This is an audio conversation from the CIO Talk Network podcast. The timestamps in the chapters and transcript below refer to the audio recording.
About this conversation
A CIO Talk Network conversation with Celso Mello (Reliance Home Comfort), hosted by Sanjog Aul.
What the conversation established
The conversation established that data is the foundation of any digital effort, and that most companies stall not for technical reasons but because they lack a defined use case and are structured in silos. Celso Mello of Reliance Home Comfort said the company processes about 500,000 service calls a year and plans roughly 800 technicians on the road using weather, traffic, contractor, and availability data, an example of starting from a defined use case rather than chasing a single version of the truth. He argued a data strategy should begin as a small satellite tied to one use case and grow into a mothership only if it succeeds, that process and data must be designed together as the Uber experience shows, and that organizations should reorganize around the customer and empower the frontline to own data. He also noted that integration between clouds and back to legacy systems remains the hard part.
Host Sanjog Aul talks with Celso Mello, Chief Information Officer of Reliance Home Comfort, about building a digital data strategy, architecture, and infrastructure. They discuss why companies stall on data, how Reliance Home Comfort improves data quality and control, why a holistic data strategy matters only to the extent it supports a defined use case, and why reorganizing around the customer and empowering the frontline matters more than the technology itself.
Contributors
- Celso Mello, Chief Information Officer, Reliance Home Comfort
Host: Sanjog Aul, Founder and Host, CIO Talk Network
On this page: Key takeaways · Chapters · Explore more · Transcript
Key takeaways
- Celso Mello, Chief Information Officer of Reliance Home Comfort, agreed that most companies rate themselves no higher than a 2 or 3 out of 10 on data, and said the reason is that the use case for a holistic approach to data is not well defined, not that the technology is missing. He tied the stall to legacy issues, disjointed applications, duplicated databases, and a fear that new digital channels could cannibalize existing revenue.
- Celso Mello of Reliance Home Comfort said data problems come down to quality and control. He described mistake-proofing data at the front end, having field technicians scan barcodes or photograph serial numbers instead of typing them, and consolidating separate billing, sales, and marketing databases into central repositories to give the frontline a 360-degree view of the customer.
- Celso Mello of Reliance Home Comfort argued that a single version of the truth is worth aspiring to but is not a prerequisite for digital. He said a data strategy only needs to go as far as the defined use case requires, citing how Apple’s pirate team built just enough supporting infrastructure for a new product rather than a holistic data strategy.
- Celso Mello of Reliance Home Comfort recommended building data architecture as one small satellite tied to a single use case, then letting it grow into a mothership only if it proves successful. He illustrated the approach with a third-party home-services marketplace that required only a defined slice of customer and technician data.
- Celso Mello of Reliance Home Comfort said process and data must be designed together, using Uber as an example where customer location, driver location, and destination are inseparable from the process. He warned that simplifying a process for the customer can complicate the data journey when the two are handled separately.
- Celso Mello of Reliance Home Comfort said the most important change is organizational, not technical. He urged companies to reorganize around the customer experience rather than departmental silos and to empower the frontline to own and correct data, noting it is often easier to stand up a separate digital business unit than to change the main organization.
Chapters
- 0:00 Introduction: building a digital data strategy, architecture, and infrastructure
- 2:27 Why companies get stuck on data
- 6:06 Improving data quality and control
- 8:20 Beyond customer data and the MDM lesson
- 12:06 Aspiring to one version of the truth
- 17:17 Is a holistic data strategy worth it
- 20:15 Data architecture as satellite versus mothership
- 30:50 Building the data infrastructure
- 41:34 Organizing around the customer
- 45:28 Connecting process change and the data journey
Explore more
- Building a winning digital culture
- Data Security for the Digital Business
- Building Unified Digital Workspaces
- Building Endurance for Digital Transformation
- Creating a Winning Data and Analytics Strategy
- Building Your Digital Business Technology Platform
Transcript
Introduction: building a digital data strategy, architecture, and infrastructure 0:00
Sanjog Aul [00:00:28]:
The topic for today is Building Digital Data Strategy, Architecture, and Infrastructure. Our guest for today is Celso Mello, who’s the Chief Information Officer for Reliance Home Comfort.
