Tim Gasper [00:00:04] Welcome. It's time for Catalog and Cocktails, your honest no- bs, non- salesy conversation about enterprise data management, presented by Data. World, coming to you live from Big Data London in London at the Olympia. We're very excited to have a special guest today. We'll introduce him in just a second. Juan, my co- host, Tim Gasper, also here. Juan, inaudible.
Juan [00:00:25] How are you doing?
Tim Gasper [00:00:26] I'm very excited. This has been a great day.
Juan [00:00:28] Finally, finally we got the infamous Chris Tabb.
Tim Gasper [00:00:34] Chris Tabb.
Juan [00:00:35] The co- founder of LEIT, Tim told me to say LEIT Data. But most importantly the mean data streets, the LinkedIn mean data streets.
Chris Tabb [00:00:43] This is a high level, high passion data layer. It's great to be here, you two guys. I think I first gate crashed one of your shows you were doing with Joe, the last 15 minutes, I inaudible.
Juan [00:00:54] I remember, yeah.
Chris Tabb [00:00:56] When I was watching you on there, I think, this guy goes as fast as me and talks as ramsey as I do. So.
Tim Gasper [00:01:03] Data ramsering crew.
Chris Tabb [00:01:04] Yeah. I'm going to try and get to the level but we're trying now with inaudible.
Juan [00:01:09] All right, let's kick it off. So tell also, what are we drinking, what are we toasting for?
Chris Tabb [00:01:12] We're toasting to Big Data London for a big success and for the high performance data debate that starts at 6: 00 PM within Big Data London in the theater called the Analytics Decisions.
Juan [00:01:24] We're live, it's 4: 00 PM UK time right now. We usually use 4: 00 PM central time but get all coordinated after this we're going to your panel. I'm going to be on one of those demos.
Chris Tabb [00:01:33] Yes, we are. Yes.
Juan [00:01:34] All right, so cheers to Big Data London.
Chris Tabb [00:01:37] Yeah, cheers.
Juan [00:01:37] This is a very packed event.
Chris Tabb [00:01:38] Yeah, it is. Epic, epic event.
Juan [00:01:42] All right.
Tim Gasper [00:01:42] You have one question, huh?
Juan [00:01:44] I have a question. So mean data streets, mean data. So what's the meanest street you've been on?
Chris Tabb [00:01:49] I'd say it was the old Microsoft mean data streets of DTS.
Juan [00:01:55] DTS.
Chris Tabb [00:01:55] That's where, if anyone knows, it goes back, it predates SSIS. But I suppose jokes aside, that's what a technology mean data streets in. I suppose where the mean data streets were born. And the term, I suppose it came from another LinkedIn rant sometime where people were talking about how tough things are and this not working and why is it so complex and why have we've got so many moving parts and create the modern data stack. Modern data stack, mean data streets. So the mean data streets were born. I've walked many, many mean streets in my life. Not all of them mean data, but yeah. I suppose...
Tim Gasper [00:02:39] I didn't fully realize the acronym was meant to align.
Juan [00:02:43] Yeah, I just realized right now, MDS.
Tim Gasper [00:02:45] MDS, yeah.
Chris Tabb [00:02:46] I'm thinking we're in the business data stack now, so we're in the BDS era. So I think I did a post about this a little while ago where...
Juan [00:02:54] On the bullshit industry?
Chris Tabb [00:02:55] Yeah. I'd say we are the honest no-bs one. Not a bad term but... I think the prehistoric data streets were back in the days, the modern... Yeah, that's the beginning of that. Maybe the days where Netezza, Teradata ruled the streets for them. You don't hear much of them. Well, Netezza acquired by IBM years ago.
Juan [00:03:27] Teradata is there.
Chris Tabb [00:03:29] Are they here?
Tim Gasper [00:03:30] No, I don't know.
Juan [00:03:32] No, I don't think they're here.
Tim Gasper [00:03:35] They're around.
Juan [00:03:36] Okay. Have you been on any mean streets?
Tim Gasper [00:03:38] Mean data streets? I don't know about mean data streets. They've been pretty nice to me but you know what, I grew up in Cleveland, Ohio. We've got some mean streets in Cleveland, Ohio.
Juan [00:03:50] And I grew up in Columbia.
Chris Tabb [00:03:52] Well, okay. inaudible.
Juan [00:03:53] I can think about streets in Colombia.
Chris Tabb [00:03:57] I think the streets I grew up in Praklaw are not quite as mean as that.
Juan [00:03:59] All right, well let's kick it off. All right, let's keep it very simple, honest no-bs. What do you mean by business value here for these data folks?
Chris Tabb [00:04:08] I did a post about this recently, is it an overused term? Everyone starts talking about value and the work we've been doing with Matt Housley. We've been trying to break down just that term, business value, and trying to articulate that into one sentence. We've done a few iterations of that and the current iteration is, business value is a evidenceable positive impact on your business performance or your company performance. We've had a few, is it all about profit? I don't think it is, when we're doing the work on this, not everything can be directly attributed to profit, and I suppose there's a time dimension to it as well because if it's intraday, sorry, within a year you're looking at adding business value within a financial year. That's what your CFO is going to want. You take the probably extreme example of Uber. Uber is operating$83 billion, I think, they'd run until it started making a profit. So that's a long runway and anyone out there that's been in the startup world, that's a lot of money to go and test the theory. So that was over 10, 12 years I think, I can't remember exactly, 12 years. So I don't think anyone would've gone, "Right, it's a really good idea, in 12 years we're going to make you some money. By the way, you want some more? You want some more? You want some more on the journey?"
Tim Gasper [00:05:39] It's pretty amazing how some companies can just kind of keep going without a profit for quite a while, right?
Chris Tabb [00:05:45] Yeah. And I suppose it's like the money picks, I think, once you've committed to it you've got to make it work, otherwise you've got no other options. But obviously they have made a profit now and everyone knows who we are.
