Drill to Detail - Drill to Detail Ep.56 'Essbase Cloud, Analytics Cloud and Cloud Analytics Consulting' with Special Guest Matt Yorke

Episode Date: June 4, 2018

In this specially-extended episode just before ODTUG KScope'18, Mark Rittman is joined by Matt Yorke from Qubix to talk about Oracle Essbase Cloud, Oracle Analytics Cloud and the business of Oracle Cloud analytics consultingEssbase Cloud is Here!Oracle Analytics Partner Forum 2018Oracle Autonomous Analytics CloudODTUG KScope’18Matt Yorke on LinkedIn

So hello and welcome to another episode of Drill to Detail, the podcast series about the world of analytics, big data and cloud data management. And I'm your host, Mark Rittman. So my guest in this episode is someone I met for the first time when I attended an Oracle partner event but for various reasons wanted to keep a low profile so uncharacteristically for me I was sat at the back and wasn't asking or indeed answering any of the questions as I normally do. However there was someone else at the front of the room also with an English accent who was trolling the Oracle folks almost as well as I do and I thought I have to have this man on the podcast at some point and I chatted to him in the event in the evening and you've got a long I suppose background in consulting and delivery and particularly with working now with some of the new Oracle Cloud

products so I thought it'd be great to have you on the show. So Matt York from Cubix welcome to the show and it's great to have you here and joining us. Hey hello mark thank you very much for for asking me to join you on this podcast um um looking forward to chatting with you actually um one thing i probably need to call you out when actually you said almost as good um i actually think if you if you'd been sitting at the front as as well i'm i'm i'm not sure the poor oracle guys would have got a word in right if both of us had been uh had been uh on it so so yeah probably just as well you were keeping a um a low profile but it was good to hear somebody else asking the questions really so it was interesting hearing your your take on things and the fact that there was someone out there who was who was so i suppose so heavily involved in some of the new products

that they've got yeah i mean it's to be honest i was i'm always surprised at these events when people um you know tend to to sort of sit sit quietly and expect everything to be served up. I've always figured it's important to go to basically any event, any presentation, whichever side of the desk you're sitting on, whether you're presenting or whether you're receiving the presentations go with and go with an agenda you know go with the agenda to learn something to you know to drill into the detail ironically on on the podcast here um you know to share something to teach somebody something to inspire somebody you know but but you know you don't ever just want to be a sort of flag flapping in the wind there you always want to to you know drive drive something forward whatever your agenda is so yeah, I'm getting a bit of a reputation for being relatively vocal at these events and hence becoming invited a little bit more

to actually do the presentations rather than ask. Excellent, excellent. So Matt, tell us who you are again. So tell us what your background is and how you got into the consulting world and the route you had into the place you work now, which is Cubix yeah okay um certainly i um well my name is matt york as you as you said i'm one of the group directors of cubics international um we are um are called platinum partner um

specializing in you know what i'll in simple terms describe as both the EPM and the sort of business analytics space, that does straddle two, you know, sort of, or at least two arms of Oracle, both the sort of apps and tech side, you know, which from an Oracle sales perspective is quite different. But, you know, but equally that puts us in quite a unique position because it means uh you know if uh you know if if there's a sniff of of uh you know of an app sale usually the customer wants some tech i if the customer's got epm or is going to buy epm they need some business analytics you know to go on it as well and vice versa so that's kind of a position we've we've sort of leveraged a bit in terms of how i got into cubics um well i i did a marine engineering degree um in lost count now uh mid 90s um and marine engineering degree down at plymouth and

one of the things i kind of realized is is um i was quite good at the computer stuff you know i was good at all of the engineering stuff but i was quite good at the computer stuff. You know, I was good at all of the engineering stuff, but I was quite good at the computer stuff as well. So, but actually what I really wanted to be was a dealer in the city. I really wanted to be a trader for one of the investment banks in the city. So I kind of got off my bum a little bit and I wrote a letter to every single investment bank in the city, you know,

and this is very early days, word two, you know, mail merge. You you know that was quite a lot of effort um and and whilst i was um kind of you know trolling the banks uh there's you know there's a pattern to my behavior here um i um i firstly took a job as a sort of bulldozer spare part salesman and started writing actually a um a sales database for them you know very early generation windows 95 um you know very expensive very low powered laptop and i taught myself microsoft access um you know to write a database that the gut that the sales guys could go around the world with you know to look up stock look up customer orders um you know and basically use it to help them