Sanjog Aul [00:00:43]:
The reason we wanted to talk about this topic is either in the last, I’d say, a month, I’ve spoken almost on all different angles of digital transformation, maybe on the live broadcast like I’m doing here with you today or at different in-person events around multiple cities across United States. No matter who we speak with, and yes, they talk about other aspects of digital, but somehow this whole data as a critical element is popping up again and again and people are expressing serious concerns that, okay, business wants us to go digital, but without data there is no digital. So what we wanted to explore here today is to see, yes, we would have a good data strategy and, we could at least work towards it and then you want to have an architecture and the infrastructure, but since there are some challenges in doing that, that’s why even if we say, we will build a data driven culture, that may not be enough. So how do we get to the brass tacks without going very technical, but still be able to build a rather holistic approach to building a digital data strategy, architecture, and infrastructure. That’s essentially a goal. So, Celso, thanks so much for joining us, and let’s talk about the first aspect, the very current state. Most companies and their CXOs, including CIOs and CDOs, both say that from a scale of 1 to 10, I don’t see us being above 2 or 3 when it comes to data. Where are we stuck?
Why companies get stuck on data 2:27
Celso Mello [00:02:27]:
I’m afraid I’m part of that group. It’s really hard at this day and age to have a good handle on data, and I suppose this happens as a result of many legacy issues that creep up over time. Successful companies, like ours, usually grow in the process. They have disjointed applications. They have duplicated databases. The suppliers actually make it a little more difficult because with the hype of big data and analytics and all of that, it creates more confusion, unfortunately. So in some, I think the use case for a good holistic approach to data is not well defined, and that’s what makes us stuck. Yes, you’re right. The digital strategy requires a good data strategy. Without data, there’s no digital, but the precise use case in terms of how we’re going to architect and put our data together in a way that is meaningful to facilitate a digital business is not very well defined, and that is not as easy to do as it sounds. It’s something that, because of the amount and the overwhelming influx of data makes it even more complicated, but I think at a deeper level, if I may, I think there’s a bit of a fear of disruption to the existing sources of income. The digital approach to companies is typically a new channel of business that might cannibalize the existing channels, which makes it more complicated for senior management to embrace it fully and take it forward, meaning that they might be putting their own sources of revenue at risk.
Sanjog Aul [00:04:28]:
And you are spot on in terms of your assessment of many companies who have not been able to make progress in this field of data because somewhere at the top, there is the risk aversion, if you will. So that they could keep doing what they’re doing, but at the same time, we know that you got to keep moving and moving ahead for you to stay relevant, and so, that’s a different conversation where how do we get the people at the top to start recognizing. So now let’s look at companies where they understand that they either move ahead or perish, and the business has understood, and they are willing to support it. They are willing to sponsor it. So now coming to where data is, and then you mentioned that you could have data coming as new form of data versus what you have been tackling. So if I were to ask you specifically do we have a good handle on the data we have been using? Because if you have a successful business, that means you have some good processes running, you’ve got some good visibility about the data so that you’re keeping your customers happy even though at a transactional level. That means you got that to a point where anything new that comes, it is not being stopped just because you don’t even have handle on your existing data. So, if 2 years ago, if I’d asked you this question versus now, would you say that you are better off at least dealing with what you have in terms of data?