Juan [00:06:00] But that doesn't mean that data teams can go off and be-
Chris Tabb [00:06:02] No. I think that's part of it. Thing is, those days are gone and I don't think they went into that approach thinking it's going to take them that long. But I suppose it's what mode. That's the term we think is business value. Then there's what business value mode your company is in. If you think of, I mentioned earlier, your CFO wants to see some gain, some value within that financial year. So you're in a fiscal mode, it has to be within that one. If you've got someone that's forward- thinking, that they've got a strategy of getting a better market share, four or five years you're in growth mode. The third mode we've named is a hybrid. So it's where you are using optimization, so cost savings to fuel or either pay for entirely or contribute towards some growth activities. As with any planning activity, people throwing money in it, you can keep... If you have some controls and make sure that okay right, we're going to look to see a balanced approach of where we can save some money to help with refueling any growth fashion.
Tim Gasper [00:07:21] So I like these kind of three modes that you're talking about here. I imagine the hybrid mode is going to be some mixing and matching between sort of the first two modes. Maybe let's focus on those two modes a little bit. Fiscal mode, growth mode. So maybe fiscal mode, that's the first one you mentioned. What would be some examples of evidenceable positive impact when you're in fiscal mode?
Chris Tabb [00:07:41] It's all about the time you can actually make change in your organization. So I've got to think about competitive advantage, but this thing has some value to it as well. But say for example you're in very good organization and you get new features out regularly, you've got opportunities to satisfy that fiscal mode scenario. So your CFO's going to say, "Right, this is the budget. I need to save some money otherwise I'm going to shop that project that you know is going to be your future for the next year." And you do that, you think, well the business is not going to survive well out there. So you need to look at what you can do with some quick wins. I hate the term, the low hanging fruit. And you look at... I'll give you one example of onboarding process. I like to use the term friction sometimes as well, is that friction is the time complexity and effort to doing something. If you are signing up to a new service or something like that and you've got to fill in so many forms, you've got to then go submit and wait to come back for a verification. That process you'll get dropped off. In banks, you're going to have KYC checks as well and depending on what people you use, how sophisticated they are, you may be refusing people that actually were good or other way around. So you may be taking people on that actually were fortunate. So if you can target those two scenarios and optimize that within a financial year, you can decrease the cost of conversion, converting new clients or onboarding new clients. You are going to inaudible clients going to come on board that year, they're going to add more money, more revenue within that financial year. There's an example of fiscal. I think the growth one is... Big Data London is sort of a good example of that. There's a lot of vendors out here all in the same place. They don't have the luxury of just waiting to see if they can organically grow. There will be loss making for a while. In the same way in your organization, taking out the technology or product, it's all about getting customers. And if you are going to have to do some deals to go and get a larger market share, and even run at a loss for a little bit or run at a... Maybe not a full loss but it's subsidized though you're not making the margin that you would expect or maybe your shareholders would even want at that point. But you know if you can get them on board, you've got a sticky product, that means that they're not going to go off anywhere else and you now have got four or five years worth of a loyal customer, that all adds up. So in year five, the profit you provide that one is going to be much more than if you've gone in a fiscal mode. I think things that fall into that growth mergers, building a data platform, building a data capability in the organization, it costs a lot of money. Go back to that business value or the term ROI that was used earlier as well is, sometimes it's very hard to justify that and you need to build credibility, you need to provide evidence of all the things you can relate to. And it may be for the first period of a company, that overhead, building that platform, getting that good DataOps capability in place. Getting a good data cataloging or getting the business working well, streamlined with your dataset. That may take 6 to 9 months or 12 months to get it working really well. Hopefully, you may be able to do some things on the side that can keep the business happy and not just have... Having incremental value while you're on that journey. But sometimes it's not really possible to actually get it within that fiscal year. I think in the book, one of the sub-articles that should be out soon, we talk about... It is that evidence piece as well. Some things you can't always evidence and some things you're going to need to... You know that they're going to be needed in the long run to build up your data boost or data plan.
Juan [00:12:15] I got a couple of things. One is-
Chris Tabb [00:12:20] Can I have a drink really quick?
Juan [00:12:20] Yeah, go drink. Got to get your... That was a lot. But we're taking notes, we're taking notes. So one thing, on the fiscal mode, I think this is kind of like straightforward, you got to make money, save money, how can you help them do that? But one thing that I'm really taking away here is the friction. When you identify where that... You should be looking for that friction and figure out that's energy, those are people, let's go figure out what's there because maybe there's an opportunity to go help. So I think that was a really key thing that you said.
Chris Tabb [00:12:47] It's a good thing. How I link that to business processes now, is because as soon... You can't measure friction until you have a business process to hang it off.
Juan [00:12:57] This is something that is coming up a lot with the folks we were talking about, it's about let's go figure out our business processes and the business processes, it's connected to everything. This is why I love crafts. Everything is connected and then what goes into the processes? People, decision, all that stuff. Data goes into these processes. That's why everything is connected. So fully agree. I think we need to get a focus on that.
Chris Tabb [00:13:19] Yeah, and we're talking, again, companies. I'll tell you what, I've seen business processes implemented, and why I think it is a missing piece of puzzle that I don't think companies have enough focus on. I've only ever seen them defined in large transformations where you need to go in there and you're merging, maybe a merger acquisition or... People have to understand it because you have to work out how you're going to separate these two organizations or combine these two organizations, or where you can combine parts of the process and see where the best ones are. I've never seen any company ever maintain those processes after the event, and I'm sure there are... I know there's been business process changes, they don't go and change it. I do think there's a... Data modeling got brought back to life, it kept the CPR, it was dead for a while in the early days, the modern data stack. It's come back and it's really good that it's come back. I think that whole taking that modeling and what you talk about, the knowledge graphs and everything all being linked, it is all linked. The easier we can collect that associated metadata so we can actually look at how everything's all joined up. Your operational world and your analytics world, sharing that same knowledge, bringing that all together, you're going to have that competitive vantage I just mentioned earlier.
Juan [00:14:42] I'm not saying it, you're saying it here that having all this metadata, getting it connected, that's critical. That's critical for your operational efficiency within the organization.