leverage the sales and i realized i had a bit of flair a bit of aptitude for this um eventually after getting enough rejection letters from from uh from the banks saying i had the wrong wrong degree to be a to be a trader i kind of thought okay i can i can do the the you know the uh you know building systems and solutions for customers so I switched my focus and joined the predecessor of Cubix, a company called Julian Owens Associates. So that was 1st of April, 1997. And then from there, fairly quickly, actually then ended up working with building an S-based system.

What version was it s-based four with a big vb5 and then latterly six um front end for ing bearings obviously uh you know that was the uh you know the organization that rose out of the the ashes of bearings bank and and nick leeson um and more or less haven't really looked back from then so that was uh when was that that was sort of 1998 with with those guys there in those early days mostly focused on the you know the s base end um and then of course starting to starting to look at where i saw the weak points were in projects and one of the things i fairly quickly realized the weak point in projects wasn't actually the s base end i realized it was the relational database end where people were spinning up business objects as, you know, as it was s base whether it was business objects the weak spot was always the database modeling the data to make efficient use so i kind of that

was where my my focus then changed a little bit then to to add in the relational side and of course then the logical next step beyond that was when oracle bought ob, or sorry, when Oracle bought Siebel, you know, the logical next step was suddenly in that portfolio, Oracle had the best database, they had the best OLAP product, and now they had a really strong, you know, BI tool as well. And so that was kind of the, you know, the logical flow and sort of the logical, you know progression there so my role these days is i'm um you know my well my kind of official title is head of customer delivery so i'm the sort of head of the uh of the delivery team so i'm responsible for um ensuring the success of of all of the the projects that uh that we win um you know and i've got the you know the rest of the um delivery team um you know that

sort of report into me so um yeah pretty interesting pretty challenging but good fun as well okay interesting yeah interesting i mean there's a few things really um when i spoke to you at the oracle event there was a couple of things there that particularly interested me about kind of yourself really and what you've been doing so the first was that you like you say you've been working with these products for a while and um you know we all know oracle software oracle products can be rather complicated and they can be rather kind of um they can be kind of sometimes kind of difficult to integrate because of the complexity of them and so on and i think the the i suppose that the the obvious way to go with that

is to think it's not very good and it's over complicated but therefore complicated customers with complicated needs and I thought what was interesting with you is that you kind of knew where the bodies are buried as such which is like I would do but you also kind of knew how to get it all working and I think you know when customers buy this kind of technology they buy something that's complex because they need it to do a complex job and what they want from a consultant and a company is somebody who can actually make it work not tell them made the wrong decision and so it was interesting that you kind of similarly come from that same sort of background as me in that area the other thing that was interesting we'll go into both these really is is that you you know the S-based side so you know my background with

with Oracle BI has been more on that kind of relational side it's interesting to see or hear from somebody's implementing Oracle BI but from a kind of or certainly in covering the S-based side as well and the third thing was that you've been doing consulting for a while we ended up spending a good part of the evening kind of kind of drinking loads and setting the world to rights about how to do a consulting project and I think all those three things are things you want to talk about really in this podcast because it'd be interesting to get your take on those and we said at the time you should just recorded the conversation we had at the bar uh and and so you know in a way we're going to do that now but probably with less drink and probably more kind

of uh you know it more kind of uh interesting opinions and so on but let's start off then really with with kind of um oh wait well our client is cloud and airspace so a lot of the lot of the people listening in this recording won't know what they are. So do you want to just very briefly, very concisely tell people what S-Base is and what Oracle Analytics Cloud is, first of all? Yeah, yeah, of course. And how they're meant to be. I mean, full enough, this is, I mean, that's exactly the, when I do a sort of a UK OUG or, you know, business analytics presentation,

that is exactly what my first slide is, actually, because on the face of it, you kind of look at it and you think, well, what is it? What is it? What can it do? What can it do? And that's actually one of the challenges is because what Oracle have done, and they've been, you know, frankly, really quite smart here, is what, you know, they've kind of written a cookbook of what are the tools, what are the components that, you know they've kind of written a cookbook of what are the tools what are the components that