Improving data quality and control 6:06
Celso Mello [00:06:06]:
I think we’re getting better every day, and I think there’s still a long way to go. So the data we have, and I’ll talk about customer data in particular because that’s more meaningful in the digital context. There’s two problems with that. There’s quality and control. By quality, I mean inputs are perhaps not as precise and accurate as they should, and control is the data is not properly organized in a way that it’s easy to access, and every day, we’re making strides towards improving both ends. So in terms of data quality, for example, you can mistake-proof at the front end and ensure that your frontline people will enter data in a way that is checked and it’s precise and is accurate. We’re doing this every day with the mobile technology influx that exists now. We can create lots of logic that allows our frontline people, even field people that drive trucks in our business, for example, to enter data or scan barcodes as an example, which reduced the risk of input errors and things of that nature. So 2 years ago, that wasn’t the case. We would rely on people to type in information like serial numbers and very detailed data like this, which, just by definition would be prone to error. Whereas, with now a phone, you can take a picture of a barcode and the data will be right every time. So that’s something that we’re better at now, and that’s on the quality side. On the control side, our intention or our strategy has been consolidation of databases. So we have multiple customer databases, for example, with the separate data like billing and sales data and marketing data and so on, and we’re bringing it all together in central repositories and trying to use it not only for reporting purposes, but also for transactional purposes, giving that 360-degree view of the customer to our frontline people. So we’re making strides on both ends, but there’s still a long way to go.
Beyond customer data and the MDM lesson 8:20
Sanjog Aul [00:08:20]:
Now you did mention that you were tackling the current state or whatever the data that you wanted to handle and you mentioned customer. So now digital if you were to think digital is only customer, that may reduce the scope of what we should be tackling. Because to give a good experience to a customer, just knowing what the customer is saying is not providing us enough insights so as to surprise or wow them. Because if you just deliver what they expect, then it’s a transaction, but when we wow them, that’s when they take notice. It becomes remarkable. They share with others, and that’s how they stay or attract new people. So that means we have to look outside the customer only and maybe internally to see what we could be doing or what data are we collecting even at the operations level, which will allow us to optimize things and give a better experience. Now that said, we could go back to say 10 years ago. We started this journey of MDM, which is master data management. Lot of uproar regarding that, lot of excitement and investments made, and at some point, we said, you know what? This is not happening the way we expected because we are looking for that 1 version of the truth with a premise that a frontline worker is ready to own their data and be custodian of their data, but when you dig deeper and start talking to them, they say, I really don’t care. I’m here to do a job. So when nobody took ownership, it fell flat on its face. That time, we could still get away without having to solve that problem. Now with digital, we cannot afford to have it stay the way it was in the past. So what do you think we can do this time to get out of it because now our existence is at stake. It’s not just a feel-good project anymore.
Celso Mello [00:10:29]:
And I think you’re absolutely right about empowering the frontline. That’s something that perhaps didn’t work well in the past, but in the digital world, that’s so important. Consider the importance of user rating in eBay or Uber or Amazon. That’s the frontline. That’s the customer or the supplier that are providing that information. It’s not curated by anybody. It’s essentially raw information provided and it could be a decision maker in terms of make or break the deal. So in a real business scenario, I’ll give you an example of Zappos, I believe. Their frontline people that sell shoes on the phone have full power to modify or provide refunds or make decisions in front of the customer without any further approvals necessary from a manager or anybody in the back office. So that’s the sort of thing that I think is required in the future to facilitate the digital businesses, and most importantly, the data needs to support that type of process decision making.
Sanjog Aul [00:11:44]:
Would you say we should still aspire for that 1 version of the truth when it comes to data or is that utopic and we have to settle or at least start with achieving something lesser than that as a first step towards enabling digital?
