My Playbook for Data-Empowered Operations (261)

This week’s podcast is a quick summary of how to use data in operations. 

You can listen to this podcast here, which has the slides and graphics mentioned. Also available at iTunes and Google Podcasts.

Here is the link to the TechMoat Consulting.

Here is the link to our Tech Tours.

Here is my data playbook.

1. Sell data as a product. Or part of the product itself. But you are selling data. This can expand into selling services.
2. Analytics use cases are valuable. Think analytics vs insights vs predictions. 
3. Data can increase the speed of management decisions. Think dashboards.
4. Rate of learning and adaptation is more powerful version of this. This can be a big strength in some businesses. 
5. Data products supporting agile teams are important. 

------

I am a consultant and keynote speaker on how to accelerate growth with improving customer experiences (CX) and digital moats.

I am a partner at TechMoat Consulting, a consulting firm specialized in how to increase growth with improved customer experiences (CX), personalization and other types of customer value. Get in touch here.

I am also author of the Moats and Marathons book series, a framework for building and measuring competitive advantages in digital businesses.

Note: This content (articles, podcasts, website info) is not investment advice. The information and opinions from me and any guests may be incorrect. The numbers and information may be wrong. The views expressed may no longer be relevant or accurate. Investing is risky. Do your own research.

Transcript

:
Welcome, welcome everybody. My name is Jeff Towson and this is the Tech Strategy Podcast from Techmoat Consulting. And the topic for today, my playbook, I guess, for data-empowered operations. This is kind of, I guess, a reaction or a continuation of what I talked about last week, which was part of it anyway. it's kind of saying, look, I don't think data network effects really exist. I don't think data scale as a competitive advantage really exists. I like 80 % of what I hear about data as an idea and a strategy, I don't really pay much attention to it.  So that was kind of negative.  So, I thought I would sort of flip it around and say, okay, here's how actually I do think about data and sort of the playbook whenever I'm looking at a company I'm thinking about. So, let's say more useful. less just throwing stones at ideas that I don't really understand or use. Now I'm actually in Shenzhen today. I spent the last couple days at basically Tencent Cloud sort of big annual event.  Lots of speakers, lots of executives, really fascinating. I'm going to write about it. I'll talk about it next week.   Huge amount going. It's really kind of the big AI engine of Tencent. You know, it's AI plus cloud and that tends to be the center of the action. So, I'll talk about what they're doing, which is pretty fantastic. I also got to interview one of their senior executives who's in charge of a big portion of this. So that'll be coming out and one of their sorts of bigger customers, I interviewed them as well. So, a couple of interviews came out of it. Yeah, pretty fantastic. So, I'll write all that up and put it out shortly. Pretty great couple of days. My brain is kind of full.  So now I'm relaxing in the hotel flying out to Shanghai in the morning, but yeah pretty fantastic Let’s see standard disclaimer here Nothing in this podcast from my writing or website is investment advice the numbers information from any guests may be incorrect have used in the opinions Express me no longer be relevant or accurate Overall investing is risky. This is not investment legal or tax advice Do your own research.  And with that let's get into the topic Now the concepts for today, really, if you look at my digital operating basics, know, DOB3, which is basically build your digital core, which is now increasingly becoming an AI and digital core, you know, lot of it sits in there. Some of it sits within digital operating basics two, which is continual customer improvements, starting with personalization. Again, that's very, very data driven. And then you could also think of rate of learning and adaptation, which is one of the sorts of digital marathons. Anyways, all three of those basically sit under operating activities, which is kind of where I fall for all of this, like data, creating data architecture systems, all of that.  Overwhelmingly, it's operating activities. It's not a competitive advantage. It's not a mode. It's important.  Digitally wiring an organization is very, important. But I kind of put it all in there. And within operating activities, there's really four to five things, I think, which is 80 % of the action. That's where I think the payoff in terms of operating performance is.  So, I'm going to kind of list those. That's what I'm calling the data-empowered playbook. It's really operating activities. Now this is actually going to be short today. I know I've said that many times and it never ends up being short.  I think actually today will be short.  We'll see if I hold to it this time around.  So, you know data you heard all the time data network effects data economies of scale data scale advantages data products. You got to build data products in your in your business. data assets, data ecosystems, data flywheels, data feedback loops. I mean, it's kind of all over there., you know, it's kind of like you put data in front of everything and it, nobody really knows what it means, but at least in the last podcast, I'll start with the stuff that I think's not real. I had a couple of conclusions from the last podcast, you know, really answering questions is data a competitive advantage? and the answer is almost always no. Not a hundred percent most of the time. Data network effects.  