different kind of manu - an auto manufacturer or a healthcare company.  So, you want to try and leverage the imagery, the words, the approach, even the slogan of your own company and focus on that sector and category to find all your analogies and ways to talk about it.

Alexandra: Yeah.  So, give people something to relate to, basically, and make it understood?

Scott: Yeah.  Because, you know, the whole company understands what it does.  How do we make data alive for that company in a way they already think about it?

Alexandra: That’s a good point.  You also mentioned that you work a lot with enterprises.  What would you say are the biggest challenges in their data strategy and also in the journey-powered digital transformation?

Scott: I found that the challenges they have are really at that foundational data level.  They’ve got - people have plenty of the vision.  They’ve got plenty of objectives.  They have plenty of things they want to do.  And then they get hung up because the basic data isn’t there or isn’t structure or managed or governed in a way that will allow those things to happen.

Alexandra: Uh-hm. 

Scott: If I look across all the kinds of businesses I’ve ever talked to and in my career, I’ve had the opportunity to really speak and work with every kind of company in every category, in every stage of maturity, in every type of geography.  That’s pretty broad but at the enterprise level.  And even though there’s lots of different sectors, I’ve found that the challenges they have are actually more the same than they are different.  You got all these multiple silos.  You’ve got disparate data.  You don’t have an internal standard.  You lack common definitions.  These things don’t connect.  So, what I find really fun is talking to a financial services company in the same room as a healthcare company and having them - you know, they know they’re in totally different businesses, but having them realize if they talk to each other, they’re going to have data problems that are more the same than they are different.

Alexandra: Yeah.  Yeah.  That’s interesting.

Scott: At that level.  Yeah.  What they call it is where the differentiation happens. 

Alexandra: Yeah.  Yeah.

Scott: You know, a healthcare company calls them “patients”.  You know, a banker calls it an “account” or a “client”.  You might have people who call it a “customer” or “vendor” or “partner” or “prospect”.  You know, you’ve got different terminologies, but they’re all still relationships.  Every business has relationships.

Alexandra: Absolutely.  So, if you say that it starts out - or oftentimes the problem’s started the basis if a company wants to get from 0 to 100, where to start and which steps to go through to achieve good data management?  Of course, on a high level so that we don’t extend the time we have available in this episode.

Scott: You’ve got to get some kind of senior-level support.  You got to get sponsorship at the higher levels because a lot of times moving into data management across an enterprise takes some form of cultural change.  You got to change the way people, not just react and deal with data, but how they input it.  Have them understand why it’s so important to, you know, use a drop-down instead of free-texting in a customer name as a really simplistic example.  But those kind of changes, if you want to instil that new data culture, it’s got to come from from the top.  So, one thing for sure is explaining in a way to the higher levels that they really have the buy-in.  And then you get right into the “how”.  A lot of the, you know, how do you organize the team?  How do you engage with the business stakeholders?  How do you deliver that ongoing value?  Those are all things that I think people are well-practiced in, so I add that layer of “You may have all that, but if you can’t explain it to your board in a way they understand it, you’re not going to get the support you need.” 

Alexandra: Yeah.  Makes sense, makes sense.  Do you also have some data management success story to share with us?  So, our listeners are especially interested in success stories from the financial services industry, insurance, or maybe also telco.  Is there any success story that you can leave us with?

Scott: I’m trying to think of an example of the particular client, but I mean, you know, you see success in the customer experience.  So, if you’re a financial services company and you don’t - you know, and you have completely household-ed data about most, if not all of the interactions and account types for this particular customer interaction, then you’re succeeding, then you’re moving towards this idea of being relationship-centric.  You’ve got everything you need to know about that relationship accessible to you in that moment to be able to deliver the value you’re looking for, to try and deliver to that customer.  On the flip side, if you’re a customer or consumer of one of these services and you’re constantly telling that service what you do with them.  Or “Oh, I also have this other account, but it’s in my wife’s name, but it’s household-ed and it’s connected.”  Or “This is my husband’s, you know, account, but it should be connected in some way.”  Or “We have - you know, I’ve got business.”  I’m making all this stuff up in terms of the complication, but those frustrations happen all the time to us as consumers or as clients of some of these services.  And I think it’s becoming a real differentiating factor.  If I have to explain my relationship to the entity that’s supposed to be servicing me over and over again, it’s not a very good relationship.  And being from a data background, I kind of know where those five - you know, you call somebody, they don’t have their stuff in front of you, you got a data problem, don’t you?  So, it’s, you know, not a specific story about a specific company, but we feel it.  A lot of times, you know, people don’t see good data, but they feel it.  They feel the result of that especially when you’re interacting with a customer or in some kind of relationship.  They go, “Okay, these folks really understand me.  They know that I called last week about this thing and…”  Whatever that happens to be.   