Aspiring to one version of the truth 12:06
Celso Mello [00:12:06]:
I think we should still aspire to that, but knowing full well that is a target very difficult to get to. As long as we chart a path that takes us there step by step, I think we’re moving in the right direction, and there’s ways to do this. There’s data you mentioned earlier that we have, the data that we already own and the data that we don’t have such as, in our case, for example, weather information is important in the HVAC business and even traffic information for purposes of our route optimization. So all of that needs to be consolidated in one place or in an easily accessible place in order for us to properly service the customer or do it in a way that not only satisfies their needs but also wows them. So we’re not going to get to the point where all the structured data that we already own and the unstructured data that we need in order to facilitate our processes are all easy to get to and fully integrated with our back-end systems, but if we know that that’s the target, we can take steps towards slowly bringing them in and integrating them into our processes such that little by little we move towards that point. So an example here is, again, back to our business. We use weather forecast information to plan capacity. We process about 500,000 service calls a year. There’s about 800 service technicians on the roads at any given time. It’s important that we plan this in a way that we don’t create overcapacity, which is costly for us, or we don’t have undercapacity, which is a problem for our customers. So weather is a critical element in that planning in addition to technician availability, contractor availability, route optimization. All of that is brought into a model that we created to help with that planning. We don’t get it right every time, but every year we look at what we did wrong the year before and try to sharpen our model so that next year we get it right. So yes, we’re not going to get the 360-degree view or the holistic nirvana of data, but if we take steps towards that every day by introducing a little bit every time, I think we’re going in the right direction.
Sanjog Aul [00:14:39]:
I’d like to ask this fundamental question that should we even aspire to have a holistic data strategy? Because if you try to go after the impossible, and I’m not saying it is impossible, then we are losing out from what we could otherwise achieve up front. Yes. It’s great to have a vision, to have a utopia, but is that truly a practical holy grail, if you will?
Is a holistic data strategy worth it 17:17
Sanjog Aul [00:17:17]:
So sometime back, I remember I did a show of, lean back to treat your customer better, lean back to do more, which means how about taking a step back and asking this fundamental question. Should there be a holistic data strategy, and could there be a holistic data strategy? One is that’s even worth it. Another is, is it even possible?
Celso Mello [00:17:45]:
That’s an interesting question. It’s a philosophical discussion. If you aim for the impossible, is that disruptive? Is that constructive or disruptive? Is it going to take you somewhere or is it only going to distract you from your goal? Different people will probably take this in a different way. Everything works as long as we’re making progress towards the right direction, which in this case is to facilitate a data strategy that gets to your business digital goals. There’s a step before that. Consider all of the processes, the applications, the people, and the technology you need to enable your digital business, whatever that digital business is. In order to achieve that goal, you’re going to need a certain data infrastructure or data architecture. So whether that’s holistic and includes structured and unstructured social media, everything, I think it’s less relevant than getting to the point of facilitating the digital business you have. If you define that use case well enough, the holistic data strategy becomes less relevant. It’s just important to the point of facilitating that use case. So I’ll give you an example. Apple created a pirate group that went off the mothership to try and create a new product. They had to create their own supporting infrastructure to get to that point, and I presume data was an important part of it. Did they create a holistic data strategy? Perhaps not. They just created a use case and a data strategy to facilitate that use case. So should or could there be a holistic data strategy? Maybe yes, maybe not. My answer here is only to the point it supports a very well defined use case.
Data architecture as satellite versus mothership 20:15
Sanjog Aul [00:20:15]:
And when it comes to the data architecture portion. So we have to kind of build a mothership when it comes to architecture, which is not only to tackle what the type of data, the variety, the volume, the velocity, whatever that architecture we have to create should be able to support. Not only what you have, but also create placeholders of what’s possible and what we see coming down the pike. There are specific architecture groups within companies, and yes, there are data architects, but when it comes to digital, what would you think is the state of our maturity in being able to produce an architecture which could turn out to be a mothership? Or do we just create satellite ships first and try to tackle what we have in bite size and then think mothership? If you do that, you might get your immediate needs covered, but you are truly not moving towards building a castle. It’s almost like laying a brick, and you need to know which castle are you building, and since people have struggled with this architecture and maybe there is a struggle related to architecture, that’s why they’re not able to build that mothership. What do you think could be done new, more, or different specifically in the data architecture context?