I don't really see them. I can find one maybe data scale. You have more scale than your rival. Okay. Now as I kind of said last week the problem with that quantity of data is not usually the key thing. Well kind of scale is about quantity. I have more than you. have more factories than you. I'm bigger than you. Quantity data is a quality problem.  And as soon as you add quantity to data, usually your quality goes down. usually scale, not so much is data a scarce resource, which is on, if you look at my list of competitive advantages, one of them is a, is it a scarce resource? we have a piece of land on the beach. Nobody else has it. That's actually scarce. There's not that much land in certain beaches that are popular. You have a big building on Central Park West of New York. Okay. There's actually not that many pieces of land that would be considered a scarce resource. Yeah. Data proprietary data almost always can be a scarce resource in certain businesses. So that's a competitive advantage. You can go for. I agree with that one. The others I don't. It's not that common, but it can be.   Is data a barrier to entry? Is it something that can make it difficult for a   new entrance to jump in your business?  Yeah, kind of. You can see libraries of data. You can see a lot of user-generated content, maybe Wikipedia, maybe the YouTube library, creating some degree of a barrier. You can see a data ecosystem where you're pooling together five, ten, fifteen companies that are all gathering data together. We'd call that a data ecosystem. All of those things can raise the barrier for jumping into a business. Okay, kind of.  Generally, it's not that hard to disrupt. know, a lot of things can be a   barrier in life.  The question is, how difficult is it for a well-funded, well-run rival to jump into your business? What does it cost? How difficult is it? How long does it take? Timing, difficulty, cost.  Most data is not that hard to overcome. there's ways to get around it. You can do synthetic data. You can buy data. You can scrape data as a barrier. It's not awesome.  And generally, when people talk about data as a barrier to entry or an advantage, it's usually about some new business, some new product, some new industry where a couple player it's early days and a couple players have gotten ahead of everyone else like Tesla with its lot of data about roads traveled. They get better performance because they have more data and it’s still the early days of the industry. But as the industry matures is it going to still persist as a barrier or a competitive vendetta and usually the answer is no. Data is kind of ubiquitous.  So, with the exception of scarce resources, I'm kind of I'm dubious. However, the other two conclusions I had from last week The real question is will data enabled learning and adaptation become a sustainable competitive or operating advantage in the future? Will it become a smile digital marathon advantage or competitive advantage?  I don't know yet. I'm watching in theory. I think it could, but the real, the real issue is not data. I had someone literally interviewed someone yesterday who, you know, he used to be CEO of Lazada Singapore and stuff.  And he kind of said, look, data of itself is not useful. It's only valuable if you act on it. Otherwise, it's worth nothing. Well, maybe not nothing, but very little.  So, data enabled learning and adaptation, that can be an operating advantage. That can be a competitive advantage. That's what I'm looking for. I don't see it much. So really the question I'm mostly focused on is the one which is the subject of this podcast.  How can you improve your operating performance with data, with data enabled learning? It may not be a competitive advantage, probably not, but that's really the question. How do I dial this into my operations so my performance is better?  That's, I'm going to give you my answer to that question for today.  I really don't know how to think about data.  Are we talking about numbers in a spreadsheet? Are we talking about unstructured data, video, audio? Are we talking about tacit knowledge, information that's sort of in the heads of your employees or your managers?  It's not in a data lake. It's not in a data warehouse. It's in people's brains. All of that kind of data.  So, what I'm talking about is like digital data wiring.  think about it almost like electricity. You put in the wiring, the data flows throughout the whole organization.  Over time, it may result in intelligence, but I just kind of think about it like electricity.  You know, when I'm looking at a retail business, I'm looking at management. I'm looking at the number of outlets. I'm looking at the warehouses.  I'm not really thinking about the wiring and the electricity that's in all of those assets. That's just part of what they are. I kind of think about data the same way. If I was going to make an analogy, I'd say it's kind of electricity. Okay, so let me give you sort of my playbook, which is really just four to five things. data when it is a product itself that you can sell that's real. You can sell data and you can kind of generate a service based on data. So, consultants do this. Consultants might come in and say look let's assess how well your personalization engine is. Let's assess how good your HR policies are. Let's do your salaries and benefits assessment.  