Alexandra: So, we could say data management is crucial to delivering excellent customer experiences.

Scott: Exactly.  I will quote you on that.  Data management is crucial to delivering excellent customer experiences.  Absolutely.  Think about customer.  What’s the data about that customer?  Do you have the right data?  Do you have it uniquely identified?  Do you have a hierarchy?  Do you have a segmentation that’s correct?  Do you understand the geography?  Those four basic things, which I actually elaborate in the book and I talk a lot about, I call those the four Cs: code, company, category, and country.  You need a unique code.  You need a hierarchy.  You need a taxonomy for segmentation.  You need a country or some form of geography.  You get that stuff right; a lot of data problems go away.

Alexandra: Yeah.  Yeah.

Scott: A lot of them.  Duplicates, most of the challenges people have in reporting, integrating data, aggregating it, interoperability happen at those dimensions.  The entity, the hierarchy, the category, and the geography.  Those four turn up in almost all report.

Alexandra: Yeah, can you elaborate on that?  Can you give a few examples of problems that you usually see in enterprises and how they relate to a basically data management problem?

Scott: Starting right with first one, code.  You know, every record needs some sort of unique identifier that will allow you to validate that entity and know it exists and un-duplicate it.  I mean everybody’s got duplicates.  It starts there.  If you’re doing a customer churn model with all this fancy AI and data science activity and you’ve got a duplicate record for that same customer, then you’ve got two different histories or two different views of the same thing.  And you might send a promotion to one view of that because it looks like there’s not a lot of activity when in fact it’s a duplicate of another one that’s got a lot of activity.  So, talking at a highly conceptual level here, but duplicates plague every kind of database.  And, you know, it feels really kind of clerical and boring and back office.  Why should, you know, a CFO or CEO care about something like duplicate data?  Why?  Because you don’t have a true view of your customer or your product.  You know, moving along those Cs, if you don’t have the right hierarchy structure, you’ve got a sales person going to one division of a company.  They don’t realise they’re related to another division of a company.  The bigger your relationships, the more structured and levels there are.  As a supplier, you might work at different, you know, bill-to, ship-to, plan-to, sell-to, all these parent-child relationships are inherent in anybody’s, kind of, you know, way they manage activity.  Segmentation, market share.  If you don’t have a clean category structure, you’re not going to get - you’re not going to be able to do targeting.  You know, “We’re really good with a certain kind of company.  Let’s find more of those kinds of companies.”  Or if you don’t have that kind defined, you can’t do that kind of segmentation.  You know, every single client I’ve ever looked at, when I looked at their database, they have a category called “other” or (overlapping conversation).  No.  Or something. They had a attribute value that said “DK”, and we asked them what that meant.  It’s stands for “Don’t Know”, so it’s same thing, you know.  And it may seem trivial and you just, you know, “Oh, I won’t fill this in.”  But it reverberates across the rest of the data you’ve got.  Your top 10 category cannot be “other”.  And, you know, I’m thinking you’ve got an account-based marketing program or segmentation program.  And then finally geography throws all kinds of people.  You know, you got a sales market.  You got a media market.  You have a measurement market.  You know, we’re doing something in the media. Alright, what does that mean?  What countries are there?  We’re doing something in the northeast, but it has different definitions of that.  So, building that basic glossary and having that common understanding of these very core definitional things is paramount to being able to leverage data.

