Jonas Christensen 2:31
Welcome to Leaders of Analytics. Today, we have Jen Stirrup from Data Relish on the show. And Jen, welcome to the show. It's fantastic to have you here.
Jen Stirrup 2:49
Thank you so much for having me along. It's my pleasure to be here.
Jonas Christensen 2:53
The pleasure is all mine, Jen. Now today we're going to be talking about how to power organisations with great BI and I think the key word there is ''great''. Before we get to that, we want to learn a little bit about you and I've already given an intro of you, but it's always good to hear it straight from the person themselves. So could you tell us a little bit about yourself, your career background and what you do?
Jen Stirrup 3:16
Yes, my background and expertise is in Business Intelligence and Artificial Intelligence. I've been around long enough to be doing it before it was really trendy. So I've been around a long time, 25 years this year, so I'm feeling quite old. But I really love what I do. So I've seen lots of life cycles and technology come and go. But I've always, always been interested in data. So even when I was very young, I was quite obsessed by counting things and sorting and organising. So I even had that streak when I was very younger and perhaps should have gone out to play more. Think I just found my area and stuck with that. I'm also a number one Amazon best selling author on organisational change. I am interested in lots of things. I am part of the data and ethics committee as well for AI and data. So all sorts of things happening in my head. So I really enjoy what I do, though.
Jonas Christensen 3:16
And I can say listeners that as I was doing research for this show, Jen is a very interesting person and has a broad remit, may I say. So Jen, what I saw there was a true data unicorn. So there's strategy. There is very technical content that you put out: How to code in RM, Python, etc, etc. There is ethics. There is TV appearances. There is BI. There's AI. There's the stuff in between the BI and AI and that's kind of what we're going to talk a little bit about today, I think, as we go through. So you are one of these people who actually has this both breadth and depth, which is very rare and hard to attain. How do you acquire all that knowledge and store it in your head?
Jen Stirrup 4:56
I think it's partly because I'm enthusiastic about it. I see the impact that data and analytics can have for organisations, particularly businesses. And I think what I find is, even though I've covered a lot in my career, there's always new things to learn. So it still always freshening my skills, I suppose. There's the whole thing about The Seven Habits of Successful people and one of them is sharpening the soul. And that really talks about, you can't just keep doing the same things over and over again. You have to take time out in order to hone your skills. So I make sure I do that as well. And I think I've been very lucky throughout my career to have a great network of people around me. So I've grown up throughout my career with people who I like and admire, and I've been really lucky to find a peer group that I can learn from as well as contribute. So
Jonas Christensen 5:53
It sounds like one of your key strengths is humility. So that's good. I can hear that in what you're saying and you also make them reflect on the fact that a lot of people who have created things in the world that we consider very impactful have often studied very disparate skills. Famous example is Steve Jobs, who studied computers and calligraphy and therefore we have all these wonderful fonts inside our word processing tools today. Otherwise, it would have been the the one and only typewriter font still, potentially. But you know, you're showing that that's possible at all levels. Now, Jen, you hit up Data Relish. Could you tell us a little bit about what that is?
Jen Stirrup 6:36
So we're a small boutique consultancy. We do all sorts of things. We produce content for organisations and thought leadership. We also do executive roundtables as well and we tend to do those privately. And we also help organisations on their data journey. So specifically, at the moment, I'm working with an organisation who have never used any analytics before. It's a greenfield organisation. They're new. They're just getting started up. And the fabulous thing about working with startups is there's so much energy that people have and they bring to that role. And it can only serve to help keep you energised as well. So I think what I really enjoy is seeing the impact data can have in small organisations, as well as large ones.
Jonas Christensen 7:27
Very interesting. Could you elaborate a bit on how does a small organisation like that start using data in a good way? Because one of the first things that come to mind for me is they probably don't really have a lot of data. They don't have to big data, at least the big tables that a huge bank got or similar might have. So how do they get started? What can they do?
Jen Stirrup 7:49
I'd recommend they look at using solutions and systems in the Cloud. So they don't have to look after technology such as virtual machines or running their own databases. The thing is if they have a process, such as a customer relationship management solution, then often they will use something like Salesforce just as an example. They can help you to build a good process as you go along. If you pay attention to their tutorials and their onboarding material. So it means that even if you're a business just with one customer, let's say you just got one, what will happen is you could use that data, record what's happening with that customer, and have that as a process as well. And then that allows you to learn as you go along while starting with good practices by proven technology. What I don't like to see is lots of Excel or Google Sheets around the place, because you're immediately starting off with data silos. The thing is, if you use software in the Cloud, you can join the dots between those technologies usually, using things like Zapier, for example. Something like Power Automate, just to try and automate the way that the solutions talk to one another. So then you start to build up a more unified view of your data right from the start, simply by using quite well known technologies that are out there already.