As part of that consulting engagement, they might sell you best practices or an index because they work with a lot of clients. They sort of aggregate best practices versus okay practices and they kind of rate you, but they're really selling you a data product. They've gathered the data, they put it in a report, they sell it to you as a product, they charge for it. Now if you can sell data as a product, okay, that's real. You can kind of go one step further and say, well, what if we turn it into like a digital service? We've sold you a tractor, but in this tractor now we have a lot of IOT, we have a lot of monitoring. we will basically sell you a service on top of that to tell you when it needs to be fixed or when you've got some issues of wear and tear or when we sense a product. Now that's pretty, you could call that a service, but it's pretty much just data. So digital services, data services, you could call them that you add directly to products can be a product in itself. You can sell it smart breaks. maintenance for factories may okay that's real and I would consider that a data product or service  that doesn't really work all the time you know if you have data monitoring in a television or an air conditioner  you're going to start talking about data and tracking and monitoring but you can't sell it as a service you know I'm not going to spend money every month to get data about my air conditioners and refrigerator even though they are digital associated with the product so that sort of me the first level look can you sell can you sell it if you can sell it okay let's probably do it that's number one it's a product you can get a paycheck for it  number two that's pretty rare actually okay number two Analytics use cases. Okay, this is kind of your digital operating basics number three. This is building your digital core. You know, we pull all the information in one area, maybe the full company, but in one area. Let's say we pull all the customer related information into one massive database.  All the CRM and customer facing capabilities, all that data comes into one place. That's what Salesforce does. And we can start to do analytics use cases against it. Fine. That's in the digital operating basics.  A lot of companies have been doing this for years. A lot of companies are just starting now. That's OK. That's usually a first step.  And we get the data together. But again, data is not useful unless you act on it. So, we ought to look at analytics use cases. That's usually backward-facing analysis. Analytics are good in their own sense, their own right. You can quickly go from analytics to insights.  Let's discover things about our customers we didn't know before. We're not just checking performance. We're looking for new insights into the market, into our customers, into our competitors, into our operations. Okay? That can then go a little bit further and we can start talking about predictions. Let’s look at our customers. Let's look at our operations. Let's look at our market. Let's look at our pricing and let's start to have AI make predictions. Okay, that's all digital operating basics three.  You can start with internal data, but a lot of times you start to bring in external data and you want external insights. We want to make predictions about the market. Okay. That's pretty standard stuff. Here's where I kind of disagree a bit. I think the whole argument about customer insights, I don't really buy it.  think analytics are very good. Historical taking apart everything, seeing all the numbers, dashboards. Awesome. I think most customer insights, they’re kind of data driven, but usually they come from the judgment of the team more than anything else.  They're pretty easy to copy.  you know Amazon has lots of data about me and then when I look to buy I don't know book A they suggest I buy book B or I'm going to buy hot dogs and they suggest you should buy ketchup as well and mustard. Okay those would technically be insights but everyone kind of knows this stuff and if you see your competitor doing that you can copy it. So, this idea that you hear so often that we're going to build a feedback loop we're going to get lots of customer data and we're going to use that to generate insights and create new sources of value and then we're going to increase our sales and we're going to get more customers and it's a feedback loop. It's mostly not real. I mean you have to do that anyways. Fine. But this idea of a data driven feedback loop that generates customer value. Any insights there are so easy to copy. It's not a big deal.  Now what I do think is real there, which is what I focus a lot on, is when you build a customer improvement engine with lots of touch points, lots of data and the ability to engage quickly, that’s real. And literally I have a whole playbook on this. which is how to build a customer improvements and personalization engine into a company. And it has data, but it has teams.  It has decision, but there's a lot in there. So yeah, but generally speaking, that's number two. Look, analytics are great. The insights I think are overblown. Predictions are very helpful.  Internal versus external data and analytics is very useful. I think this whole customer insights you know, feedback loop. I don't really buy it. I think you have to do what I do, which is you have to build a sort of an improvement engine, which is a combination of tech architecture, data, teams, and operations. They all have to click together for that to work. That's number two.  Number three, and I've only got five,, giving management,  ,  dashboards that enable them to make fast decisions. That's very useful. That's pretty powerful actually.  And this is the idea of like, look, speed is a big deal. If you can make decisions quicker, that is, that is really going to change your operating performance. And I actually ask this of clients all the time. You know, how quick do you make decisions? How quick can you, we have an idea.  Let's go test it and let's go.  A company that can make decisions quickly on the spot because you have the data in front of you, dramatically faster and better than a company that has to generate reports. Reports take a long time to write.  They got to go up the chain to traditional hierarchy of management. Management needs to read it. Then they need me to make a decision. Then it comes back down.  Usually what you want is you want your management team around the table. You want a nice clean dashboard. operating performance.  