Alexandra: Yeah, I can see that.  Especially if you mentioned sales reps having to put in the information.  How to actually get people also on all levels of the organization to better understand why data management and collecting all of this information is so crucial and paramount.

Scott: I call it the “golden rule of data”.  Do unto your data as you would have it to do unto you.  What you put in the data management is what you’re going to get out of business intelligence.  And it’s an inextricable, you know, undeniable physical law of the data space.  Most people just call it, you know, “garbage in, garbage out”, “GIGO”, “rubbish in, rubbish out”.  But we’ve got to describe these things in ways that people take action on it.

Alexandra: Understood, understood.  One other aspect that’s super interesting for me and our listeners is privacy.  How would you advise businesses to integrate privacy in their data strategy?  Is there a bulletproof approach for that?

Scott: I’m not too much of an expert on sort of that particular space in terms of privacy and data regulations, but I sit back and watch a lot of these trends and themes going on and go, “Okay, there’s a data management piece to that.”  Obviously, privacy is protecting data, is governing it, is making sure that access to it is monitored and controlled.  All that happens in - a lot of the core pieces of that happen in data management.  And a really easy way to destroy your brand today is to, you know, have the data breached, have your data released in a way that it shouldn’t.  And it creates yet another area of value that the data management community can bring in helping mitigate that risk, in helping control that data access.  Obviously, there’s the technological controls in terms of being able to get in people’s systems, but there’s also - you know, a lot of that’s driven by authenticating the identity of the user, having a common data structure that people understand, and certain - you know, I get too technical because I can’t, but a lot of those principles are within the data management realm.

Alexandra: Yeah.  Understood, understood.  One other thing that I wanted to discuss, you mentioned it in one of our previous conversations.  And what you mentioned was that sometimes data scientists even brag with or complain about how much time they have to spend with cleaning up data, preparing data, and so on and so forth.  What’s your perspective on that?  Is this really necessary?  Or should data scientists bend the work rather on the tasks they were actually hired to work on?

Scott: I would hope they would spend time - I would rather they spend time on what they were really hired for.  These data scientists that almost brag, as you suggest, that they’re spending 60 to 120% of their time munging and wrangling data - first thing is, if you spend 60% of your time, 80% of your time doing something that isn’t what your job is called, then what are you?  They should question themselves on that.  But a lot of those things that they work on are being managed already in the data management side.  So, I’d reach out to that data science community and say - and the data management community and say, “Okay, you guys should get together as much as possible.”  If you’ve got a really strong data governance team, and they have already structured the data in the way that the business understands it, they’ve already created frameworks and integration techniques that allow you to leverage this data across a bunch of different systems and different use cases, that’s your golden source of this content to do the data science work.  There’s some munging.  There’s some wrangling that has to happen.  But it just strikes me, if you’re spending more of your time managing and cleaning and structuring that data than actually working with it, then there’s a problem that’s got to get solved somewhere.  And the prob - and you don’t solve the problem by teaching data management to data scientists because they’re doing things, in a lot of cases, kind of experimentally, ad hoc, maybe offline before it goes into production.  If they can reach out and say, “These are the things I need.  You know, we’ve got a product hierarchy that market already understands.  We’ve got a customer segmentation structure that’s already being managed.  We’ve got un-duplicated data in our customer master, vendor master if you go and find it.”  Let’s stamp out munging if we can.  Or wrangling.  I don’t even know the difference between those two.  But it just strikes me as way too much time spent on something that isn’t part of their job, and I really feel that there is a solution out there or at least part of the solution by reaching out to data management community and getting that better content. 

Alexandra: Yeah, yeah.  This absolutely could be the case.  And I think for all these organizations that are struggling with finding new data science talent, maybe it’s also advisable to first look into the data management practices and whether they could optimize there to free up some resources on the data science side.  So that’s definitely one point to consider.

Scott: I would hope so.

Alexandra: One other thing that I wanted to discuss with you.  For companies that in general are looking to build a stellar data management team, what’s the secret to good hiring practices and what type of people do they need?