Chris Tabb [00:14:52] Again, I always refer to posts I've done because that's how my brain farts and how it gets things out. But yeah, I call it the meta metadata. I think you've commented on that as well, is that everyone knows on sound data now, the benefits it can provide. I think it's now that educational piece of okay, all of the data about the data and how that data about the data is used collectively together. There's so much value in that. We go back to business process optimization, fault detection, frauds happening, impact assessment of changes happening. The more you've got things connected, the more you can be confident that if you're going to do a change, you're not going to break anything. Going back to that competitive advantage bunch again is if you are able to make a release every month, you are going to stay ahead of the game. You are going to... I have this thought, I talked to Craig quite a bit where he is talking about how you actually measure the associated value and it is hard to do that.
Juan [00:16:06] So the other thing that you brought up about on the fiscal mode and the growth mode is... You brought up the growth mode, you were saying, we need to go, for example, we're going to invest to have a data capability organization. How would this work in the hybrid is like, well, you can have a push and pull, it's like well, I'm going to go do this very quick ad hoc, it doesn't scale but it doesn't matter if inaudible now. But then you're not investing in the data capabilities. So it's always like the balance of it, what I call efficiency resilience. What are your comments?
Chris Tabb [00:16:36] There's another view is that a salesperson, if they have three bad runs or two bad runs in a row, they're out the door. I think as a data team, a data department, you can't wait... The CDO tenure time of how long they last is 18 months. Reason why is because they've been given that much time to not deliver what they said. If you're in that hybrid mode and you need to be able to create that incremental value, you need to be able to have a good story at least every quarter. All right? To keep the stakeholders on board.
Juan [00:17:12] This is a really good point.
Chris Tabb [00:17:12] So it may slow you down and that may not be the most optimized path to get to your vision, but you're more likely to achieve it because you can keep the stakeholders on the journey. If you make sure there's some quick wins. So there's something you've built that maybe it's deviated slightly off to the, I suppose the quickest route that you could have gone. But you've had that good story because if you wait for that good story, 9 months, 12 months, 18 months, everyone's lost interest by then.
Juan [00:17:49] This is an excellent point. So it's not like on one side I'm only going to do low hanging fruit quick wins or it doesn't mean I'm swinging the pendulum, I'm only doing the hard things, getting to invest in it. It's like no, keep some quick wins and balance this out. Now what is the percentage? Because when people are asking, " What have you done for me recently?" You're like, " I can show you the quick wins." Better than showing nothing instead.
Chris Tabb [00:18:07] And it links back to this, the thing we were just chatting out a bit earlier. That word evidenceable I used earlier. If you take the legal use case or a legal world, when you are going to take your legal case to... You've got all the evidence, you've got to have a really good case to actually win that case. There's different sort of levels of courts of law and some will be a judge and a jury, and some will just be a civil case where the judge is actually this one person, he or she will make that decision. So you need to start building up good stories that get you to that last, a big criminal case. You've got the judge, the CEO, you've got the CMO, you've got the CFO, you've got the CTO, you've got the CIO, you've got your risk officer as well. You have revenue office, you've got the whole C- suite as your jury.
Juan [00:19:09] Who's the judge?
Chris Tabb [00:19:12] The CEO. If over that time you've had a number of small cases that you've won over the CRO, you did that for him, he loves you now. You then won over the CMO, seemed really happy with what you did for her as well. She's great. Then you won over the CTO because you helped with some... You've helped all of them at some point. When you go and put that business case forward, you obviously have to make it a good business case. But the jurors, you've got credibility, you've got reference cases. We go, " As with case law 65 on this day here, when I delivered this capability for you, was it success?" "Yes." "Did I deliver it on budget?" "Yes." Have I got credibility now?" " Yes." " Judge, will you sign off this project for the next 18 months or 5 million?" " Yes." But you'll never get that unless you've got all those things. And that may have taken you 18 months to get that credibility or get that assurance that you are going to spend that money wisely. I think you'll see CDOs in longer tenures. I think you'll see a CDO becoming five or six years because they are worth more. They've made a difference.
Tim Gasper [00:20:24] I like this analogy a lot. I feel like it applies well to the politics that you have to play as you're trying to get your ideas through. And for those who are newer to the organization, who don't have that credibility yet and things like that, you have to work that much harder to prove, " Hey, trust me. This investment that I'm going to do." Especially if you're in a hybrid situation and you're talking about some growth oriented investment and activities that you want to do, you're asking people to go on a journey with you.
Juan [00:20:55] So thinking about who heads the data teams, the CDO itself, should they have a tech background or should they not have a tech background? They should be more for the business. What background should the person have to be able to succeed, if we follow this analogy? By the way, it's interesting, we've had many-
Chris Tabb [00:21:17] It's a great question. My best lawyers, is that what you're saying?
Juan [00:21:17] We've had many episodes where we take an analogy and we just keep banging on it. Let's see how far we go into this one.
Chris Tabb [00:21:22] Well, I think I'm going to give you, it depends.
Juan [00:21:23] It's a hybrid.
Chris Tabb [00:21:24] No, no, I'll go through with it depends. I'll give you two scenarios. So it's not quite a cop out, full cop out, it's half cop out.
Juan [00:21:36] Appreciate the honest no- bs- ness.
Chris Tabb [00:21:38] It's honest bs data. Let's keep it real. The things it would depend on is the size of the organization, maturity of organization, size of the team.
Juan [00:21:49] That's a fair fit.
Chris Tabb [00:21:52] But any CDO out there, they need to have some technical capability. They really need to be business focused and the more I think about the role of the CDO, I don't think we've had the right CDOs or the right people.
Juan [00:22:08] Yes, this came up recently. Actually, I was at the CDOIQ conference, and so the CDO of Visa, and he's non- technical. He said, " Everybody's talking about 18 months or whatever," his point was, " I think we're hiring the wrong people."
Chris Tabb [00:22:26] Yeah. Well, having the word data in it may steer you down a particular route, but if I go back to that, the things we talk about, about actually achieving this, you need to know what maybe the components are. You need to have a good architect working it, that's giving you that vision. You go, " Right, okay." They need to be mapping on how do I achieve each of those little steps with the guidance I've got, which is the right order to do these things in. The main thing is selling, marketing. You want to create that FOMO experience in your whole organization and go, " I want what that team department's got. I want what that department got." This leads me on to the other thing that... I'm probably jumping ahead on some things you're going to say, is creating that FOMO experience. [inaudible 00:23:220] into it.