Jonas Christensen 9:22
We often say that it's never too late to get started. You're telling us it's never too early to get started. So, I liked that. And I think a lot of people could learn from that. I have started my own business once and I know that there are so many priorities that this could easily fall to the wayside. Thank you for explaining that, Jen. Now, let's get to the main topic of today, which is how do we do great BI and what can it do for an organisation? Now, I think one of the good things to do before we get into a topic is always to sort of establish the basics of it. So could you explain to us in your view, what is BI versus Analytics versus AI and machine learning?
Jen Stirrup 10:03
So the way that I see business intelligence, it's like a rearview view of your business. So you know, when you're driving along and you look in the rearview mirror, you see what's just happening or what's just happened usually. Now, that allows you to think about what's happened in the past. And that actually can be really helpful when running a business. And some of that reporting you need to do anyway. Usually, for things like tax returns, for example, just as a starting point. So business intelligence is really a descriptive methodology for describing your business. So how many sales did you make? What was the average amount of sales that you made over a certain time period? These sort of very basic questions. Now I call them basic, but you'd be surprised how many organisations can't answer the very basic items like that. One question I asked customers is: Who is your most valuable customer? And very often, I don't get an answer to that because they don't know what that means to their organisation. So, for some organisations that might mean turn-over. For others, it may be influence and how they encourage other customers to purchase from you as well. But even a simple question like that, organisations should be able to answer it. Business Intelligence will help them to do that, based on previous data and they should be able to at least think about or come up with some different answers. But I often don't see that happening. So analytics is more about taking the data that you've got and answering more complex questions with it. So if you look at it, business intelligence is very good at answering business puzzles, where you've got a very well defined question, a very well defined answer. Tell me the number of sales you made in the last year in this region. That kind of thing. But then you've got a business mystery and that's where people can't formulate the question very well. And they can't formulate the answer, either always. So they may say something like, ''Why did people stop buying Nokia phones and start buying Apple or Samsung?'' or some other brand new mobile phone. So people don't really know the question. Because to answer that question, you have to prove that that's what happened. People started buying one kind of brand of phone rather than another, just as an example. So they can't really formulate the question precisely because they're not sure what it is and then they're not sure what the answer looks like. So why did people move away from Nokia, just as an example? Well, you need to find out more characteristics about those people, but you don't know what those are. So you just have to kind of start somewhere. So I see the difference between business intelligence and analytics, the same as the difference between a puzzle and a mystery. A puzzle: you have a well defined question and answer. The mystery: you don't.
Jonas Christensen 12:59
I like that definition. That's very good. The puzzle and the mystery. And then we have, of course, AI and machine learning, which is further up the maturity curve. We're going to talk about that later, I think, because some of these technologies emerging, which is quite interesting. Now, Jen, what does great BI look like and what does that do for an organisation when you have that?
Jen Stirrup 13:21
I think great business intelligence is all about collaboration, transparency, ownership of the data and also a stewardship process as well. So when I say that business intelligence should be about collaboration, what I mean is that people should be able to share and collaborate the data with the data to use that data in order to help them to do their jobs properly and better for the purposes of helping the organisation. So with bad business intelligence, what I see is people not sharing data. I see people not verifying the data either. So they're not closing the loop and they're not testing it properly. And I also believe that by business intelligence is where you have almost data anarchy and energy to anarchy. Nobody's in control of the data. There's no auditing process, so you don't know who to watch. It's a bit of a free for all, to be honest. A bit of a scrum, to use a rugby term, with very little structure and probably quite a lot of pain. So that's what a bad business intelligence looks like. Everyone is pointing fingers at each other as well. So they can be quite a bad atmosphere. But there's also the case that people can empire their knowledge about the technology, so they build a little empire for themselves. And when the business starts to grow, they need to let go of the empire a little bit because you've got new team members. They don't always wants to do that. So the impact of that, it's not always the technology. It's the people and processes around that technology. I think bad business intelligence can also mean that people are buying technology to solve a process problem or a people problem. And that's not always a good way forward. It's easy to buy something shiny, but it doesn't really solve the problems.
Jonas Christensen 15:12
Okay, so if we break down what you've talked about here, there is data quality: poor data in, poor data out. Poor data and wrong information. In some examples, there is the way we built the information and what it tells us how - It tells us something. And then there's the people and process involved in that. And all those three elements can break down and therefore your output is not as expected. And I have experienced all of the things that you've mentioned there. I often hear, ''The BI we have is not good enough. It must be the tool. We should change tool from Power BI to ClickOff. From Tableau to Power BI'', or whatever it might be, or ''We're not getting what we want. Here's the report that I want. I have designed it myself in a spreadsheet. Can it produce this?'' or ''The data is just not accurate. It can't be true, what's in there''. So, these are common outcomes of where we haven't done the right thing to produce the right BI. So what should be the process for a BI team to lead the organisation through this and co-create with the business, so that you get the right stuff and the trust in what you've built?