Scott: I’m not an expert in that, so - again, I’m not a practitioner on that side.  But I found that if you look for folks who do have those soft skills, who do have those storytelling techniques, who can collaborate with the business stakeholders, who have that emotional intelligence to be able to describe in very simple, clear fashion, the value they’re trying to bring, just like any role, that’s going to build camaraderie and teamwork and a better internal relationship.

Alexandra: I can imagine.  Coming back to the book that you wrote, “Telling Your Data Story”, what else super valuable takeaways that readers can find in this book?  Is there anything else that you could share with us today?

Scott: There’s all kinds of tips in there, especially around selling.  So that’s sort of a unique aspect, trying to highlight in the data storytelling space because I’ve determined the kind of story you need to tell when you’re telling that data story about data management, it’s a pitch.  You got to pitch it.  You got to sell it.  And some people kind of bristle, especially if you come from a more technical practitioner world.  I come from just soft skills, obviously.  All I do is talk about how people talk about it.  Selling may seem like it’s something that’s kind of foreign to them, but selling at its essence is convincing somebody of the benefits of what you have to offer.  You’ve got to put a narrative together that captures their attention, that develops their interest, that builds their desire, that leads to some form of action.  You want people to take action on what you do.  And that process is a selling process.  Even though you’re not, you know, looking for ink on a contract and delivering a product, you’re still working on trying to convince somebody of something, to take action.  So, there’s a whole section in there, kind of my fast tips on - it’s not a sales book by any means, but just sort of opening that up.  Some presentation tips.  Some ideas on how to talk to senior-level executives.  I’ve spent more than my share of time talking to CEOs.  And I remind people again, very seldom will a CEO ask you how you’re going to do it.  They might interrupt you and say, “Why are you even telling me this?”  Focusing on that “why” is important.  And more forefront in the book, I talk about - I have a whole list of things that I don’t know about.  So, like if you bought this book to understand how to put a data management team together, or how to be a CDO, or how to create a data governance program, I’m not your guy and I list some other folks who I think are really good at that.  If you’re looking for ways to convey this value in a story-type format, in more of a narrative, then I’m - you know, I feel like I’m one of the best out there in terms of being able to get people excited about this topic that tends to not get enough attention.  And when it does, it tends to just bore people.  So, I’m going to get some more excitement out there.

Alexandra: Yes, that’s definitely good advice, how to help people with a more technical background to speak to the business side and get their buy-in.  One other point I wanted to touch upon, we also spoke about this in a previous conversation.  Today, many organizations really want to engage with artificial intelligence, data science, machine learning, and so on and so forth.  But what we see in the industry is that many struggle to actually scale their AI practices and oftentimes end up only being in the POC stage.  Is there also some connection to general data management practices in your point of view?

Scott: I think so.  If you end up only in the POC stage, part of the reason you don’t put it into production is because you don’t have the data to back it up.  The data isn’t there.  It is structured in a way and governed in a way that can really put those wonderful AI ideas, let’s say, into practice.  So, making sure that you’ve got those sources even if you’re only doing a POC with a little slice of data.  That’s why I keep recommending that the data science community engage with their management community.  You can build all these wonderful things in a Petri dish, but if you don’t have the rest of that content - you know, segmentation, I’ve seen before.  “Segmentation analysis, we’re going to be segmentation.  We’re going to drive industry sectors.  That’s how we’re going to focus.”  And then you look at their data seg - you know, the way they’ve segmented their customers and it’s really blunt, or as I suggested, a third of it says, “other”, or there’s duplicate - always go back to the same thing, for me.  

Alexandra: Yeah.

Scott: Is that content, you know, fit for that purpose?

Alexandra: Makes sense.  And since you have the luxury of being closer to so many enterprise organizations, do you also see some correlations in regards to the emphasis and the importance that senior managements put on data management and the overall performance of these companies?  So are the more competitive and successful organizations better at the data management practices?

Scott: I think any major technology disrupter - you know, let’s name some examples.  You know, Amazon, Uber, Airbnb, all these wonderful disruptive providers, they all work off of - you know, they’re all built off of highly structured, well-governed, expertly managed data.  It just doesn’t happen without it. 

Scott: Yeah.