Juan [00:23:21] You're looking at our notes.
Chris Tabb [00:23:24] I don't think the business understand what the data team does. Again, I did a post about this other day. But if you went to the business and say, " What does HR do?" They manage the onboarding of our staff. Okay. A bit more than that, sorry if it upset the HR. But yeah, you have a good understanding of what they do.
Juan [00:23:44] Yeah. They can give a crisp answer immediately. This is what we do.
Chris Tabb [00:23:47] It's a clear... Finance, in your accounts inaudible. Sales, they sell our products. Marketing, they market our products. What do data team do? They run extracts and do some reports mostly. No, no, that is not right. What the data team does is provide you the best opportunity with the use of data to make a positive impact on your business performance. They are there as a service to add to that business value that we talked about. Increase your business performance. They should be seen as a capability, not as a function that is just writing an extract and building a new dashboard.
Juan [00:24:27] What you just said, I agree. Now, wouldn't that fall into operations? What I'm driving towards is, today, a lot of the CDOs that we see or data teams, they work to the CIO, the technology side. CIO, right?
Chris Tabb [00:24:46] Yeah.
Juan [00:24:46] Which is a cost center, that's why they're not... It's hard to make it close because it may make these cases for innovation even things out. What if we started moving towards the operation? So we need to get data teams, CDOs should start reporting to the COOs or...
Chris Tabb [00:25:04] The reporting lines, it always comes up. I think we have confusings by having a CIO, CTO and a CDO. I think organizations that have the whole lot, I'd love to see the responsibilities between them all and actually understand, actually see that there are clearly no overlap. I worked in situations where they have a CDAO, because they can't decide which one's doing... And then do the underlying technology that's on there, and it becomes very hard to understand what's their remit and what's on their remit. So yeah, any advice out there when you're sitting in front of a... really map out what activities each, inaudible each of these people have. And then also, when you define that, are you setting people up for failure? Maybe it's another reason. Why CDOs fail at 18 months, because they've been set up to fail. They have not been given the ability to succeed or option to succeed because they don't get the budget for the technology, they don't get the budget for the teams. They're not in control of a cost center. I talked about that hybrid mode of... You need to be seen on the business value tax as well or putting your socks on tax, should I say. Is you need to take that money and own that and be responsible and accountable. That's the other-
Juan [00:26:30] This is the accountability.
Chris Tabb [00:26:31] You've got to have accountability stuff because it's not cheap signing up to all this stuff. So make sure you're getting your best bank for your buck. Going back to that selling rate. Make sure you don't commit to large projects that you haven't proven it's going to work. So yeah, I think it's that iterative way of providing value but also an iterative way of building up your vision, your data product vision, your business value stack or whatever you're going to call it, the business data stack. Be accountable for it, and don't have an emotional attachment in decisions that are already there. We need to use this we've already bought it. Do the right tool now.
Tim Gasper [00:27:25] Don't get caught up in some cost or what happened in the past. I like that you're kind of implying here that you give the CDO a massive agenda and then if you don't give them the people or the budget, how could you expect them to be successful? Right?
Chris Tabb [00:27:38] Yeah.
Tim Gasper [00:27:38] I mean of course you're setting them up to leave in 18 months, beyond even potentially the skill mismatch. You mentioned the data team and how they're a service to help other groups to help achieve business performance. In that sense, let's come back to our analogy that I'm hoping we can continue to extend here. Is the data team not really the lawyers, are they the paralegals that are just helping out?
Chris Tabb [00:28:02] Okay. I like it, inaudible. The lawyer is your representative of the data team. So yeah, he's had his whole team of experts, which is a whole team of experts that are helping build up that business case. I'd like to think it's the CDO who's the lawyer, and then the rest of the team are doing all the background, the research, verification, advisory as well. You need experts and I suppose they'll be your expert witnesses that you've spoken to that you'll wheel them up if you need to.
Tim Gasper [00:28:44] Interesting.
Chris Tabb [00:28:44] You'll wheel them up at the right time.
Juan [00:28:47] The expert witnesses are other people within the organization or the-
Chris Tabb [00:28:50] I think it's both. You're right actually. The problem with using the expert witnesses from within your organization, they can't go in and... Well, they can become the expert witnesses if they have proof, credibility and have delivered value previously, then they're your reference cases. Otherwise, it's wielding someone in front of an organization. There's that, or maybe a consultancy company. So someone that's going to go, " We've seen this, we can tell you if you do these things you'll have this impact. We've seen it before."
Juan [00:29:21] So these are external people outside company?
Tim Gasper [00:29:23] Yeah. Expert witnesses. So in the court case you bring in the psychologist or something to say-
Chris Tabb [00:29:27] Exactly.
Juan [00:29:27] ...here's how the mind works and why the criminal thought this.
Chris Tabb [00:29:30] Exactly. The sleep expert.
Tim Gasper [00:29:32] So here's how governance works and here's how you should actually do it. Right?
Chris Tabb [00:29:35] Yeah. Well, the lip- reading expert. Someone is an expert in a particular field.
Tim Gasper [00:29:39] Yeah, body language expert.
Chris Tabb [00:29:39] Yeah, exactly.
Tim Gasper [00:29:43] That's interesting. Two things I think I'm hearing here, is one is that sometimes it's the data team providing the evidence and it's the business kind of advocating. Sometimes it's the data team advocating and the business is providing the evidence. So it's kind of a mix of things. It feels like you're advocating for a pretty strong role for data teams around business value versus the flip side being kind of what I mentioned around the paralegals who were like, " Hey, we're just managing the bits. You're the business." Right?
Chris Tabb [00:30:11] When you go back to the... You've got forensic experts in that sort of DNA evidence or footprint evidence or even go back to the Dexter days of blood splatter, trying to say he's very dark. But there are experts in each of these areas that will be able to provide historical evidence from how they've done this before to prove and provide that, find that KP. I think your data team, external inaudible. Not everyone has to be... When I say the data team, it could be some external people you brought in as part of that data team. So it's a CEO's data function. You can even brought some partners in, but it is that full capability at its control. The other thing we talked about, business processes and the logging the metadata, you need some experts in that to go and analyze that data to give you the results back then. And you need to make sure that they're accurate. Because if you go stand up in that court law, and you go and present inaccurate information, you can be struck off the bar or whatever that legal terms. You can no longer have the credibility in the industry because you've seem to be fraudulent or not delivered or don't have that credibility anymore.
