Azure with Tableau

So hello everyone to this session of Tableau Like a Pro. This is a fireside chat between Interworks and Tableau, where we going to take a look at Microsoft Azure and Tableau and how they interact with each other. We want to look at how Interworks overcame challenges when implementing projects there. We're going to look at how to how they managed to get agility into this project and yeah, try to find out best practices and yeah, really looking forward for this chat. But a short introduction for me. My name is Andreas Terberde. I'm a senior solution engineer in the EMEA team based in the super rainy Hamburg in northern Germany, but I'm not here alone. I have the pleasure from Interworks, the regional lead of Benelux. I have Paul with me. So hi, Paul. Hello, Andreas, and thank you for the invitation to join you today in the chat. Yes, really, really looking forward to go over the projects that you worked together with. Yeah, just a short introduction for me. I'm working here at Tableau as a solution engineer, so kind of the technical counterparts really help our organizations to really understand the platform, how to deliver the most value out of it. I'm working here since two years. Prior to that, interesting backstory, I've worked for a big Microsoft partner, did a lot of implementations in Microsoft Azure, so I have some some background in in the Microsoft field as well. And yeah, I'm happy to have even more knowledge and more expertise with Interworks with me. So Paul, maybe a few words from you about yourself. Thank you very much, Andreas. Yeah. That's quite a quite a tall bar you're setting there. Yeah. My name is Paul. I'm a consultant with, Interworks. I've been doing that for around eight years now. And I'm based, as you mentioned, in the Netherlands, right in the center of the country. And in this role I work with lots of different types of organisations in different sectors of activity ranging from finance, manufacturing, public sector And before I did that, I was actually in various different commercial roles, product management, sales management, in some places general management as well. But I always had a kind of a nerdy streak in me, and, I loved working with data. I always wanted to know if the stuff we were doing worked. And, I thought it was a bit of a dirty secret until I got into the Tableau world and figured out there's actually quite a few people just like me. So, and Interworks, you know Interworks very well. It's, Tableau's first gold partner. We're a worldwide organization where we've got operations across the US, Europe, Asia and various different practices which is really good for me. I'm a kind of generalist but I can always go back to our specialist practices for help if I want to in areas like, you know, analytics, enablement, data preparation, infrastructure and, strategy. So that's a little bit, my background. Yeah. That's that's quite quite a lot there from Interworks. A lot of expertise, a lot of, yeah, human manpower there. So really, really interesting. Thanks thanks there for the short introduction. Yeah. We I know Interworks, a a lot. Like, we worked together within with a lot of projects here at Tableau and, yeah, and I think there's there's some some super interesting, best practices, yeah, that we that we can talk about. Like, over all the years now in my my career, I always devoted my my career kind of for self-service analytics. This is always what the the the core of it was. And I'm I think it's it's a beautiful way of interacting with data of getting data into the hands of everyone. And for this, like organizations really need to to take like kind of a switch from their old, maybe a bit more IT focused analytics department to a more, let's say let's say broad and more open self-service analytics environment. So for me, it's always interesting how do organizations go with it. Like, you work together with a lot of Maybe you can like, a few words there on on on how they do it and how they usually approach this kind of switch from from, IT to self-service. Yeah I think that's really a key question there Andreas, I couldn't agree more how important that is. I see the same kind of thing I think in all kinds of organisations that I work with. A wide range of people with wide ranges of ability and interest. You find people that are really deep into data, they've learned to do coding, they're doing stuff in Python, they're into data science and stuff like that. And there's a kind of a broad tier of people generally doing some kind of analytics. And they may be doing it themselves, taking stuff out into Excel and working with it. So I think what we see in organisations is there's always a stream of formal reporting and ad hoc reporting and the ad hoc reporting tends to get in trouble when the data factory can't keep up. If you've got a number of analysts that are trying to support a sales team they just get overwhelmed with the ad hoc kind of questions and that often gets in the way of other things and that's one of the big drivers and I think it's also getting quite hard to find people with those kind of skill sets. So to my mind self-service is all about taking something that's really out of the domain of the specialist and expert if you like the nerd and bringing it down into them, making it available to the masses. And I kind of think it's interesting you know I like to ask myself the question why does this actually matter? We talk about doing it but I think that if you think about it it's a process, self-service is a process of really making it your own. I always used to come into this in my old roles as a product manager. I'd spend hours and hours and hours, I'd make fantastic presentations to describe my products, I'd spend hours on it, it'd be the perfect story and I travelled with the first sales guy that I was with there and I find they'd ripped it all to pieces, taken the photographs out and done something themselves with it. I never thought it looked as good but I realised later it was always their way of making the story their own. It was their way of digesting it, being effective in communicating it, effective in understanding it and it's