Jeff:  They’ve got, you know, I’m a geek in this part of the stuff.  So, you know, Amazon has a unique code for every single product.  They sell so many things that an industry-standard code in a certain sector doesn’t apply because they sell everything.  There’s no common data - you know, there’s no common identifier structure out there for that.  There’s UPCs for consumer products, as an example.  I look at that as little codes called an ISAN , and it’s on every single product.  You go, “Okay, their stuff doesn’t work -” I don’t know anybody in Amazon internally, I just know from an outside, you go, “Okay, there it is.  There was that - that was the little nugget that I was looking for to say, okay, they’ve got all this.”  You know, it’s obviously well-structured.  They know every product.  They’ve got that attached to a category structure and a geography and an owner.  So, when I look at a particular product, they’ve got the data to suggest - that’s where it comes from, right?  How do I know - you know, I’m buying this kind of shoe, they’ve got a categorization on all their shoes that can serve me up the next kind of opportunity.  Same thing with Netflix, that micro genre of all the entertainment content.  So, if you watch, you know, romantic comedies about dogs, then they know that they’ve got that - they can serve me more of those kinds of things.  If they’ve got it, that’s all working off of well-structured data.      

Alexandra: Yeah.  Absolutely.  So also, being able to give better recommendations.  What else would you say are the most relevant business benefits of proper data management that you can pack into this pitches toward senior management?  What gets senior management excited when talking about data management?

Scott: You know, three ways that data management brings value to a company - and I think they’re really the only three ways that a company looks for value.  So, helps to grow the business, improve the business, and protect the business.  And every company’s trying to grow, increase sales, market share.  They’re trying to improve operational efficiency.  They’re trying to protect themselves, mitigate risk, validate relationships.  The beauty of data management and well-structured data that comes out of it is it can do all three of those things for an organization, sometimes with the same record.  And that’s the core of what businesses do.  You know, I think every business is trying to provide value to their relationships through their brands at scale.  And you cannot do things at scale unless you’ve got technology.  Technology is hardware, software data.  If you don’t have - if you have data, you need data management.  Again, you know, some people might think this is an oversimplified view.  Anybody can make it more complicated.  But explaining in that way, I think, highlights the importance of this part of it. 

Alexandra: Yeah, yeah.  And especially since you mentioned how important it is to get buy-in.  I think it’s really good to also bringing it on this high level and really make it understandable for people that are not living, breathing, and thinking about data.  So, thanks a lot for sharing all of these insights with us, Scott.  For our listeners, if they - I think they by now already agreed that data management is something that’s very important to have, so if they’re looking for tips on how to actually convince others within the company, then definitely check out Scott’s book.  Before we end, we usually like to play “This or That” game, so just answer with what comes first to your mind.  Are you ready for the game?

Scott:  Okay.  I’m ready.

Alexandra: Wonderful.  So, first question, data is the new oil or the new ammonium nitrate that can suddenly blow up on you?

Scott:  Data’s not the new anything.  How about that?  Neither.  Neither this nor that on that.  So “data’s the new oil” is a phrase I hate because it’s a - it doesn’t convey a consistent message to people.  So, people say, “Oh, data’s the new oil.”  Well, does that mean it’s like oil so it’s got to be refined then it’s fuel?  Or it’s like oil because it’s - you know, “It’s not like oil because it’s not really sustainable.  And, you know, data doesn’t used up like oil.”  And I hear all this debate going back and forth and go, “It’s a metaphor.”  The purpose of a metaphor is to shortcut the time to understand what you’re talking about.  So, I don’t care what side of that debate you land on, mine is don’t use “DITNO”, as I call it.  Say no to DITNO.  Data’s not, you know - and it’s not new either, alright?  Data’s been around since before computers.  It’s been around since before electricity.  If you - some people claim that caveman drawings were the first data, so in Flintstones’ timing, data’s actually older than oil.  So, say no to DITNO.  You picked a good one for me to start ranting on.

Alexandra: Sounds good, sounds good.  I mean of course the cave drawings potentially weren’t as structured as we would want them to have, but could arguably be the first data point or one of the first.

Scott:  Exactly. 