Jen Stirrup 16:18
I think it's all about communication and being very business focused. So understanding the vision of the business is really crucial. So every business, large or small, is all about serving a customer. And sometimes what your customer thinks they're getting from you is not actually the same as what you think you're giving that customer. So understanding and putting the customer first ideally should make everyone in the organisation point forward towards the same common goal and the same ambition. But often I see bad business intelligence where people have forgotten that goal. That they have a common goal. They are pointing fingers at each other, or the process or the tools. And I think if you're very customer-focused, everything should follow from that and that's really the best step that any organisation can take. I do see that some organisations get very bogged down in their technology and it's a sidestep to what the organisation should actually be doing. So, they can get very bogged down in the data, for example, because the processes are wrong, and it means the organisation is diverting effort in that area, rather than back to serving the customer. I'll give you an example. I was presenting at a conference and at the end, a gentleman was hanging around. And then he said a question to ask about Excel. And I thought, ''Okay, fine. Go ahead''. And he said, ''How do you know where Excel has gone?'' and I said, ''What do you mean? There's always a question behind the question, right? So this sounds really intriguing. I need to know what's happened here''. So it turns out that his organisation had rented some marketing data, as many organisations do. But the restriction with the licence was it was only to be used by a particular department and you can imagine what happened. Somebody decided it would be really useful for one of their friends in another department to have it. And lo and behold, certainly the Excel workbook is all around the organisation. So the first lawyer's letter that this gentleman's company received was actually, ''Get our data off your website page''. For some reason, it actually ended up in their website. The second lawyer's letter, said, ''Get the data out of your SQL Server databases''. Because some clever clogs had imported the data into a database. It was not useful. The third lawyer's letter said, ''Get that data out of your SQL Server backups''. And the company had data centres, three data centres in the UK, and the backups were held for 30 days. So you can imagine how many backups they had to take down because of this data. Now, the purpose of telling you that is really to explain bad business intelligence can be about people accessing the data in a way that they don't understand, although not supposed to, or perhaps potentially not even to have. But in that example, that scenario really derailed this organisation. Because instead of focusing on the customer, the IT department is running around for three months, before they could get all the data out of the organisation, out of the email, out of the backups, out of SQL Server. And I am absolutely convinced that happens more than people realise. I think they will just call it because for some reason it ended up on the website. So I don't know if that helps to get serve as an illustration of what bad business intelligence can look like.
So the issue here of course, seemingly is that they have done something and breached a contract with a third party, which they shouldn't have. And that's how it's been discovered. But the underlying issue is that they actually don't know where the data is and it's living all over the place and it's living in disparate files. And if that's in house data, then people short the shoulders and go away or whatever. But it's potentially just as problematic from an intelligence point of view and a data security point of view. That could have been someone's customer data or other things. Fantastic example and it made me both smiling and cringe at the same time. So that's a good sign, I think.
Yes. And I'm sure there's people listening to this, curling their toes thinking, ''Uh uh''. Because people copy things they shouldn't all the time. You get so accustomed to an open source type of environment, which is great that people are swiping stuff because they think it's nifty. And then they just take it and then don't credit people, for example, OR they using it and they really shouldn't be. And I think if something's tugging at your internal barometer, your moral compass, then you shouldn't be doing it. And you know that. I think organisations know that.
Jonas Christensen 21:12
Yes and increasingly, we're seeing that if people don't take responsibility for that, and apply their moral compass, and what implicitly, is their responsibility, because it's no one else's responsibility, then these sorts of things will happen. And it's getting worse, because data is getting more pervasive. So we can easily sort of spill it. I call it the ''Oil spill of data''. You can easily have an oil spill with data that's very personal as well, these days.
Jen Stirrup 21:38
Yes, once it's out there, very hard to get back. And good business intelligence will be helping to drive your company forward, and helping you to serve your customers better, not derailing you completely like that. It's quite tough to get right and very easy to get wrong.
Jonas Christensen 21:55
So Jen, we've established here that a lot of organisations struggle to get their BI to serve them in the right way. And that's not because there's a shortage of reports and dashboards floating around. In many cases, there are too many ways to get an answer to the same question. Why are so many organisations lacking good BI and what should they do about it?