usually the best sales guys that are doing this and I think in organisations too it tends to be the brightest and sharpest people that are really on the cutting edge of doing this. So if you can bring that across to a broader audience I think you're actually reproducing your best and and and raising the game for everybody. And I I just think that's incredibly expensive. Not expensive. Important. Expensive is when you when you buy a lot of stuff and you don't use it. That's that's, I think, the the the most important thing. Like, really, the people get to use it because it's it's so important not to only have tools, but really to be able to use the tools because not everyone, as you said, not everyone is into IT and everyone is the nerd. But like, I don't know a job that doesn't need to answer questions based on data. Like everyone needs to do that. And so I think the, yeah, the importance of just the self-service, which is huge. Do we have, like, specific projects that you're currently working on where we have this this kind of switch from, yeah, as we see, like, a more IT centric, the BI departments of the world into, let's say, like a hybrid where where everyone can then really participate in this in this data culture. Yeah. I think we see this kind of process going on in a in a lot of organizations that we we work with. I think there's an example that comes to my mind actually which is really quite a nice example of a medium sized privately owned organisation in the business of cosmetics and beauty but they're kind of representative of a lot of medium sized organisations and the way we approached that they were finding that their practices like I said before they were struggling to keep up. They wanted to know more and more about what was going on. They were struggling to keep up. And we actually started working with them just to build a number of dashboards and that's quite a common way to start. But as we worked with them, we kind of got to know them better and we kind of realised that the difficulties they were getting were compounded by the fact that the data was difficult to work with. It was difficult to get it performant enough, it was difficult to actually get it into Tableau to work with in the first place and they needed a more robust, more flexible infrastructure for that so we moved in to help them with that and then they were definitely keen on enablement. They wanted to get people trained up but training people up isn't enough. You need to show people what's possible. To my mind it's the high definition television effect I call it. I was always happy with my television until I saw one day visiting a neighbour, I saw his picture was far better than mine. Kind of maybe my picture's not as good as I thought it was. So I think you need this kind of community being developed to inspire, to show what's possible and then to help people to get there and we see that happening. And I think one of the ways you recognise it happening is if you're in an organisation and the sales manager wants to know something and he goes to his analyst and says can you build this for me? You know you're kind of in that older model. If the sales manager comes to you and says, I'm trying to build this. How do I get this done? You know you've really cracked it. Yeah. Yeah. This is the really the the the switch, but the the interesting thing that you say there that it's not only about, like, education and not only about giving the tool because it's it's like a whole set. Self-service analytics is always like this this huge approach of kind of change management in the organization, that we also, for example, address with the Tableau Blueprint just to show everyone like it's it's a broad set of things that you have to work on. And one thing that you that you talked about is is also the the infrastructure part because there this is what I saw in like in my previous role where we I did that and but but now also working together with you on the projects there this switch from then coming from like a more more old, like more IT driven environment that can be maybe like on on premise, which is not bad in per se, but but a lot of times this meant like this iterative numbers or just the numbers until something gets done were quite high, so a lot of months until the new VM was there. So a lot of times I see with the organizations that they not only switch to self serve, but only, also move their their infrastructure to a more agile, more cloud driven way. How do you see this this kind of move? Do you see this as kind of the same that is like a joint approach there? Definitely, yes. We've seen a really big trend to move to a different kind of infrastructure model, bringing in cloud data a lot more. I think what we saw happening really was there was a kind of an explosion of movement into the cloud shortly after Covid hit everybody. Mean that's no surprise to anybody that as soon as all these people had to work remote it got very difficult to support that. I think it really did open a lot of eyes to just how important it is to be able to have that kind of flexibility today. And we've seen as well it's not only the move to self-service that's driving this but people are looking for broader pictures. They're trying join what's going on in my store with what's going on in my factory with what's going on with the customers online to get a more holistic picture. So people are trying to bring in information from different places and I think One of the challenges I think that you have as well when you're working with an organisation is there's just so much choice out there. How do you know what's the right thing to do? And with the cloud infrastructure what you can do very easily is you can set something up just try it and say well does this actually work, does it bring us any value? And you can shut it down. So it's all of these kind of pressures, remote work, to get more data from different places, more options coming in that you need to test out that are driving these movements. I think acceptance has gone up enormously. It's not that long ago I remember talking to clients particularly here in mainland Europe where you have this kind of