um you know make a business analytics project um successful uh you know and they've not they're not forcing customers to have all of it but they're saying this is you know this is in the toolbox and therefore you've got all of the tools you need to build a really really kind of powerful uh in a bi project but if you don't need bits you don't use those bits that is both incredibly powerful and this aligns with what what you just said both incredibly powerful um but you kind of got to know which bits to to fit where so s base is you know s base is is um you know to be honest when we when um i mean when i was originally using S-Base, it was originally Arbor software, then Hyperion, then Oracle bought it. And to be honest, a lot of people at the time thought, oh, what's going to happen?

But one of the things that Oracle actually have done is they've know, where there's a lot of sizzle, you know, you can build beautiful dashboards. But often one of the things when you look at a BI project, what you always find is the guys are struggling to get performance. And also the cost of change can also be quite high a new new new requirement comes through or let's say a spike in the data comes through and people are instantly scrambling to say well what's going on what's going on and they don't want to wait those sort of you know um let's say for argument's sake weeks often months if it's a particularly complex to to get a different dashboard that that tells them that that spike was nothing to worry about. And that's where the S-Base end really, really hooks in.

Because S-Base's history is it's a very, very high performance aggregated reporting tool, but not just aggregated. It's also an ad hoc reporting tool. And that means that there's typically two camps of people in an organization. There are the detail guys, the analysts that really want to jump in and get all of the detail and look for spikes in the data and look for new patterns and new trends. And once they've done their bit, and that might be a finance department. It could be a marketing department. It could be, you know, kind of more or less could be any department, but they've got that role. Once they've done that, they then know how they want to model and track and monitor the business on a day-to-day basis.

And that's when your BI piece comes in. S-Base bridges that gap perfectly because one of the things you always want to do is make sure that you have a single golden source of data. You want to make sure that reporting is really, really fast. But you also want to make sure that if somebody says, so why has that happened? You don't want to have to phone up the, you know, the bi team and get them to start writing um you know new reports and you know we could certainly discuss also project delivery methodologies because that's not ideally what you want them to you know what you want to have to do anyway but that's typically how things work with customers so s base gives that very very high performance um you know

aggregated view of an organization whether by period whether by product whether by client whatever you know whatever that whatever the sort of customer wants to to however they want to view their organization by entity um but it's an and because that then layers on top of the um classic sort of data mart reporting layer that you'd have on the BI end as well. And you actually end up with a bit of a pyramid where you can say all of our leaf level data that typically would stay. So, you know, pure customer level data or doing some work with an airline at the moment, you know, the booking level data, the real data that can stay in relational. You can then load that data up into S-base where you aggregate it. And it means that the cost of asking new questions of your data via, you know, via Smart View is very low. Next to that smart view sort of financing analyst type person who wants a dashboard, they can consume the same data from the same source at the same time.

And therefore, they're actually having the same discussion about what to do with the business, about the performance of the business. Whereas classically, they're quite separate. And that's actually one of the problems. So what you're describing there, I mean, what you described there could equally, you know, being kind of devil's advocate, could just describe a data warehouse. And, you know, if you talk about, say, with Oracle these days, you've got in-memory database. You know, as somebody who bought one in the past, an Exalytic server, you know,

that was a particularly astute purchase of mine at the time, which I kind of, you know, which was interesting. But, you know, why S-Space? Why go to the trouble of using a completely different technology to what most people are used to, when you could use, like, in-memory database and so on? Does it meet certain use cases but not others, or is it universally better? How do you kind of deal with that kind of question? I'm tempted to say it's – I'd like to say say it's universally better but the truth is it is it is different and and and the fact that it's different means that it

can answer a different type of question very effectively and an obvious um example there is um if you look at any kind of um you know relational reporting approach whether we're talking in memory whether we're talking you know designing our our data mart or or whatever you know whether it's inman kimball you know so on and so forth um classically um an organization in terms of how it views itself does not like to adhere to um structured um levels in their dimensions and their hierarchies. Let's say the way the entity, let's say the way entity rolls up or the way product rolls up. Some bits of the organization might be quite small, you know, a single entity itself might be quite small. But other entities, you know, such as I don't know, let's say an EMEA entity could be very large as lots of sub entities