Jen Stirrup 22:18
I think they're often using the wrong tools for the wrong jobs. And they're often copying and pasting, rather than having proper automation and a proper automated process to smooth the paths of the data journey, as it goes through the organisation. So, to give you an example, you may have heard of an organisation called Enron. So Enron went bust, I don't know many years ago now. They're one of the largest US organisations and that Enron situation was something that nobody foresaw until it was too late. And it wasn't the fact that they didn't have enough data. The problem was, they had too much data. They were drowning in data. And I'll never forget the pictures and the television of those people leaving the building, carrying big boxes of files and papers. the problem was not enough detail. It was just everywhere. Nobody knew how to analyse it. They're using the wrong tools for the wrong jobs. So say, for example, in that scenario, your business intelligence is consisting of workbooks which are disconnected. People are manually changing them. So you could produce equality. No auditing. What can happen then is the organisation needs to think more mindfully about the right tools and what to do with that. So ideally, they should be automating. Organisations sometimes don't want to automate, because they worry about the impact on the team members. They're worried people might lose their jobs, essentially. The way I tend to find is if you're automate, it allows us as humans to make more advantage of our creative skills by taking away some of the automated, repetitive, quite boring tasks that we as humans are not very good at. And that's where low-code solutions can come in. No-code, sometimes they're called that. And automation tools can help us to move data around, join it together, and then make it into something productive. So I spoke to a gentleman recently, in one organisation. He spent four days out of five a week, working week, four days copying data, copying and pasting data into places. That's not why people go to university and incur student debt. Right? I mean, anybody can do that. But when I spoke to him, he was reluctant to change actually, and I was a bit surprised because I thought that must be terrible, you know. But he didn't want to let go of the task and I think he'd got so ground down by doing it. It felt like he almost couldn't do anything else. And that's a difficult position for somebody to be in. Where they lose their confidence. Because all they're doing is copying and pasting. So sometimes there's a resistance to change in the organisation. And I think people, managers, leaders and employees and team members need to be given opportunities to be led into the light about data. Give them a good example of what good looks like and what that can do for the organisation. You can do that by starting simple. Prototyping, building up, learning from the failures and moving forward.
Jonas Christensen 25:32
So, how do you do that when you're out with your clients? How do you show them what good looks like and what are the elements on a business intelligence tool, piece or output solution that are always there when it's ''good''?
Jen Stirrup 25:48
Yeah. And so I tend to try and understand better. First of all, do the business need silver platter reports or self-service analytics? We hear a lot about self-service analytics. People are going to be given a data set. They can pick the data that they want. They're going to click and tick and they'll be able to get visualisations that are based on good integrates data. The reality is a lot of people don't want self-service analytics. They don't want to click and take anything. What they want is to click on a link and get the data that they need straightaway on a silver platter. So I call it ''Silver platter analytics''. They don't actually want to do any self-service at all. And they want it how they want it on a silver plate and served exactly perfectly. So I try and work out through the businesses: Do they want silver platter analytics or do they want self-service analytics? There's no point giving them self-service analytics tools, if they're just not going to self serve. And then that example, I can try and give them something that's more static in nature. I can give them the vision of self-service analytics. But often people are not interested. They're moving too fast in their business. They don't want to analyse. They just want data, information, and move forward quickly and usually operationally. So I think that's the key thing for many organisations. Where I see successful organisations is the ones that automate the data transfer from the source systems to somewhere else, where it can be used for analytical purposes. Now that can be, here's the buzzword: it can be a data lake or a data warehouse, or wherever. A data lake house is the latest new thing. The thing is that these terminologies tend to come and go. They can also be interpreted and misinterpreted by vendors as well. So we have to cut through the vendor pieces of information. I don't know what to call them. I have a few words, maybe not very publishable. So we have to cut through the vendor speak. Just keep it as that. We have to cut through the vendor's speak in order to try and understand really what the organization's needs. And do that through quite a good structured lens, in order to really deliver the business intelligence, analytics and AI that they're looking for.
Jonas Christensen 28:11
You're making me reflect on many past experiences here with the example of silver platter vs self serve. And one particular example was in a previous organisation, we had 30 reports that all pertain to the same portfolio of products. And for me, that was terrible. Why do we need 30 different versions of the same information, slice and dice for different people? It's mushroomed out of control. No one knows where to get the answer. And I started going around the business saying, ''We're going to collect this into a solution that might have 1,2,3 reports, not 30 and that'll give you everything you need. You go in there and you log in and you can filter and slice and dice''. And the head of sales for this product portfolio said, ''But hang on, I actually like just getting an email in the morning. And I get five different emails about five different areas of this product portfolio and it's a static PDF cut out of the table that I get in a gmail and I might forget to look at it. But when it comes like that, I always look at it first thing in the morning and it's there''. So that really made me really reflect on what is great BI to me, which is a very technical solution that looks whiz-bang and has lots of graphs that can interact and so on. It's not necessarily meeting the needs of the end user who potentially has a much more basic need that fits in with a process that I might not appreciate. Is that something that you experience out there?
Jen Stirrup 29:41
Absolutely. I think we can try and take people on a journey that they don't want to go on and they don't need to go on. I did some work for an insurance company years ago. The boss I was working for at the time said, ''You need to go and speak to this gentleman in such and such department, because what he's got is a piece of paper stuck on his noticeboard at his desk. And we need to give him something visual''. When I saw this table, and it honestly had about seven lines of data and it wasn't very much. It was insurance, and it was things like the number of hurricanes they expected to happen in the world that year. This kind of thing. So I remember asking him, ''How can I improve that?'' and he said, ''You can't''. He says, ''All I want to do is look up from my desk, look at the table, dot down the figures and look back down again''. And give a yes or no answer to the question he was being asked. And that was that. And so what are we thinking? ''Okay, so I could have done something visual''. So went back to them and gave them a Word, a table in Word, and said, ''Here's a table we've done. Here's some weather symbols. I've given you a hurricane symbol. I've get what the other ones were. Just remember the hurricane''. And he said, ''Thank you very much. It's in colour now instead of black and white'' and he was really pleased. There's no business intelligence involved in that. I just gave him a pretty Word document. So I think sometimes we can spend a lot of time overanalyzing because we are analytical people. That's what we do. But we may be doing it in the wrong direction. I gave this man his Word document. He's more than happy. Reps don't use black and white copy. Stuck that up. Job done.