conservative I don't think we should go into the cloud safety and control and now when you look and you see major government finance institutions saying tell me why you need to do this on prem because our default is now to do it in the cloud. Okay, things have really switched around. There's an acceptance there. I think seeing larger organisations go into the cloud just makes that more acceptable and there's not this kind of anxiety to the level that, that there was. So all of these things I think, are coming together to, to drive it. Yeah, definitely. Like this this this caution that you described a few years ago, like, I was an intern at Microsoft where I I saw for a lot of this caution, like, on the German markets, and, this this was back in the days, but now, like, every big organization also that I work with is pushing for the cloud. Like, the main one is, I think there at least for the German market, the central market, it is Azure, like Microsoft Azure as their as their go to cloud platform because of the smooth integration into their enterprises. So I see a lot of moves there. Like how do you see there for example for the the Tableau server deployments? Working together with a lot of customers that currently doing this move that have like their big on premises setup, but then decide at some point, okay, we want to be more flexible, but we want to be more agile and to push that into into a Microsoft Azure environment. How do you see this kind of Tableau server deployment moves there? Do you also work together with your customers on that kind of projects and how do they currently or how do they usually look like? Yeah, it's actually very the timing to have this conversation with you is really good because I've just been working on a project to do exactly that with a fairly large enterprise account. They're moving from a very closed, sealed down kind of situation to putting a server on Azure in the cloud. And I have to say that whole thing does seem super easy to do really. I mean one of the things that helps it is that the Tableau technology is platform agnostic right. So I could do this, we're talking about Azure today but if I wanted to I could put it on AWS or a GCP architecture, it'll all work fine. And as far as the user is concerned they won't notice any difference, they won't know that it's moved from an on prem to a cloud. The only thing that might be different is the address they use But it's all the same and you've got lots of options as well as to how you go about doing the installation which makes it easy. The options really range from we shouldn't forget one of the options that you have and it's a very popular option is not to move it to your own architecture at all but to come to Tableau and say listen we just want to rent space on a server and we want you to take care of looking after the machine. And that's really getting more and more popular. Think that's worth saying as an upfront and we do see a huge increase in interest and uptake in the Tableau SaaS offering, the Tableau Online Server. But having said that if you want to go to your own Azure deployment you could do that by going and using a template to have it installed for you. You can choose to set up your own machine and install it yourself. Or we've done scripted installations with some of the larger organisations. We did recently with a major healthcare public authority and we were actually able to do that with scripting it which was really really useful because we had to spin up tens of nodes, I think it was something like about eighty nodes or something on this installation. Was really quite big. And you can imagine that's unmanageable if you've got to do that by hand. So that was done with set up a script, have the first node installed and then roll those out and hook them all together. And the thing I love about doing that kind of thing is whenever I'm working too I tend to try to script things that I'm doing because then if I've done a practice and I have to repeat it I'm taking my own fallibility out of the equation. I'm able to just repeat that process and it'll run. And then once you've got your server installed, migration lots of options there. You could just back up and restore if everything's the same. If I need to change stuff like a domain I can export the site and re import it. And there are data management add on options that will let you actually go through and pick content that you want to move and process it on the way too. So again lots and lots of different options there and It's pretty straightforward. Yeah, I just thought about the thing that you said with the manual or the scripted version. This is really the benefits of having such an environment on Azure. The ARM scripting in Azure is brilliant just to shoot out a few scripts and then basically have the whole Tableau deployment rolled out. So I think a lot of benefits there compared to where you really have to, like, on your on premise where you have your racks and your servers. So way easier to scale it up and down in a really agile way. Yeah. Once you've got it installed as well on Azure, the whole thing works so seamlessly. As a user experience, as consultants we get to work with a lot of different organisations and if I look at Azure installations that we work with, the experience for a user is great. You log on to your machine in the morning, you maybe get a ping on your telephone to confirm that it's really me and then I'm good to go. Can go on to Tableau server without logging on again. I can log on to the data resources. I may be hooking up to Azure Data Lake Storage Gen two or any of the other Synapse products. You know, and that all that all works seamlessly. I'm I'm just right right in there and it's it's secure and and setting it up is, is pretty straightforward. And it's, yeah, we we always we always take a huge invest in in those connectors to to then make it seamless. Like our connector ecosystem is, I think, pretty, pretty huge. Like we connect to a lot of these these data cloud services, of course, Azure Data Lake Gen two, or for example, the Azure Synapse, which I see also for a lot of customers as they are now main cloud data