You put that together with everybody's brain and you make decisions in the room. Right. That's kind of the Jeff Bezos, Amazon approach, which is, know, we have people, we have meetings for the purpose of making decisions. So, we feed data into the room, everybody, and we make the decision. Right. I've said this more like the, the analogy is like, you know, if you're playing chess against someone and you get to move twice for every time they move once. you're almost always going to So but the key there is speed. You want to have a metric for how fast your team is making decisions.  And yeah you should track it. I like to have a number for it. OK number four when this all starts to become much more interesting, everything I kind of said is sort of digital operating basics and you know, I call them the basics for a reason, but number four is where things start to get really much more powerful, which is when we start talking about rate of learning adaptation. You know, when we can start to feed data into the system, find customer improvements, do personalization, and start to impact our customers, if not in real time, fairly quickly. Ideally, it's in real time. We see what they're doing online today.  We hit them for promotions immediately.  That is pretty powerful. So, if you can go from digital operating basics to customer improvements to a personalization and customer improvements engine, that's a big deal. And that's when I start to talk about rate of learning as an operating advantage.  That is pretty much where I spent half my time.  literally spent half my time thinking about this question and the other half thinking about how you build a moat in a business. How do you win. But it's pretty much these two things. That’s 70 % let's say. Okay. Number five   data products that you can give if you can move your business to agile team structure sometimes are called pods. McKinsey talks about this all the time. You know you want to have 20 30 50 pods are just teams right teams of five to ten people that you can basically throw them a problem and they will run with it and fix it.  Usually for them to function effectively you got to have the right people which is difficult. know, big companies have hundreds of pods, but smaller companies, if you have five or 10 pods as a medium or smaller business that you can throw problems with and just trust them to execute and implement without it coming up to the boss, that is a fairly powerful way to do certain things. But for those pods or agile teams to work, you've got to give them usually data products. So, you need some standardized data products that draw, they don't just have a data lake and a data warehouse.  They're drawing that up into a very useful form, which is usually called a data product. And your customer for your data products is your internal teams. That's pretty powerful. you see a lot of companies like especially the larger companies, the multinationals, know, fortune 500 companies. When you start talking about digital transformation for these large companies, you spend a lot of your time talking about building data products and then training your teams because you got to get the skills and then putting them into pods. Like that's a lot of the playbook. Now, to do the data products, generally have to have a technical architecture underneath that. there's a, there's another layer, but yeah, that's kind of the standard number you always hear from a McKinsey or a BCG is digital transformation is like 60 to 70 % about people hiring people, training people, upskilling people and putting them into pods. 20 % is data. you know, 10 % is sort of architecture and other stuff. Now that's not exactly true, but it's actually a good way to think about it. So, if it's about people, 60 to 70%, you got to give them very usable data products they can use to execute with. And within these customer improvements, feature improvements, launching new products, all of that's real. And also on the non-customer side, the operating efficiencies.  operating changes, creating smart factors. That's all real too. So, but all of that depends on sort of the data products.  So that's kind of the basic playbook.  I would say up until for me up until last year, two years ago, and then generative AI sort of comes on the scene and the whole data architecture data question just gets turned on its head. And this is really the last topic for today. And I don't have any answers on this one yet, but you know, once you start talking about generative AI, the whole data thing's crazy. You know, suddenly we're talking about lake houses, you know, combinations of different types of data, a lot of which is unstructured.  We start talking about things like tacit knowledge that isn't even in cameras or spreadsheets or emails. It's in people's brains. the technology starts to change because it turns out, you know, the generative AI tech stack is kind of completely different. It seems like you're building another one.  Huawei has some outstanding, when you start looking at the infrastructure, the servers, how the chips work together.  And on top of that, how the data is processed. I mean, generative AI, you need an unbelievable amount of data, right? It has to be like a river. that's just flowing in there all the time and you know the algorithms are running and coming up with predictions.  So, Huawei which is really in my mind one of the leaders in Asia if not the leader when it comes to hardware server racks cooling systems they're crazy they have all this stuff and they have a data architecture that they're building which is really global. It's big in China now they already have sort of these main clusters they put together but   they're building across Southeast Asia right now and they talk about data oceans versus data lakes versus data rivers.  It's really awesome actually I can't publish it yet. I've asked multiple times can I write this up and they're like no don't talk about it.  So, I can't really do that but I will soon.  Ant Financial talks a lot about how business if businesses water it used to be like a well. and you had a well on your own company. And then we started to share data between companies and that was like handing bottles of water.  