Jonas Christensen 31:18
Very great example and it is always horses for courses, isn't it? It is really hard to move people up the maturity curve whilst respecting where they're at. And that maybe what we call more mature, more advanced solutions, more technically advanced solutions is not necessarily going to solve the job just because it is more technically advanced. You see the tools that we use having more and more features, and then more and more abilities to slice and dice data in a different way. But it's important to remember that we have spent a lot of time on understanding various diagrams, whereas if you're using, like a Sankey chart, for instance, it might be the first time that someone sees that and it can be super overwhelming if you don't know what it's telling you.
Jen Stirrup 31:59
Absolutely. It has to be something really meaningful for the business. So, to answer the last question better, what I do with organisations is I give them - I ask them for a key metric that they haven't implemented yet. What would you really like to know? And that could be anything. It could be, ''Tell me the number of support calls that were, say, priority one within the last week?, for example. Just something really, really - Sounds basic, but not always that easy to get. So I ak them for the key metric, and ask them why they haven't done it yet. And that tends to be because they don't have the data or the data is buried somewhere. They don't know how to get it. Or sometimes they just don't have access to that data because a vendor wouldn't let them write their own reports. They have to use what they've been given. So with that one key metric, I take that as my challenge I try to map out everything to try and get just that one metric on a dashboard. It doesn't have to be fancy. We're back to our gentleman with our seven line table with a picture of a hurricane on it. Just if I can show the data, then that tells me I can do end-to-end in that organisation. I can go from the source to something that the executive can look at. And I can do all of the steps in between that and more. So I think that one thing for organisations to take away is that you can do that. But start really simple. Start with something meaningful for the business and make it actionable as well. Make it happen. And then people align themselves with success. So when they see that number in the dashboards, suddenly everyone's responsible for the success of that dashboard, and how great it was. People will naturally gravitate towards success. So if you can even show the smallest success, start there, and then show another small success. And a small series of sustainable small successes is a better way of persuading an organisation to get behind analytics and business intelligence. If you try and do something too big, and it fails, or is perceived to fail, then people will become very averse to getting on board with it. And again, back to people and not technology.
Jonas Christensen 34:16
It is a lot about people. And the thing that I often reflect on is that the data has a lot of psychology in it. So, the way you showcase a data point may mean that it gets interpreted in a different way, both consciously but also subconsciously. And if you sit back and think about how various people are using data to sway our opinions in all sorts of areas, you can see that that is actually psychology at work. And we should think about that when we do design our business dashboards, even though it might not sound like that's a place for psychology. There's a big difference between a dashboard that has a map with all the dots on it, where you've sold your products and that probably is kind of useless. But it looks pretty. Versus a, say a leaderboard where you have all your salespeople running up against each other, who are all competitive and wanting to be the best salesperson in the company and having those guys run against each other could be another way to do it. But seeing the underlying data, very different way of influencing how the organisation takes that data and uses it for forward progression, I suppose, as we call it. So, Jen, data is eating the world. And every industry is impacted by that. And we're also now seeing that these, what I call, you could call them traditional BI tools like Tableau and Power BI, are becoming increasingly AI driven, which again, transforms what these tools can actually do for business. And we've just talked about some users needing some simple solutions. But at the same time, we're putting all these superpowers into these tools. What are the trends we're seeing in this space and what should we expect BI tools to do in the next one to three years?
Jen Stirrup 36:06
So, I think before we answer that, we should perhaps examine what AI actually is. So the definition of artificial intelligence is that it's computing inspired by human thinking. So when we look at machine learning, it's not really inspired by humans. It's more statistical and it's more about navigating a search space to come up with a result rather than being specifically inspired by humans, and the way that we use language, or we make inferences, for example. So I do think AI will become much more prevalent in many systems, including business intelligence. If you can imagine, you've got a dashboard and the AI, the heuristics behind the scenes, presenting the data to you that it thinks you need to see. I see AI sometimes being used to do things like create a multitude of reports based on the data. I don't always think those are very successful when I've seen them in play. I think you will see more AI. I think there'll be more natural language processing. So you can ask for a report and you see your data. So a bit like the minority reports with Tom Cruise appears in front of the glass screens with his gloves on and he's swishing results around. And the reports appear. He gets access to his data. But I think that that is in bubbles for the users. The thing that I'd really like systems to tackle and use AI in machine learning as well, specifically would be in the areas of data cleansing and data integration, because these are the parts of their business intelligence system that are not really very trendy. And people don't question the data enough, actually. So, they trust but they don't verify. And I think if we get more intelligence around the way that systems need to talk to one another, that would give us so many more benefits about enriching the data.