warehouse. So we have a lot of, like, partners there also that we work together with, and also, for example, of course, like a Snowflake or Databricks. There are a lot of options that we can connect to. But one thing that I find quite interesting, like, how does Interex go for kind of an architecture? What are kind of patterns that you try to to set up when you have so much freedom? So what in in the kind of data layer that you want to set up? Because this needs to also address a lot of these points of self-service analytics. So this agility that we that we talked about earlier. So how does does Interworks go go through there? That's a really interesting question. We do have what I would call a kind of a dream team. So if somebody comes to us and says what would you recommend? Tableau is obviously a long standing, longer standing favourite of ours and we've been working with you guys for forever you know it goes back a very long way and we love Tableau because it's just to work with, it's very very powerful but it's also fun to work with and we see people you know they can learn it fairly quickly, they can do some amazing things with it quickly so that's why Tableau is definitely right up there. Then for the data coming into Tableau would you know given a carte blanche we'd typically turn to snowflake. And the reason we do that is that it's just enormously flexible. It hit me the other day when I was thinking about this call with you. I remember a migration that I kind of did, was a reverse migration, this was a step down migration on the data platform from an organisation that had split off from their parent company as a result of an acquisition and they had to move away from a very, very powerful data platform and use something much more affordable for them. And it was a nightmare. Routines that would run on this powerful database platform, we had to sort of take them apart and run them in steps and do bits and pieces of it. It was a horrendous thing and it kind of really brought home to me the difference in power between a humongously expensive data investment and an everyday Ferrari versus the Volkswagen Golf if you like. But today you can take a platform like Snowflake and you can say it doesn't matter if your budget is a credit card and a student loan and you're trying to set up a startup or whether you're a big enterprise and you're going to turn up with a freight train full of money. It's actually the same solution because it's so scalable. Snowflake is a great one and we tend to work with data processing tools like Matillion that lever the power of Snowflake. Let Snowflake do all the heavy lifting and they basically make that easy to control and manage. And then in the other direction, once you've got your dashboards ready, a big, big key thing is help organisations drive adoption. And one way to do that is make sure people can find stuff easily and they can bring stuff that maybe they have other things going on than Tableau that they need to bring together. And we make that possible with a very highly customizable but out of the box portal solution that, you can bring all of this, together with. But that's if you like, that's the tools behind the dream team. Like we said at the beginning, the tools are at best half the story. And what we always try to do is really understand where an organisation is, what's their skill level, what kind of habits have they got, what are they trying to achieve or avoid and really kind of get engaged with them wherever they are on that journey and help them actually get traction. Because otherwise it's so sad to see deployments where all a company does is spend more money from more powerful tool to do the same thing as they've always done. I think it always needs to be this kind of hybrid where you have these old environments, these historically grown environments on on premise that we need to maybe then take some parts. But, yeah, if you have this this carte blanche, like, this this empty landscape where you can basically then do whatever, then, of course, tools like Matillion and Snowflake just a just a great way to easily scale to whatever demands and needs you have and I think this is the great the great solution there for for self-service analytics that you have with for example Microsoft Azure. Yes. But a lot of lot of good insights and best practices there from you, Paul. So it's a really interesting, I think, for a lot of people just to have this as kind of the best practices that you can then now write down, look into it, maybe also set up a small environment into Azure and see as, as you said, how fast you can install it, how fast you can scale it up. And I think, yeah, as you said, Interbrex has a lot of expertise in that part. I think also there for a lot of people super interesting that they come to you, whatever they have as on premise or their fully cloud environment that they have, like, partner that that can look through it and help them really to to see and, yeah, and understand data. Definitely. And a lot of the things we do with clients all start off with very small seeds. We often start with a client just doing one little piece of work. Maybe it's helping them with a proof of concept or they want to get some training done or they're struggling with performance. And very often what we find is as we get to know each other better, we'll uncover other areas that are maybe equally interesting for them to work on. And then it kind of organically, grows into helping them really raise the whole game. And that's when it really gets fun. Which is which is, I think, a very, very healthy, healthy approach. So, Paul, really thank you for your your great input there. I think super, super valuable for, yeah, a lot of organizations, how ever the big they are or how big the projects are, if they're small, if they're big. I think good insight. So thank you there for for the fireside chat for your input. And yeah, I hope this this was valuable for our listeners. And yes, so thank you for that. Have a nice day. And Paul, we definitely will see us around in one or another project with Interworks. Thank you for the invitation, Andreas. Paul, take care. Bye bye. Bye bye.