Juan [00:31:34] Now, I'm excited we're going on this analogy. I was seeing how much we can take it further. It seems like, if we follow this analogy, you're always in court. That's not really where I want to always be, in court. I want to actually never be in court.
Chris Tabb [00:31:52] Exactly, exactly. That is the end of objective and this is why I spilled to another analogy.
Tim Gasper [00:31:59] I love mixed analogies.
Chris Tabb [00:32:03] What you want... I'll go to an electric car scenario. Okay? You need to get to your journey, but you know that you're going to have to refuel regular, and you want to make sure that that refueling process is the most streamlined activity that you have to do. So the scenario you meant to report before it's optimized, you have to go to the... What you'd have do, you'd have to drive your car along, you go and plug it into a standard plug, it's going to take you 12 hours to recharge your car because you're using a suboptimized process and you haven't got all the facilities that make it a very quick recharge capability. So you take that, the full point that I ran, what you want to get to is that recharge... That recharge thing is the CEO is, " By the way, need a bit more cash. You know that stuff we were going through earlier, it looked great, you made a load of money, you've got your bonus, can I have some more?" " Yeah, no worries, come, come." So you just get that monetary cost sent via email. That is the data oasis. We've been living in data mirages for too long. That end game is the oasis, it doesn't exist.
Juan [00:33:16] To quote you, " Data oasis, get money by email." Jim, this is another T- shirt.
Chris Tabb [00:33:22] Yes, please.
Juan [00:33:23] Open, " Data oasis,"
Chris Tabb [00:33:25] Get money by email.
Juan [00:33:26] "Get money by email."
Chris Tabb [00:33:27] And on the back, " Data mirage. Go to court for three weeks."
Juan [00:33:31] This is beautiful. Okay, we need to get to the T- shirt business with all this stuff.
Tim Gasper [00:33:38] Yes. inaudible.
Chris Tabb [00:33:38] By the way, I do have one of the Honest BS T- shirts, I've got given it earlier, but there was no changing facilities so I saved the surrounding area not having to change.
Juan [00:33:50] I'm now curious, where else are we going to take... Now we're talking about mirages and... Where else is this going?
Tim Gasper [00:33:57] Tell me about the data desert.
Chris Tabb [00:33:58] Data desert, okay.
Juan [00:34:03] inaudible.
Chris Tabb [00:34:03] It's a very, very dry and long walk and it is very hot and unenjoyable.
Juan [00:34:06] Okay. I mean, another topic here, given what you presented earlier today, anything you want to say about data products?
Chris Tabb [00:34:16] Yeah. I think it does lead into this ways of working or changing inaudible.
Juan [00:34:23] What is the data product with respect to the court?
Chris Tabb [00:34:29] We're on the fly today. That's not something I even thought of. Okay, I think in that court analogy is your business case you are putting forward is to go and deliver that data product. So it is a project in that case, and an object needs to provide associated value. Business matters. Otherwise, why are you signing off a project? I think if you use that word project, project could be enhancement to a product or it could be a product on its own. Let's try and unpack the word product because I'm part of it. There's many groups I'm part of that are discussing this, and there is that purist end of the spectrum and there is a pragmatic view as well. If I was to probably go for that... Maybe a purist view, maybe I'm guilty of this myself. And it isn't a product until it's a skew, it has a skew, it's in your pin, your...
Juan [00:35:36] That's the pragmatic? No, no, no purist.
Chris Tabb [00:35:36] I think it's a purist.
Juan [00:35:36] Purist.
Chris Tabb [00:35:37] Purist. That's using product in its true, true sense. When we talk about data products, normally we're talking about that. It could be you're selling that data if you're the Nielsen's or Experian of the world, it is a skew. You go and buy, " I want this data set," or is it the marketplace? inaudible.
Tim Gasper [00:35:59] We like to call this, data products have boundaries. It's a thing, right?
Chris Tabb [00:36:06] Yeah, I like that. It's a thing. Taking it as its truest form of product has a skew. And then a data product, I don't think... When we are talking about it normally, we are talking about is a line item or an invoice. What we're talking about is it is something that provides business value, it's something that solves a problem. MIT have got... I take this slightly to monetization and if anyone's watched a show early, listen to this, you're going to hear me repeat myself. I'm sorry. It maps to... MIT of coined the terms wrap, sell and improve. Sell is very easy and is that skew scenario. That scenario is you're selling your data, you need to invoice all that data, you need to get money for that data. And then improve, it is all about that business process optimization. So you are improving and becoming more efficient. If you can't measure it, you can't improve it. So you need to have that business process that you are measuring against. And then the last one is the wrap, which is not quite as obvious, first of all. Wrap is a symbiotic capability you're giving to maybe a core product that makes it stickier or more user- friendly, or maybe you could charge a bit more for that additional feature on it. Everyone has apps on their phones, they're banking there, and you'll see that they'll probably have maybe an FX conversion there or have this, whatever little calculator for this. They're not make you money from that, but they want to keep you on the app or the platform you're on. So they're trying to make it sticky, but make it sticky in a nice way. That you like using it. Again, in the book we've got this analogy of... There's many sticky products out there. They're sticky for the wrong reasons. They're sticky because they're so hard to move away from or you're locked in with a license agree... The sticky that I refer to this is that... I much prefer the scenario that I'm in a deluxe holiday resort and I don't want to go because it's so great, rather than to have Stockholm syndrome and actually be held captive by my vendor and having to enjoy it. I have no other option of being here. So each of those things could be seen as a product. Anything that is providing a tangible evidenceable value, and that business value as we've defined it now. It doesn't have to be directly attributed to fossil mining. It could be, we call it carbon footprint, zero emissions or...
Tim Gasper [00:39:15] Entities.