and other bits and bobs underneath that. That's classically somewhat harder. I know you can do it, of course, but it's somewhat harder to model effectively in the relational world because you end up with either having to take a top down or a bottom up approach. You end up with gaps and this and that. S-base allows you to model the business as it sees itself. So it makes it fantastically accessible. It makes the data fantastically accessible to the users. It's also, and I actually run our S-Base training course as well. And one of the sort of learning points that I share with people is speed of thought as well. So you've got a very pure model of an organization in there. Now, that's not saying it's different to the way you'd model it in the world of relational AI. It's not a different organization model, but it's a very pure model of

how the organization will view itself. And the users can ask questions of that data um instantly i mean the the you know i did a i did a i did an explicit example the other day i'm gonna say no cheating perhaps you'd call me out on that where i took um oh i don't know what was i looking at um 150 million data points i think i was looking at in a data mart now i didn't push the relational data mart as far as i could i purposefully asked a hard question um you know i i asked a question you know that basically said give me the total uh in this case give me the total number of bookings across um five years worth you know of data um how fast is it from a relational data mark perspective we were running uh 22 seconds forcing it to do a full table scan the exact same report um and in this

case i was i was using actually data visualization cloud service to to do the report the exact same report coming using an s-based data source was 0.1 of a second um and i think that's one of the really the you know the really compelling things is asking the exact same question but you know if you're having to wait 22 seconds each time versus you know 0.1 of a second the whole user you know, becomes a bit slicker and people become, it helps the organization move themselves to a sort of better state of maturity. Because when things are fast, they can explore more, do more analytics, they can start to innovate rather than waiting for ports to run. Okay, so one more question about S-Base and we'll get into kind of the way you do projects in a second. So again, something that's kind of happened, I'm conscious that since S-Base, my days,

when it was either block storage or it was aggregate storage, we then had hybrid storage, and then we had hybrid storage running on the Exotic server I had. And now there's obviously S-Base cloud service. So for anybody that kind of knew S-Base maybe sort of five years ago, just give us a very kind of very pot-potted history of where it is now really. So how has the product itself evolved a bit over the last few years? And what's the kind of state of the play like now really for S-Base as an engine and so on?

Okay. So, well, I mean, there are quite a few patents sitting there with S-Base. I mean, the really big change to S-Base, which you will have seen, which you mentioned there, is when they added in the ASO storage engine. The reason why that was so powerful and such a massive change

is what that did is rather than the traditional way of where it kind of pre-calculated everything and in fact there was an art form to sort of tuning tuning the old the old bso cubes and it's well known it's well documented you know that all of the best practices are clear now it's not an art form anymore everybody knows you know knows exactly what to do and how to do it gave businesses an organization sorry an opportunity to model a lot of complexity and to write calc scripts and do all sorts of things. Then along came ASO and what that added was capacity and performance and scale. So in old S-Base, because I tried to do it one time, loading 100 million data points into

a BSO cube and giving giving good reporting performance was easy but it might calculate for two days you know that that was uh you know and they went okay so we managed to really uh push the boat out you know use a bunch of techniques we turned two days into four hours but even four hours is is not acceptable you know customers want to be able to uh do anraday, you know, intraday update, you know, latest snapshot for the GL, you know, of the latest balances, bang, push it in. And just a few minutes later, you know,

look at their YTD performance for, you know, let's say, you know, particularly when they're trying to close the books on month end, you know, that type of thing. They want to fire an update in, bang, get the answers out, have a look at it. So that's where ASO, you know, really, really end you know that type of thing they want to fire an update in bang get the answers out have a look at it so that's where aso um you know really really uh you know starts to win what hybrid then adds on top of that is it gives that um dynamic and sort of high performance aggregation capability but it puts it onto the bso type of cube. So you can add high performance for big sparse dimensions, that type of thing,

but still have the complexity of if you're trying to run complex calcs, for example, you know, multi-layer allocations, you know, top-down forecasting, that type of thing as well. So that kind of bridged the gap between the two i would say in the in the bi space most commonly it's um it's a it's an aso cube that you use um principally because um it's normally it's almost always a bottom-up calculation it's always a bottom-up aggregation of some type that you want to do things that they've added that have really moved that even further forward is you can actually run calc scripts in ASO now. What that means is whereas before you could only

do a bottom-up aggregation, you couldn't, for example, run an allocation. You couldn't, for example, take last year's data and copy it into this year's to create a seed forecast. You can now do that in the world of ASO. So BSO is still there. It's still, if you wanted to run a multi-layer allocation, for example, or run top-down forecasting, that type of thing, you'd probably still use that. And it's still the backbone of PBCS. But the power, the performance, the capacity, the scale that you can do on the ASO side is increasing and increasing to the point where it kind of starts to then touch or certainly overlap what you would have traditionally done in relational.