Celso Mello [00:21:48]:
I think, to continue with the analogy, the best way to approach this is to build a satellite, not many, but 1 little satellite that will eventually become the mothership if it’s truly successful. So it would start small. You create a data architecture or infrastructure that supports the specific use case I said earlier. It will start small, start with a pilot and try it a little in a small market, start to be successful, then you start to incorporate more data. The data architecture increases and eventually it might outgrow where you started from and it becomes a mothership by itself.
Sanjog Aul [00:22:32]:
And if suppose you even if you did that, what does it do to the big picture thinking? You’re not just getting one thing done at a time because that’s what we’ve been doing all along. Digital is forcing us to think or to take a step back, look broader, look bigger. Is it a matter of practice or is it a matter of us never being exposed to it, that’s why we are not able to take a step back? Or are you saying that’s not even a good idea to begin with to start taking a step back but then lose out on what you’re doing on the ground? The concern is if you just look at laying of the bricks and not thinking about which castle you’re building, we might build a bridge instead of a castle, which is not what you wanted.
Celso Mello [00:23:17]:
True. Let’s take a moment here and discuss what digital is. I think there are companies where the product is entirely digital and where it wasn’t necessarily used to be, but it became digital. An example is Netflix, truly a digital product, compared to Blockbuster, which rented the physical product. In other cases, the product uses digital just to facilitate the performance of a physical service. Uber is an example of that – a physical service transporting passengers from point A to point B, but the way you go about getting to that service is all digital. A company like Amazon does both. You can buy a book through a digital platform, but you can also download that book, which is an entirely digital product. The distinction is important because when it comes to the data architecture that you need and how you go about it, whether that’s a satellite or the mothership, it’s very related to what the product is and whether the whole thing is going to be digital, the product itself is digital, or the delivery of the experience is digital. There’s a different set of data requirements that you would need in either case. For example, in our HVAC business we sell and install furnaces and air conditioners and water heaters and plumbing. By definition, that service is always going to be in the physical world. However, we can deliver that experience in a digital way. We can facilitate our customers reaching out to us for a service request in an entirely digital experience, be it through an app or online where they can not only request service, but also find availability of our technicians, even see where the technician is, who the technician is that is coming to their door and when and where their truck is. That entire experience can be made digital, but ultimately the service is going to be performed in a physical way. In order to facilitate that digital experience of getting a physical service, I need a certain amount of data about the customer, about the technician, about the equipment, and all that. That will drive my use case, which in turn will drive my data architecture. If, however, my entire product was digital, I would need a lot more information about the product itself, and to provide customers with that experience it would require a different kind of data architecture. I don’t know if I answered your question or went around it. I just felt it was important to talk about that digital definition.
Sanjog Aul [00:26:22]:
No. That is the definition which most people use. So there’s no doubt about it. We just wanted to make sure that we are not losing sight of the forest for the trees. You see what I mean? And the issue here is this, people have a tough time defining or seeing the forest to even think about. So they’ve always been seeing the trees. In the digital context, what you are trying to do on the data architecture side, the challenge is that they’re having a tough time seeing the forest. I’m not saying forgetting. You have to see it for you to envision it, but for them to do that, that’s where people are challenged because they’ve always, in the data context, always played at the tree level, and I’m trying to see if they keep doing it, then you would not get the full potential. You mentioned different ways people use digital. So you could either be using it for operational improvements, or you could be doing something fundamentally different in terms of a business model or a value chain reconfiguration, or you could be coming out with product innovation, but all of those things have to be supported by something which will allow you to connect all different pieces wherever you’re going with using data as a resource, as a key element of how you exist as a company, but that’s where if we don’t have our arms around it, we will be using more of a shotgun approach. Is there a way to prevent it?