Juan [00:39:15] Yeah, did you see Apple's latest commercial around the sustainability and they're like that type of stuff is... We were talking earlier is like, that may be a goal, but I can't say that that's profit or anything MRO. But time goes on and now I can say, if we don't do this, you're going to lose money. But actually, you don't know how much the relationship inaudible.
Chris Tabb [00:39:38] Exactly.
Tim Gasper [00:39:39] Apple's probably one of the best at this. I mean, with privacy as well, they were one of the first companies to be much more on like, " All your information's going to be private."
Juan [00:39:46] They did that through Google.
Tim Gasper [00:39:49] Oh, yeah. How that was purely turning into profit versus a longer term investment? One last topic I think we want to unpack with you is around return on investment. We talked about business value, we talked about data products, wrap and sell it. Is the ROI equation a simple one or is it more a complicated one?
Chris Tabb [00:40:10] It is a complicated one. And I think this is why I've moved away from using the term ROI to describe business value.
Tim Gasper [00:40:17] You don't like that term too much in this context?
Chris Tabb [00:40:20] No. I think, yes, in some... I think maybe not in the larger scale. In some small use cases it is not a bad... But I think it's overused because there's never evidence. So it'll tell you, okay-
Juan [00:40:33] That's a good one.
Chris Tabb [00:40:33] You'll say, okay, and you see it here. If you walk around the stands here, you'll say, it'll make your teams 20 times more efficient. Okay, you're going to get 20 times ROI of the same team. Are we? Where's your evidence? And I don't want to see some white paper you've got some company to go and say-
Juan [00:40:54] Well, the typical thing there is that we save you X amount of hours, how many people you get to salary. But it's all kind of theoretical.
Chris Tabb [00:41:00] I think it's too theoretical. Going back to that, you are then asking the right questions of how does it affect my business going forward? Return on investment, look on its own, it's nothing. It needs to be linked to business strategy, business vision, and that is the top level. Everything needs to hang off that. If it isn't aligned to why are you doing it, why have you built a dashboard for it? Why have you built a pipeline for it? It's now an overhead that no one's looking at. I had another concept, a KPI of KPIs. There's so many KPIs out there, who's actually tracking the KPIs of these KPIs to see how valuable they are or how used they are?
Juan [00:41:45] Tell you, that's the meta meta.
Chris Tabb [00:41:46] Yeah, yeah. I don't know how far we're going to take this in the KPI of KPI, I don't know, I didn't go that far. How are we assessing the usage of this? How are we looking at... On the show earlier, the thing we were talking about is that you need to deprecate stuff. We have stuff and I think the product mentality, products going to life, products expire. But we've all seen our data landscapes. There's a lot of stuff that's just left around for quite a long time. I think that's because we haven't looked and patch it up. We haven't removed that code when it's no longer leading.
Tim Gasper [00:42:35] Sometimes we're missing the metadata picture around that. I sense that a lot of companies, they're afraid to get rid of dashboards or KPIs or transformations because they're like, "We don't know what that's going to affect and we don't really want to... We don't have time to do the work right now. inaudible."
Chris Tabb [00:42:53] If you had a good observability product or monitoring product that understood the usage of it. Usage, usage.
Juan [00:43:00] Then you got to connect it to the people.
Chris Tabb [00:43:02] Yeah, exactly.
Juan [00:43:03] At the end of the day, this is all centered around people. If people are not using this shit, then why are we actually doing it?
Chris Tabb [00:43:07] Exactly. I don't like the word observability too much because it's-
Juan [00:43:12] Honest no-bs take. Please open up.
Chris Tabb [00:43:14] Okay. The reason why is because if I go and speak to five different vendors, they all have a different view what it is. I like the term monitoring and monitoring has a number of levels. If you capture every single level, I think then that is your observability capability. But I think what you have out there is a lot of monitoring products that are monitoring different bits. They're all calling themselves observability.
Juan [00:43:40] So what are those bits that you're monitoring?
Chris Tabb [00:43:41] I think there's six levels. I think that the first level is your application performance monitoring. So your APMs, your Datadogs, New Relic of the world. It may not be relevant to all data platforms depending on what you've got. I think the next level is your SaaS products. So you're going to have your software as a services. It's looking at how they're used, how they're performing. And I think the next level of that is the... No, next one is connectivity. I think that's basically the movement of data between each of those locations. So does it fail in point? Otherwise, it's content. So content is what's being moved from one level, so you're not losing roles. No, losing rows or losing data. And the next one is, I think, quality. And quality comes in two types. You've got quality ingest and you've got quality over time, degradation. Over time data drifts, it's getting worse over time. It was all right to have the odd null now and then, now I'm getting 90% nulls. The algorithm were trained before, they only had 10% nulls, is now worthless because it's going to be brought onto and actually costing money to make any incorrect decisions.
Tim Gasper [00:45:05] Inequality injustice, quality as part of the transmission process?
Chris Tabb [00:45:09] Yeah. Some of these you can check when they come in and for example, ranges or not. But over time you can look at deviations in distinct values between things. You only have 3000 customers but then suddenly it goes up to 9, 000 inaudible.
Juan [00:45:25] Okay. So let's connect all of this, because I'm actually surprised we got into this observability.
Chris Tabb [00:45:29] I don't know.
Juan [00:45:30] Which is interesting. Connect it back to business value. So we'd be observing everything?
Chris Tabb [00:45:35] No, no. I think there's the right path of what you need to observe at different points that will enable you to make the right decisions. I think what's relevant is the last level I mentioned, which is the usage. Who, why and how they're using it? So you go back to that KPI of KPIs little side where we went on to, is if you're knowing who's using it, when they're using it, you know whether you can start deprecating it. And then if you actually go to people that are using it, would this be better? Let me deprecate these things because I've got something better you may not have known existed. If you had a good catalog out there that would be able to show the knowledge, how it's connected. Maybe say, " Oh." You're getting this data from the same place. There's also another dashboard that has all these other metrics on it you didn't even know about. It's, " Hell yeah, I want that one. That one's so good to me now." I've seen scoping up and sizing migrations and it costs a lot. If you've got 4, 000 reports, isn't it better to go and have some monitoring, observability capabilities that are looking at the usage part to remove how much stuff you're going to migrate? Because the cost in putting in that monitoring of that metadata is going to be a lot less than migrating figures and reports you've never used.