Jonas Christensen 38:08
Hi, there, dear listener. I just want to quickly let you know that I have recently published a book with six other authors, called ''Demystifying AI For The Enterprise: A Playbook For Digital Transformation''. If you'd like to learn more about the book, then head over to www.leadersofanalytics.com/ai. Now back to the show.
So we're definitely seeing some trends here of some really interesting features being introduced to the end user, which is where I suppose a lot of the interest is going, but just as importantly, is all the stuff that happens in the back-end before the data enters into the dashboard. And potentially even more important than one click solution to visualise data in a different way. Now, these trends are giving business intelligence tools and the users that use them, some superpowers that also helps to, we call it, democratise AI, democratise data, and the techniques. There users that are typically no-code business users. So, these new users will be able to generate outputs that they would not have previously been able to get to and they would have relied on analytics teams to produce those and also to validate the accuracy of the data, etc. So what are the benefits and pitfalls of that sort of evolution and what does that require of data professionals?
Jen Stirrup 39:33
So, when we look at things like a data democracy, there's different ways that we can look at that. You could see a data democracy is released, where we have all the data audited and the data is transparent. We know what it means. There's the data dictionary. We have both silver platter and self-service analytics and we're in a good position because everyone in our organisation can access the data that they need. The thing that makes me a bit uncomfortable about the phrase ''democracy'' is democracy makes us think that everybody has equal access and equal power to control and change the data. However, when you go into the area of data privacy, it feels very much that is not a democracy. The balance of power is actually in favour of the company, rather than the individual. So in the UK, we have the ICO ( Information Commissioner's Office) and what they do is they hand out fines, for example, to organisations that are not behaving well with people's data, and you still see so many big name organisations not treating data and people's data as if they are adults, and not treating the data in the way that people deserve. So data democracy makes it sound like every business person is in power and access to this in detail. But we have to remember that the data that's been held by these organisations is perhaps not always been ethically obtained. And that's where the democracy part starts to fall down. It's a bit of an anarchy. Companies apologise when they get caught doing something. But afterwards, they just - It's not enough sanctions to really stop them from doing that. I still sure many people don't really understand GDPR, for example. One gentleman said to me that he wasn't worried about GDPR, because that's what his cybersecurity insurance was for. And I thought, ''No know, your policy won't cover you for GDPR breaches''. And he said, ''Well, everybody's doing all sorts of things with data''. And if you look up the ICO, you can see lots convictions by companies that should really know better.
Jonas Christensen 41:47
Yeah. Now, I would also say: Don't rely on insurance to fix your problems. Insurance is a survival mechanism, when stuff really hit the fan. But you're not going to leave the gas on and light a match in your house, just because you have a good insurance policy. Don't think like that when it comes to data either.
Jen Stirrup 42:03
You're right. It was a sales meeting and then when he started talking like this, I couldn't wait to leave the room. So they said to me, ''Oh, yeah. Are you ready? Would you like to work with us? You know, we're doing on script projects''. And I thought, ''I've just sat in a hour with your business leaders telling me that your GDPR breaches are covered by your cybersecurity''. So I just went back and said to them that I was busy for the next fraternity and I wished them all the best with everything. Because I thought If they are already looking to blame other people, I don't want to be involved.
Jonas Christensen 42:33
Yeah, I can imagine as a boutique consultancy, there is definitely some 80/20 principles that need to be applied as to what clients should take on in terms of who's worth the effort. So Jen, giving the whole organisation access to data and tools doesn't make everyone a great analyst. And that's actually one of these risks of democratisation. And what you've described is that sometimes it's not just a democracy, it's an anarchy. Is it possible to democratise data and potentially AI now in these tools, businesses, and what does that take?
Jen Stirrup 43:08
So, in terms of will it be possible to have a democracy, I think it comes back to how do we determine democracy. We know that some people should have the power and control over the data and that's freely handed to them. The only thing I'd be concerned about is what that means to the external stakeholders of that organisation for whom the company actually holds data. So that could be data about the customers or their children. And if you look at the Facebook example, that was more of an anarchy, but it was like a democracy, because Facebook had the data and allows people to access it. And once they lost sight of that detail, they didn't know what was happening to it. There was no oversight. So I'm not sure we'll ever get a totally clean and pure democracy because I think the auditability between companies and what they're doing and passing data back and forth, I think that's going to be really hard to track. And I don't think there's probably much appetite to do either.
Jonas Christensen 44:07
Now in terms of just using data and being able to interpret it correctly. That's a skill that is just probably not for everyone. So there's a risk there, too. I compared it a lot to where we are now with data analytics, data science. I compare that to computing and maybe 30-40 years ago, so late 80s, early 90s, where we're moving away from typewriters, and everyone's getting a PC in their office instead. And we started to use Word Processing and all that. And at that time, you could still have some people who were using a typewriter and some would move to computers and then have to learn but we still have people who are not computer whizzes, and they should not be responsible for the most events, tasks on the computer. Similarly in data, we have a similar cohort, I think that won't be able to do it, but they can do other things. So this complete democracy, everyone should be using dashboards and be just as good as the next person is unrealistic, I think.