Celso Mello [00:28:20]:
That goes back to your very first question in terms of where we’re stuck. We’re stuck in not seeing the forest from the trees. You’re absolutely right. I think the way to get out of it is to define that use case that I keep talking about. I’ll give you an example of how we can get out of it and take a small step towards seeing the forest from the trees. Some time ago, we decided to partner with a third-party company to provide our customers with the ability to create a marketplace so that customers could order HVAC services as well as any type of house service such as painting and plumbing and electrical. We wanted to create that marketplace through a partner where customers would feel comfortable ordering a service through a series of reviews that they would get from other customers that did the same type of experience, similar to what you would see on a TripAdvisor-type experience where you have various hotels or travel experiences all rated and you would go based on that and feel comfortable ordering because you saw value there. We wanted to be part of a marketplace like this. Going back to the data comment, what data would we need in order to facilitate that experience? Customers would need to know that they are our customers so they can order. We would need to know the information that they have and how we can get to them, their address. They would need to know information about our technicians, their location, and things like that. That became, I guess, seeing the forest from the trees. We had a very well defined use case that allowed us to just segment pieces of data from the large universe of information that we have to facilitate just that 1 digital product that we wanted to provide to our customers. Once we knew that’s what we needed to do, it was just a matter of executing. So the hard part, as you say, is just the scene defining it and then the execution follows.
Building the data infrastructure 30:50
Sanjog Aul [00:30:50]:
Let’s move on to the data infrastructure that we should be building so that we can collect, process, store, analyze, dispose, separate out, distribute anything and everything that we do in terms of operations on data. We are able to do it in an effective and seamless manner versus patching in disconnected data sources in a kludgy manner and where things fall through the cracks. Because, of course, we need strategy. We need the architecture and the infrastructure so that when we do hit the road, we are able to execute well and realize the very strategy we started with.
Sanjog Aul [00:33:39]:
So let’s talk about the data infrastructure. Since we have a data journey, and data will flow through different plumbing and pipes, and it’ll be utilized, stored, processed, distributed, you name it, as a resource. You got to have an infrastructure which will support it and not only what you’ve been doing all along but if there’s anything new that comes our way which in this digital age, you have to adapt to, you’re not ripping apart things. So that means you have to think through the data infrastructure for the future, and it could very well be fundamentally different than what you’ve been tackling till now in the pre-digital era. So what have you been doing?
Celso Mello [00:34:29]:
It’s interesting you use the word data journey, and I totally relate to that. Traditionally, we used to talk about data in terms of inputs and outputs. Data would come in and there would be a process that modified that data or assigned that data and then there would be an output from that. It’s very different now. You have a ton of input from many different sources, some structured, some unstructured, and not all of that data gets processed or at least doesn’t get processed right away. It goes through a journey of sometimes maturing or getting combined with other data or becoming relevant until such time where it gets used towards some product or a report or a customer decision. It’s truly a journey rather than a process as it was before. What we’ve done here is, we have data coming in to our company from many different sources. Our own users are obviously a very important source. We also have GPS data, for example, that comes from our trucks on the road in terms of where our technicians are. We have traffic information. We have weather, as I said earlier. We have billing information. We have banking information, financial information. All of that goes to different databases and gets loosely integrated and sometimes stays there for years before we actually have a use case where we need that information to do something more meaningful with it. I talked earlier about an example where we create models that factor in weather forecasts and traffic patterns and technician availability in order to produce shift information to satisfy customer requirements. We already had that data in our databases for years, but it didn’t become relevant or necessary until we matured enough to create a model out of it. When such time came, we had to add even further data in order to make that model meaningful. I would like to finish that statement talking about the cloud as a facilitator for infrastructure of data. Since the cloud came around, it became a lot easier to talk about infrastructure, whether it’s to support data or applications or anything. You can, with the cloud, do things very quickly with a small investment and ramp it up or down without needing too many people or too much investment, which is fantastic. However, there are concerns with that too. There’s certain data that doesn’t lend itself well to the cloud, not so much because of security. I think security is something that can be handled, but in terms of integration, some of that data needs to be integrated back to systems that may not be on the cloud or may be on separate clouds or private clouds. Integration between various clouds is not trivial. So we have to be careful when segregating data or deciding where the infrastructure is going to be, whether it’s on-prem or on the cloud based on the need for integration with other data elements when making that decision as to where to host your information.