Juan [00:46:57] We've gone through so much.
Chris Tabb [00:46:58] All right.
Juan [00:46:59] And you know what? I finished my beer. Tim, you're really behind.
Tim Gasper [00:47:04] I'm behind but I'm taking great notes.
Juan [00:47:05] We're taking notes.
Tim Gasper [00:47:06] I promise.
Juan [00:47:06] He's taking more notes than me. All right.
Chris Tabb [00:47:08] I've got one hour to another two hour inaudible.
Juan [00:47:10] All right, let's head to our AI minutes. So you got one minute to rant about AI. Go.
Chris Tabb [00:47:20] Hallucinations are great.
Juan [00:47:24] Expand.
Chris Tabb [00:47:26] Well-
Juan [00:47:26] Thank you. You still got inaudible.
Chris Tabb [00:47:28] The reason why they're good is because it gives us storytelling. It does fill in the gaps. So knowing when to use hallucinations in AI and when not to. You don't want to hallucinate about things that are on mission- critical. But you may want to use it to actually create a story that then you can go and use in your court of law, that then can go and get you the sign off you need.
Juan [00:47:49] Beautiful way of closing us out. Connected it back to the court of law. Actually, I agree with you very much on hallucinations. You went about ideation, creativity, is this perfect for it? And my rule of thumb is if it sounds like a fact, then you got to go check it out.
Chris Tabb [00:48:03] Exactly.
Juan [00:48:03] That's right. That's it. All right, lightning round questions.
Chris Tabb [00:48:07] Oh, wow.
Juan [00:48:07] Tim. All right, quick questions. We'll see. Number one, do data people sufficiently know what data value means for the organization?
Chris Tabb [00:48:15] No.
Tim Gasper [00:48:16] All right. Agreed. All right, number two. You mentioned that the data and chief data officer might lead us down a certain path, maybe a technical path, should it be different? And secondary question, are you a fan of CDAO?
Chris Tabb [00:48:31] Yes, I'm a fan of CDAO. One can't operate without the other, the more control we put over an individual. And if it means that it's going to prevent another role perforation, we've got both them in an organization which means it's less likely you're going to achieve the outcome desire or desired outcome. No, it's probably good to combine them. I think that leads back to the product thing quite well, is because data and the analytics part is the product.
Tim Gasper [00:48:57] How about a chief data product officer?
Chris Tabb [00:48:59] That's not bad. I like that. I like that as well. CDPO. If it does sound like it's Star Wars.
Tim Gasper [00:49:06] C- 3PO.
Juan [00:49:11] All right, next question. Is the data team going to play a key role to drive business value around AI?
Chris Tabb [00:49:14] A hundred percent.
Juan [00:49:14] Around AI.
Chris Tabb [00:49:14] Around AI? Yeah, a hundred percent. I think one good thing that's come out of this generative AI boom, hype or whatever you want to call it, is it is rubbish, shit, without good data. And the only way of getting good data is good data management capabilities, good data quality capabilities, good capital capabilities, good data governance in general. Without any of those things, you're not going to be able to have a successful AI implementation.
Juan [00:49:48] Tim, you got a final one?
Tim Gasper [00:49:49] Final question. Can every company benefit from a data product approach?
Chris Tabb [00:49:56] I think so. I think so. I think it's an educational piece. I like to use the term, the after the possible. And it goes back to, I talked earlier, is that the businesses don't know what data teams do. The data team's role is to demonstrate what the after the possible is for them to say that's the value. Don't ask what your data can do for you. Ask what you can do with your data.
Tim Gasper [00:50:18] Don't ask what your data can do for you.
Chris Tabb [00:50:23] Yeah. Don't ask what you can do with your data, ask what your data can do for.
Tim Gasper [00:50:28] Yeah, there you go.
Juan [00:50:30] Yeah, there you go. All right Tim, takeaway times. Take us away with the takeaways.
Tim Gasper [00:50:37] All right, takeaways. Hey, there's a laptop over here. Hey, everybody.
Juan [00:50:39] I got my iPad over here today.
Tim Gasper [00:50:41] So great conversation. Key takeaway. Started off with honest no- bs, what is business value? And you said that you've been iterating on this a lot. As an aside, I'm curious about some of the previous iterations, how you decided that something wasn't quite right. But you've got some very specific words here. Evidenceable, positive impact on your company performance. And it's not always going to be profit, was a key thing you mentioned, right? It depends on kind of what mode you're in. A couple of different modes that you mentioned, including a hybrid kind of option. First option of a mode that your business might be in was fiscal goal mode, trying to save some money on the budget. This is a situation where maybe quick wins are more important. Maybe cost savings is a little bit more of a focus. Maybe it's just really A plus B equals C, kind of quick math. You focused on, I think, a very useful definition. You said friction. Friction is the time complexity and effort to do something. And when you have friction, you're going to have dropoffs in whatever funnel for whatever business process that you might be looking at. And these dropoffs are things that if you can address them, if you can optimize them, that's stuff that can affect this fiscal year, which is kind of I think how you were thinking about the fiscal mode. You've said this really then links to business processes. You can't measure friction if you don't have a clear understanding of your business processes, and we all don't do that enough well. You mentioned that, " I've never seen anyone maintain their business process models beyond sort of the initial events that maybe triggered them documenting it." So it's very important for you to do a better job of documenting those critical business workflows. You also talked about metadata being critical because it represents how the business works and you can improve the operational efficiency by focusing on it. And also the criticality of meta metadata, the data about the metadata, which is also important. Then you mentioned the second mode, which is growth mode.
Chris Tabb [00:52:39] We did cover a lot, did we?
Juan [00:52:41] We did.
Tim Gasper [00:52:42] And you mentioned a lot of the vendors that are here at the conference and a lot of them are in growth mode where it's not necessarily about, " Hey, let's get profitable." It's, " Hey, here's a bunch of money, here's a bunch of investment, get competitive differentiation, get market share, get number of customers."
Chris Tabb [00:52:56] There's value in being inaudible.