Alexandra: I think that’s also something that’s challenging for today’s business leaders, is that data more and more has to be managed as an asset, but it’s such a different asset than, for example, money because you can share it and still keep it.  And I think they’re from so many aspects, it’s so much different to what businesses are used to working with.  So therefore, I think there’s still some adaption that needs to go on.  But to our next question.  Big data or small data? 

Scott: Small data all the way.  Big data needs little data.  Little data is small data, structured data.  Structured data works harder than unstructured data.  Period.  End of story.  People are trying - part of what makes big data big - you know, there’s volume, right?  We got to collect it all.  There’s ways to do that.  There’s velocity.  Sure, the speed makes, you know, a difference.  But it’s the variety that kills you in terms of the three Vs.  So, to get value out of big data, you must structure it.  That’s - you know, little data, small data comes in there, so I’m a small data guy all the way.

Alexandra: Makes sense.  AI regulation, yes, or no?

Scott: Sure.  Why not?  Why wouldn’t you?  Control the stuff that gets out of control.  So, I’m not an AI expert, but the last thing you’d want is a machine that does things that you can’t stop.  I had a non-AI but what looked like an AI-related problem.  My ice maker in my fridge would not stop making ice.  It was like - this is a small example of what - and was like pou - I’m looking at my fridge over here.  I hear this pounding.  I hear all this stuff falling in the freezer.  I had to take the whole thing out.  I go, “This is, you know - this is like an AI-based, you know related problem."

Alexandra: I know. I now imagine your kitchen being filled up with ice under the ceiling, which is something that I actually envy considering we have 30+ degrees we currently have here in Austria, so would love to have a cooling kitchen like you do.  But of course, probably quite annoying.  Last question from our “This or That” game.  Working in the office or remotely, what do you prefer?

Scott: I’ve been working remotely for decades.  I love it.  So, I don’t work in an office any more since I’m a, you know, boutique consultancy.  And after the last year and a half, I’ve really, I think, transitioned - I mean I had to transition all my business to digital, you know, to virtual.  I do a lot of events.  Like I think, you know, we talked before about five things I’m doing today, and one way or the other, I could never do that if I had to physically go places.  So, it’ll be fun to get in front of people again, but, for me…

Alexandra: It also has a value.

Scott: There’s a wonderful by-product to not having to travel everywhere all the time. 

Alexandra: Wonderful, wonderful.  So, I’m happy to hear that you’re happy with the current situation of possibilities of the home office.  Wonderful, Scott.  Thank you so much for your time and all your valuable insights that you shared with me and our listeners, and looking forward to talking to you soon again.

Scott: Great, Alexandra.  Thank you so much for having me today. 

Jeff: That was so much fun, Alexandra.  Scott really knows how to tell good stories in a fun and entertaining way. 

Alexandra: He sure does.  It was refreshing to hear someone who is so excited about data management.  And we received some great advice on making enterprise data projects happen.

Jeff: Those will come in handy for sure.  So, let’s collect the takeaways.  Why don’t you begin?

Alexandra: Sure, let’s do that.  So, first takeaway, focus on where the data journey begins.  It’s really about determining the truth with your master data.  You also need to care about metadata, data governance, and data stewardship, but Scott highlighted that master data is really the most important data source in your organization and that all the other data sources are basically about your master data.  So, it’s so important to get this one right and have truthful master data. 

Jeff: Yup.  And he also talked about the three Vs of data management, which are vocabulary, voice, and vision.  For vocabulary, you need to consider and align your words and terminology.  With the voice, this is the way you talk about your data, not only as an individual, but as a group, team, and organization.  Harmonize your voices to make it sound good to the business side.  And I love that.  Vision covers strategy.  Data management should be enabling the strategic intentions of the enterprise, determine what’s the direction and what kind of data is needed to get where you’re going. 

Alexandra: Super valuable insights that we got here.  Then of course stories are important.  Data storytelling helps to connect with the business side.  You should really avoid to go into the technical details and really concentrate less on the “how” and more on the “why”, especially when talking to senior-level decision makers.  What also helps is using simple analogies to describe data, and that’s really effective because images stick much better with people.  And if you really use imagery that’s already familiar in your company, it’s a great way to find analogies that work for your business partners.  We have the example that you really could explain data management if you want to implement it in a bakery, with analogies of bread, sandwich, and also the ingredients that you need for that.  So, find something that people can relate to and never become too technical.