Jen Stirrup 45:06
I think that's fair. So one of the best data scientists I ever work with couldn't actually code at all actually. She could use Excel, but I worked with her in one project and what her skill was the ability to ask really great questions. That was the first skill. A second skill was to know when to stop asking questions, when we were going down a rabbit hole. And she now knows how to code and Python actually. She went to learn in order to try and understand developers better. But I think that's what I learned from her. It was just that asking questions is such a unique skill and quite a rare skill. People are not always very good at asking the really insightful questions and she was fantastic at it. She really kept us all as a team and a great direction. And she will also know when to stop asking questions, when do we need to come back out of that rabbit hole and then think about something else. And I think you can code as much as you like. Really lots of people can code. But asking yourself, the bigger questions about why you're doing these things is a rare skill in some ways, I think. And I think to be a really good analyst, you need two things: When to start asking really fantastic questions and secondly to know when to stop and to move on to another topic and start to examine that in the same detail.
Jonas Christensen 46:27
There's a term that often get used, which is ''Decision scientists''. I like to compare data scientist and decision scientists. And I think decision scientists have much more of a career runway because they are, in my opinion, more likely to have a big influence on their organisation. Because if we cut the word scientist out of it, we have people who are using data vs people who are actually helping to facilitate decisions with data. And I think people should reflect on how they become decision scientists, as well as data scientists in their careers. They're going down that career path. Now, Jen, what should people do, who are not traditional data professionals, to keep up with all these trends and expectations that inevitably will fall on to them?
Jen Stirrup 47:14
So, I think my advice would be to read a lot. There's loads of great books out there around business intelligence, and analytics. And you don't even have to read and know how to read a book about code, for example. I'm not expecting that. But I think there's some great books that you could look at. One I wrote a review for it or reviews for it and Jordan Goldmeier is the name of the gentleman who wrote the book. He co-authored with someone else. And it's about how to become a data head. I'm sorry, I don't have the link at hand. But what I liked in that book was that they were really trying to look at it from a business angle and then diving down to the details. You mentioned decision science a moment ago. I remember that was a big turnaround in the 80s. And it seemed to go out of favour and now it's coming back against. So that's one of the trends we should have maybe discussed earlier. But what I like about that approach is it's coming back to recognising that a good business intelligence and data science programme must include the business front and foremost of what they're doing. It's not enough just to say, '' I'm gonna run this algorithm. It's gonna save me money. It's going to optimise productivity''. We have to also ask ourselves, ''What harm might this do to our reputation or to our customers?''. Such as data privacy: Is there a potential for a data leak? Or is the algorithm actually helping and benefiting the customer in some way as well? So I think that's an important thing that sometimes gets missed when we think about data science. So I'm hoping that decision scientist's aspects will be dialled up again and not go away this time as it did before, I think it went away because it was complex and hard. Maybe people start to accept more easily the data is hard.
Jonas Christensen 48:59
It is really hard. And I think it's actually one of the hardest professions to work in at the moment, because you are in a world that's very technical but you have to translate that into layman's terms to a group of stakeholders who typically aren't as proficient as you in that. And the higher you go on the organisation, the less experience they have, because they are typically and I'll say typically, from a generation that didn't have that when they were doing the doing at the lower levels. So they just don't have the applied experience that you might have in wrangling data and using data day-to-day. They see lots of data output. But the the understanding and the vision for it is something that needs to be created out of those data teams, but also we need to lead the organisation along on a journey to actually build this out in the organization. It's a very big technical challenge and a very big leadership and communication challenge at the same time. Those are skills that are hard to house in one person typically. So we do have a big job ahead of us. And Jen, one of my last questions for you, because we are closing in on time here, is we have all these clever people working in the data space and there are so many tools coming at them left, right and centre. They're getting more and more advanced and we need to learn more and more to keep up with this space. Which technical skills should data professionals develop to stay relevant in the 2020's?