And then it starts to go, okay, maybe we have a couple pipelines here and there between a couple companies. And now they talk about it like major water works, like dams and just data flowing everywhere, all throughout industries and people are capturing and using, it's crazy. There's all this stuff that just sort of comes and then it very quickly goes from, know, this idea of, we're going to have feedback where we're going to have people use machine learning and that's going to generate data for us and we're going to get smarter. So, data starts to become a system of intelligence and this is really what by dues focused on by do know the four AI cloud companies. am really spending a lot of time studying our 10 cent Alibaba, Huawei, and Baidu. Baidu was one of the first, because they're really good at AI, to sort of talk about feedback loops between, as people use our algorithms, as they start to build custom models and deploy them in industry, and they are very industry-focused, that's going to create a feedback loop that's going to improve our knowledge maps. And we are going to start to build models and knowledge systems. that are industry specific that are smarter. And they're focused a lot on manufacturing.  So, this whole data river thing starts to become a system of building knowledge and building intelligence. So, are we even talking about data and generative AI? mean, there's a lot of architecture to build.  Or are we really talking about operating intelligence and knowledge systems? That's really the question I'm thinking about.  And that very quickly turns into a discussion about let's build vector databases.  Like forget all these SQL databases, all that. No, but we're talking vector databases because they can actually capture knowledge. And you start to put them together and that's a whole world of expertise. So, we're building intelligence and knowledge systems. OK.  that's cool. How do we build that into a business? Well, this book I mentioned the other week by Sangeet Choudari   reshuffle. You know, he talks somewhat about building an organizational brain within your company. When you build knowledge systems, which are based on data, and you start to build real expertise in your organization, you know, it's like you have a brain that's in your organization. And one of the first ways you use that isn't you feed it to your agile teams. So, your pods instead of having a data product it's like they have an extra team member that is virtual that is the organizational brain of the company and it knows everything the company knows and it knows a lot of knowledge from outside but that's really a member of your team and suddenly your agile team your pod is dramatically smarter than it ever was before. So that would be sort of an organizational brain. The other place you can start to maybe put this in, which he talked about was the problem with agile teams and these pods is they get up. They're very effective as standalone pods that can do things, build features, fix problems, address pain points, prototype new products, anything. but then you have what he calls a coordination tax that you have to pay because all the pods have to link to the other pods in some form to capture the knowledge that each team is doing.  And that's usually done by reports. It's done by lots of meetings. It's done by lots of zoom calls. That's a coordination tax. Well, an organizational brain and intelligent agents can basically do that across the company. In theory, you don't need meetings and reports anymore because these agents will capture everything and then every team will have sort of a, an organization brain presence and they will handle all that coordination. So, you don't have to do all the meetings and write everything up and put it in the, you know, the file cabinets. It'll just get captured and shared easily. So that's a couple of things. Okay. The organizational brain makes your teams more effective. You can have organizational agents that handle all the coordination and knowledge sharing across the organization.  The big topic for the last two days here at Tencent was AI agents, multi-agent systems.  They are leaning into this in a major way.  And they're already starting to deploy them. So, AI agents that have the knowledge.  to do things internal to the company, do things external to the company, to be present on teams, to be sitting next to management. When you start to put intelligence in agents in your company or external, that's a big deal. And that's kind of where things are. going to, I'll talk a lot about it, but it's pretty much the, you know, the big topic for 2025 is AI agents and the fact that they're becoming smart very quickly.  that's kind of it. So that last point, I don't know, it's not really in the playbook. It's just kind of my,, the questions I'm thinking about right now, but yeah. So that's five points that I would put in sort of a standard digital operating playbook,, related to data. I'll write those up. I'll put them in the show notes. I won't go through them again, but, yeah, pretty awesome. Really. I think it’s,, it's fascinating. And so Anyways, that's it. I'm flying out tomorrow. I'm going out to Shanghai. I'm going to go to the Huawei Connect conference. This is like literally my favorite Huawei conference all year.  This is the conference basically for developers. So, you know, it's all the architecture and it’s deep in the tech. Basically, it's deep in the tech and it's deep in the business use cases. It's really for developers, which I am not, but I love the, mean, I carry like literally like suitcases worth of documents away that I end up reading.  And some of it, I understand that some of its way to, you know, engineering for me, but yeah. So that's going to be three days in Shanghai.  Fantastic.  So yeah, it's, it's been a great week. Anyways, that is it for me., I'll be back with a lot of content very shortly, but I hope everyone's doing well. Talk to you next week.  Bye bye.