Jeff: I love this.  And I think storytelling is something that I can definitely improve.  And I really appreciate the visualization of these stories because they really stick with you.  Most enterprises have challenges also at the foundational data level.  And as Scott pointed out, the lack of good data management stops them from really realizing their aspirations.  You have data silos, disparate data, lack of internal standards, lack of common definitions are also common.  And these data problems tend to be the same across industries and across verticals.

Alexandra: Yeah, that’s right.  So, Scott really emphasized that it’s so important to get the basics right and that still this is something that’s lacking within many organizations, no matter which industry you’re looking at.  And then we also talked about senior-level support and that getting this for your data project is essential because they need to happen through cultural transformation.  And to get this support, it’s really important to be able to explain things to the board.  And we already heard this, making it simple is key to let them understand why data projects are so important.

Jeff: Totally.  And data management success is seen and felt also on a customer experience level.  Financial service providers, for example, need customer interaction data to become relationship-centric.  Customer frustrations are stemming from data issues all the time.  We’ve all experienced this.  And customers don’t see the data, but they feel the result of good or bad data management.

Alexandra: That’s so true.  And here we have it again, you need to get the basic stuff right.  A lot of the data problems go away in reporting, integration, interoperability if you get the basics right.  Here Scott shared the four Cs with us that you should focus your attention on.  And those are code, company, category, and country.  Code stands for un-duplicated, unique identifiers.  Sophisticated customer church models won’t work unless you have clean, un-duplicated data.  Company stands for hierarchy structure.  Your data needs to be consistent across departments.  Category stands for segmentation.  If you don’t have the categories right or have useless categories in your data like “other”, you won’t be able to segment your customers and create efficient targeting.  And then we also had country, which stands for geography.  And getting geography right in your data is also paramount for success of any of your data projects. 

Jeff:  I really like the golden rule of data, too.  As Scott explained, what you put into data management is what you will get out of it in business intelligence. 

Alexandra: That’s right.  And then we also talked about privacy.  And according to Scott, the core of privacy protection also happens in data management.  Data privacy is another benefit that you get if you do proper data management.

Jeff:  And for best results, data management and data science teams should work closely together.  In too many organizations, data scientists spend a large portion of their time on data preparation and data management tasks, which, obviously, isn’t a good value or a good use of their time, I should say.  And the solution is not to teach data scientists data management, but to give data scientists well-managed data. 

Alexandra: Absolutely.  So, this would really free some organizations about their concerns of not finding enough data scientists off the market if they allow those data scientists they’d already have in their organization to work more efficiently and more effectively for what they were hired to do.  And then Scott recommends also hiring data people that have soft skills, like excellent storytelling and emotional intelligence because those are really important to be able to communicate with the business side.  Data projects need to be pitched and they also need to be sold to the businesses or the business side of your organization.  And it’s worth looking at them as products.  So, if you’re looking for ways to convey the value of data projects, to get people excited about the topic, then check out Scott’s book that we mentioned.

Jeff: Of course.  Data management brings value to companies in three ways.  As he outlined, you can grow, improve, and protect the business.  And technology, of course, is a tool to help you do these things at scale.

Alexandra: Absolutely.  And remember data is not a new oil.  It has been around since the dawn of times.  Also keep in mind that structured data beats unstructured data every time.  So, for Scott, small, structured data is always more valuable than unstructured, big data. 

Jeff: Indeed.  Well, thank you, Alexandra, and thank you, Scott, for this educational and entertaining episode.  Scott and Alexandra, we thank you.  To our listeners, we thank you, too.  Hope you can take a minute or two to leave us a review.  We appreciate you coming by and look forward to seeing you on the next episode of The Data Democratization Podcast.

Alexandra: Bye.  See you soon.

Read More

What's your use case? Find out how synthetic data can empower your organization!

Talk to a synthetic data expert!
magnifiercross