Jen Stirrup 50:25
Still believe that statistics is an important tool. I saw a tweet yesterday, where somebody said something along the lines of ''You no longer need to know stats in order to be a data scientist''. I think if you want to make decisions based on data, statistics is absolutely crucial. Even this basic descriptive statistics will get you quite far Sometimes. One organisation I worked for, they found that they were calculating the median incorrectly in Excel. So you know, the median: You take the data, ascending order, you pick the middle number. What they were doing was they had the data in a column with some case study numbers, and column A and the data in column B. And instead of running the median command or something like that in Excel, what they did was they picked the middle number of that column. And anytime you switched the order of column A, that changed the values in column B. So the median swung wildly back and forth. So ask them why they've done that, because I thought I cannot find any reason why anybody would want to do that. And then eventually, they said, ''We didn't know how to calculate the median''. And I said, ''Okay. So why are you going to the median in this case, because maybe the mean might be better?'' because they didn't have any significant outliers. It was quite an even dataset. And they said, ''We didn't know how to calculate the mean. So choose the median?''. So then I said, ''Well, how long has that been going on for?'' and they said they didn't know how to calculate the median for two years, properly. And for two years before that, they couldn't calculate the mean and I stop them and said, ''So for four years, you've not been able to calculate the mean, or the median in this two column table of data. Right?''. And I did my best consultants piece. I really did. I was all sympathy. So then, I said to them, ''You know, you didn't have to do that. It's actually really simple to do''. I showed them. And then I said - I mean, sometimes I should just honestly just shut my face. Right? I didn't. I asked the next question, which was, ''Is this happening in this spreadsheet only or is it throughout the organisation?'' and the minute the words came out my mouth, I knew the answer, right? I just should have stopped there. And they looked at me and the leader actually hit himself, hit himself really hard on the head and he said, ''It's everywhere. We haven't been able to do this''. So it's not that I blame Excel. I blame the fact that they weren't able to develop or test it properly. And we're making business decisions based on the middle column and however, it was sorted at that time. So statistics will get you really far. Because you can be that person in the room that says, ''Hang on a minute. Have you double-checked?'' and the number of people that don't double check is really surprising to me. Now, we don't always double-check absolutely everything. Sometimes you don't have time. So I tend to do the work, and then give it to someone else to test a bit blind, for example. So I tend to give people with that, and then give them the opportunities to tell me if anything's wrong, because I prefer having more eyes on something than having no eyes on it at all. So never forgot that example. But I think that's why I always say, as boring as it sounds, knowing, at least basic statistics will get you quite far. If you can calculate the mean and the median, you're doing much better than that organisation. So you've already got a head start, right.
Jonas Christensen 54:04
So I take away from this here: Go back to basics. It's not about Python vs R or whatever. It's about knowing the basic foundations of your tool or your trade and the tools are just what you use to facilitate that. Thank you for that, Jen. Now, Jen, I have three short questions left for you. The first one is I've noticed you publish a lot of book reviews on your website. So you must be a big reader. Do you have any book recommendations for us and why would you recommend the book that you might recommend in a minute, or books?
Jen Stirrup 54:41
So I would recommend Gordon Tredgold. I read everything he writes. Gordon has got a book called ''Fast'' and it's about his methodology for success and business. It's also data savvy. He's got some amazing examples in there. I won't spoil it for you. It's a book that's quite short. So even if you don't read much, it's a short book. There's no excuse. It's also very insightful. And it's also quite funny some of the examples, because when you read it, some of the examples, you think, ''My toes are curling. I'm cringing. I know exactly what he means''. So I'd recommend the book called ''Fast'' by Gordon Tredgold. And that's why I recommend it. I think it's very relatable. And he's got such a great way of explaining things. I've had the pleasure of chatting with him now and again. He's just got an incredible way of explaining things. He's one of the Forbes top 100 leaders or something like that. Incredible guy.
Jonas Christensen 55:37
Great. I have definitely seen that book cover somewhere. I can imagine that word and book cover, but I haven't read the books. I will be picking that up very shortly. Thank you for that recommendation, Jen. Now, one question I always ask the guests on the show is to pay it forward. So who would you like to see as the next guest on Leaders of Analytics and why?
Jen Stirrup 55:58
I'd like to see Donald Farmer. He is a true leader. I think that I've been very lucky to - Anytime I've spoken with Donald, I have learned something and learned a lot usually. He's got a great way of communicating. He's one of the people that inspires me to read a lot, actually. Because I'd like to be Donald when I grow up eventually as well. 've been very pleased to have him as a mentor and a friend. I don't know if he knows that he was a mentor to me, but I look up to him enormously. So, yeah, definitely Donald farmer. I didn't see him on your list of podcast guests, but he would be my first choice.
Jonas Christensen 56:34
Very good. I was afraid that when he said Donald, another last name was going to come up but Donald Farmer sounds good. I will definitely be looking at Donald and trying to get him on the show. So thank you for that recommendation. That's brilliant. Now lastly, Jen, where can people learn more about you and get a hold of your content?
Jen Stirrup 56:52
So I blog technical material over at jenstirrup.com and attend to do more business related material over at datarelish.com. I'm also on LinkedIn. So if you look up Jen Stirrup, I'm on there. I'm very accessible in Twitter as well. I probably spent too much time on it. So if people want to get in touch, plenty of ways to do that. I also have a calendar facility over at datarelish.com If people want to take some time to vote out. Thank you so much for having me along today. It's been great fun to chat with you join us.
Jonas Christensen 57:26
It's been brilliant. I have appreciated this conversation so much. Jen, all the best for the future for you and for Data Relish and we look forward to seeing more from you in the future, content wise as well. And have a nice day.
Jen Stirrup 57:39
Yeah. Thank you so much.