Tim Gasper [00:52:57] And that's important too. Even building a data capability sometimes can be more of a growth mode consideration because of how much investment it requires. But probably the majority of companies are really in this hybrid mode, which is, hey, you're trying to get these quick wins so that you can invest in that growth and you're trying to take the savings to invest in your growth, keep your stakeholders on the journey, make sure you get those quick wins so you can satisfy the fiscal needs while also progressing the growth needs. And then finally, before I pass it to Juan, you had a great analogy around the legal system. You're doing this legal case, you need evidence to back up your arguments, you're moving your way up through the court system and eventually you get to the final judge, that's the CEO. And you have to get kind of good at this process. Small cases or other precedent or other credibility that you built, that's going to help you to be able to make your arguments better and to move up that system more effectively. Who's the best lawyer? Well maybe it's the CEO, maybe that's actually the person who should be really driving this. And the data team is not the paralegals. They're deeply involved.
Chris Tabb [00:54:03] I think CDO is the lawyer.
Tim Gasper [00:54:05] The CDO is your lawyer. Juan, what about you?
Juan [00:54:08] Well, first of all, the business doesn't understand what the data team is doing. This is a big problem, because if you ask HR, what do you do? Sales, what do you... Marketing, what do we do? Very clear answer, very clear crisp answer. We don't have that for the data teams. Your answer to that is they are in service to help other groups improve their business performance.
Chris Tabb [00:54:24] Using data inaudible.
Juan [00:54:25] Using data inaudible. But there's all this confusion when you think about the CDO, the CTO, the CIO and all reporting, it's really hard to have a clear structure. You really need to go map this out. It's very key to really understand how that works in your organization. The other issue there is that you realize that we're probably setting up for failure a lot of these CDOs. CDO, you have this massive agenda, but you're not giving them the people and the budget. I mean, they're not going to be successful and then that's why they may leave in 18 months. There's a lack of accountability around this stuff. Now, what role does the data team play for business value? Going back to the analogy of the legal system, the lawyer is the representative of the data team, which is going to be the CDO. The rest of the team are doing all that background work to be able to provide all that strong work that's needed to provide the business value. Also, talking about expert witnesses. You need external folks who need to show their expertise outside inaudible work. Maybe it's even consultants to go do that. But taking this analogy further, you don't always want to be in a court. At the end of the day it's like a journey. So we're switching now to a car EV journey like, well, if you need to get to a destination you need to plan it out, so you can plan out that journey. So a data oasis is just getting money by email and the data mirage is you go to court for three weeks or more around that. We jumped into data products and it's the case that you're making in that whole legal analogy, right? You're putting this forward to accomplish. It's a project too. The product itself is a product or an enhancement to an existing product. You can think about it from a purist point of view. It's a product in its true sense, it has a skew, it has boundaries to the thing. But the other way of seeing it, it just solves a problem. It provides that business value. And you're taking this kind of approach from MIT, the sell, improve, wrap. So selling, the definition is that you're selling, making money there. You're improving, so you're taking a business process, you're understanding that you need to be able to make that better. So that means that you need to have a way of measuring where you are today so you can improve that. You know what you're improving. You can measure that improvement. And a wrapping is that you take the product that exists, just kind of make it stick here, make it more user- friendly. Then on ROI, it is complicated about this whole ROI. In small use cases it's probably okay, but the issue with ROI is that it's overused because it's not evidence. And I really, really like that. At the end of the day, this all has to be linked to your business strategy and your business vision. If it doesn't align to the why, then it's an overhead that no one is looking at. It's like this extra tax that people have no fricking idea why they're paying it. Who is tracking the KPIs of the KPIs? We're now in the meta meta space around here. That's why it's critical to understand the usage, at least a data assets and data products. And we wrapped up talking about the usage, about observability. A word that you do not like because if you talk to five different vendors, you get five different definitions. You prefer monitoring. There's six levels we talked about. Application performance monitoring like the Datadogs, the New Relics. SaaS products, connectivity. If the moving is working, the content, what has been moved? The quality and ingest and in overtime, if there's data drift. And the usage around it, right? How did we do?
Chris Tabb [00:57:34] You did well. I mean, I'm surprised, you've written up better than I could have. I might ask your notes so I can check them out later.
Juan [00:57:41] Take the transcript of this and put it in GPT. All right, wrap up because we need more beer and we got more stuff to do here.
Tim Gasper [00:57:50] We got one, yeah.
Juan [00:57:51] All right, quickly, what's your advice about data, about life? Second, who should we invite next? And third, what are the resources you follow?
Chris Tabb [00:57:57] So, my advice, I suppose anyone listening on their Data Grid, don't do a job you don't like. Do it because you're learning something new, you are adding value or you learning a new domain. After that, they're getting more out of you than you're getting out of them. Who's on right next? I'd have to have my data value wing man, as I call him, Matt Housley. Yeah, me and him are working together on this book coined The High Performance Data Paper. Where a lot of this stuff we are unpacking, what business value is and how you're going to achieve it.
Juan [00:58:36] And finally, what resources do you follow? People, LinkedIn inaudible?
Chris Tabb [00:58:40] Obviously, yourself, Juan, and this show. Joe does some great stuff. Obviously, he's got his nerdy rants, the money brought in data chats. There's a lot of... I mean, Ben Rojan is great, a lot of his stuff he puts out there and if you're engineer, you won't find any better stuff on YouTube for that. Yeah, many people. The good thing is don't just focus on a few. The more different views you can take on, the more you can shape them into your own. So yeah, don't just live into one vendor or one individual. Have a selection that maybe give different perspectives on things. Definitely have some business folks in there, because we do always go down into technical weeds.
Tim Gasper [00:59:32] I think that's good advice. Yeah, I think sometimes it's easy to listen to maybe only data, podcast., expose yourself to some other disciplines, business, et cetera. Yeah.
Juan [00:59:41] And with that, I'm done. You're done. We're done.
Chris Tabb [00:59:44] Great show guys. Great show.
Juan [00:59:46] Cheers.
Chris Tabb [00:59:46] Cheers.
Juan [00:59:46] Cheers.
Chris Tabb [00:59:47] Cheers.
Juan [00:59:48] Thank you Chris.
Chris Tabb [00:59:48] Thank you.
Special guests