Sanjog Aul [00:38:20]:
Now when you do mention the data, so whatever infrastructure that you created in the past, if you were to not necessarily rethink, but maybe extend or in some cases fundamentally rethink, I would not totally rule rethinking out. What changes would you make knowing that you could get more unstructured data? It’ll come at a faster pace. You might have data generated at the edge, and that type of data is better off not stored on the cloud because there’s too much lag, latency, or other costs. So the volume, the variety, the velocity that you have, and on top of it, you want to do analytics on it because data itself has value, but really you’re trying to derive value from analytics, the insights that you get. So if you were to take your organization, for example, and you start thinking digital, what would be your incremental shifts in the way you saw the data infrastructure? And also on the ground, what would you change? What is your logical thinking about data infrastructure, and what would you do physically?
Celso Mello [00:39:38]:
Well, again, I think it starts with defining where data needs to stay based on volume, where the data is generated, how it’s going to be used, velocity, and all of that. All of that needs to be characterized and well organized in order for you to make decisions. In terms of changes, I think you need a well-defined infrastructure strategy on your own data center side in terms of redundancy, for example, and making sure that data, whether or not you use it immediately or use it for future purposes, is categorized in levels of how proprietary it might be, how secure it needs to be, and how much redundancy you need for that particular portion of your data. Likewise, if velocity is an issue, you may not put it on the cloud. So you may need to change your policies and procedures with regards to what stays on the cloud and what doesn’t. We haven’t really done much in my company in that regard. I think we’ve established procedures around security of data, confidentiality of data, what can go on the cloud and what cannot, but to a larger extent, I think we’re in the infancy of the infrastructure side of it. We are leveraging the cloud as much as we can, but we’ve bumped into obstacles with regards to integration back to legacy applications when we deploy data and applications on the cloud. So that’s an additional concern for us.
Organizing around the customer 41:34
Sanjog Aul [00:41:34]:
So, in the next level of what you’re trying to do, whether we spoke about strategy or architecture and infrastructure, it is all about people driving it. The leaders are the ones who have to see that what I did to come this far may not take me to the next level without fundamentally looking at how my organization is designed and structured. It could be the people, the reporting relationships, the way you are scaling back and up and what capabilities you have, whether it is in-house capability or what you have from outside. So if someone is getting ready in the context of data, what organizational changes do you think are recommended, or what type of organizational review do we conduct to find out what changes we need so that the organization becomes conducive to building a holistic data strategy, architecture, and infrastructure?
Celso Mello [00:42:35]:
I honestly think that’s probably the most important aspect of this entire conversation, the organizational changes. While data seems to be a technical concern, it’s not. The problems we’re dealing with have to do with the fact that the organization is structured in a way that perhaps is not conducive to proper data health. Let me be more specific. Organizations, when they grow, maybe because of scalability pressures, have to be organized in silos. You have the billing department, you have the sales department, you have the operations department, which in turn can be broken down even further, but customers don’t see companies that way. They don’t see a customer as, right now I’m in the process of purchasing something or I’m paying my bill or I’m getting a service, and as such I should be a different customer and deal with different people. Customers expect, especially in the digital world, that they’re going to be treated in a seamless way. It doesn’t matter what touch point they’re having with the company. It’s almost like organizations need to be organized around the customer experience as opposed to departmental silos in order to facilitate the digital delivery of the service, which in turn has to do with the data. If you had the organization around the customer process, maybe the data becomes less of a concern because it’s naturally owned by the same people that are dealing with the frontline people who are dealing with the customer end-to-end. So I think this is about rethinking the organization around the customer and empowering the frontline. Given the frontline, the part of your company that is dealing with the customer firsthand, to make decisions in terms of updating data and getting data right the first time and correcting as they go, as well as processing and making decisions along the way. Now this is easier said than done. Typically, it might be easier for companies to segment a digital business outside of there. We talked about the mothership and the satellite before. It might be easier for a company to create a separate digital unit or business unit that is organized around the customer, that empowers the frontline and treats data in a seamless way than trying to do this with the main mothership because that’s so structured and stuck in their ways that it’s really hard to change.
Connecting process change and the data journey 45:28
Sanjog Aul [00:45:28]:
Let’s talk about the processes that we have been talking about changing as a prerequisite to digital transformation, and then when we think about process changes, of course, it’s going to have an impact on the data, whether in terms of how it gets created, how it gets morphed, how it gets stored, processes, you name it. When we are thinking of data as a fundamental key element, and process changes have to be reviewed and that becomes a prerequisite, should we keep the two separate? Because typically people say process changes should be kept independent of what data it’ll create, but should the data now drive also or should it have an impact on how much tweak we do to the processes to make the data journey better? Should we tweak the processes to make the data journey better? Or should it be abstracted from it till the time you look at process as the most optimized process you could create to get something done, but it makes the data journey more complicated. Should we connect the two or still keep them separate?
Sanjog Aul [00:48:39]:
So, for digital transformation, we know we need process changes and optimization before we go ahead and automate or do anything else with respect to digital transformation-related optimization. Traditional thinking always has been, okay, let’s simplify the process. Let it be optimized or simplified to a point where it gets the job done the fastest, least cost, etc, but we have never usually, at that time, inventoried the data entities or attributes; we will never think about what it will do to the data. That was an afterthought, but now with digital coming in, should we invite the data discussion even at the process rethinking level so that we are not left holding the bag by saying, okay, we simplified the process for the customer, but we further complicated our data journey and in turn made that whole data-related effort difficult and in some cases impossible?
Celso Mello [00:49:47]:
Look at, imagine what went on at the Uber office when they were thinking about what the digital experience would be for someone ordering a taxi. They probably thought about, okay, someone’s going to pull up their phone and the app is going to know where that person is. They’re going to need to know where available drivers are and everything has to be meshed together. So the process they were thinking about was part and parcel with the data. The data in this case is location of the customer, location of the drivers, destination of the customer. All of that is part and parcel with the definition of the entire experience, which I think in this case is process. So I think absolutely it has to be taken as one, both data and process, in order to facilitate the digital experience. After all, digital is manipulating the data. It’s getting data in and out and providing data sometimes as a product, or sometimes to facilitate the delivery of the product. That’s part and parcel of the delivery of that service or product.
Sanjog Aul [00:51:08]:
So, we see our Uber example? Celso, your Uber example is a good one, which is why digital natives have a leg up over regular organizations because they started knowing that data will be a foundational critical element for us to get to the next level. Talk about an airline. Talk about other complex events that happen, say, even in an airline. They did not start keeping digital in mind. They tried to make a good customer experience. Now try going behind the scenes and look at their processes, how many different elements play. If they were to start thinking of optimizing the process, that might optimize it for the customer but not necessarily internally, which actually prevents them from taking it digital all the way. This is just a question I’m throwing out for you and the listeners: are we at a point where we say, you know what? Whatever processes we change, it should not cannibalize by making the data journey more complex.
Celso Mello [00:52:12]:
I share the same problem here in my company as in your example an airline would do. We have a ton of data, and we have physical experiences as well as digital experiences which we are trying and struggling to turn into digital to keep up with customer demand. As you look behind the scenes, the data itself doesn’t lend naturally to the digital processes. So how do you do it? I think what we’re seeing is we have to take bits and pieces of that. To go back to your question, process and data need to be looked at together. There’s no way for you in a true digital experience to optimize the process make it faster, better, and cheaper without looking at how the data is going to facilitate that in the digital world. They’re very much the same. I’ll use my company again as an example. So, service for us, it’s a very important part of what we do. It’s critical that we know the data of the equipment the customer has, the billing information about the customer, the address, and when we facilitate and optimize that process of service delivery, that data needs to be part and parcel of that entire transaction.