And that's really nice because it means when you're looking at solution architecture there's not a gap anymore there used to be a gap where you were um you know when in the particularly when it was bso versus you know bso cubes and and sort of you know relational data marks there was always a little bit of a gap in terms of the capacity you could put in the cube versus how quick you could get aggregations to run in in in your data mart aso has really really bridged that gap um but not taken anything away as well it's it's been it's been very very additive so you know so that was kind of the last few years of of s base and of course cloud now um you know they've um you know they've done a you know frankly they've done a you you know, frankly, they've done a really nice job. It's still early days, but it is S-Base 12 up there in the cloud.

You know, and for example, one of the things that always used to be a little bit challenging was S-Base had a very old security model. What they've done is they've re-architected all of that. now integrated into basically an Oracle single sign-on platform, IDCS, Internet Directory Cloud Services, I think, where all of your BI, Publisher Cloud, or Bix, Publisher Cloud, DVCS, and S-Base, it's all single sign-on now as well. But it's single sign-on with Active Directory integration to your customers on-premises Active Directory environment as well. So, you know, that's the kind of direction they've been going on and it works generally, it works really nicely. Okay, so imagine now, imagine I hired your company and imagine I hired Incubix now to do a, you know, you've won a tender to do an implementation of

OAC with potentially with that space but at the moment it's just OAC and there's a bit of kind of maybe there's a bit of a bit of kind of up it's a bit maybe of migration of on-premise stuff into the cloud and there's a bit of kind of new development what's your general approach to going into starting a project so you're going in to start a project on OAC how do you typically kind of like phase that project and what are you looking to find out at the start and generally how do you deliver these kind of projects really um yeah good question the in fact it's particularly good that you kind of said so it's an oracle analytics cloud project but may or may not be s base and that you're absolutely bang on the money so how do you make how do you make that

decision so one of the one of the questions i often often used to get from from architects inside the companies i'm working with is how we make a decision do i go for this this or this how do you make how do you make those decisions and decide on the architecture you're going to use really so the first question is um well one of the ways we position it actually let me let me let me present it as what my approach would be rather than you know questions to ask is if the customer has an existing estate you know of uh you know on premises obi 11 let's say for argument's sake the first thing we would do is we'd actually run a health check of that what that does is um or the outcomes that we want to generate from that is understand what the customer's pain points are

because if they're running that existing system and they say, we love it, we love it, everything runs really fast, everything runs really well, we're going to take a good look at what they're running at the moment and look for recommendations. But what we're less likely to do is say, okay, maybe we need to review your solution architect

at a bit more of a fundamental level versus if a customer says, well, yeah, we love all of the dashboard and we kind of love the reporting, but we don't like how long it takes to do data updates. We don't like the fact that maybe BI is disconnected from the S-based guys. I mean, that's a classic sort of problem,

a classic thing where the guys might, they're running on a different schedule. The guys are reporting different numbers. Maybe they're just saying that the reporting performance is slow. So we kind of look at the pain points that the customer has. And then off the back of that, we can then put together a number of recommendations. And those recommendations could be a straightforward lift and shift it could be a lift and shift with some augmentation or as in you know in the case of of you know what we've what we've seen with with some customers where they're saying well actually it's really slow and you know that the you know the the cost of building new reports is very very expensive that's when we

tend to you know sort of do more of a bottom-up uh review of the solution architecture overall and actually understand why and you know as i'm sure you well know there's a whole bunch of classic classic mistakes that that um you know people make mistakes that i'm sure you made early days mistakes certainly i you know i didn't make them yeah yeah exactly exactly i don't learn i learned from other people's mistakes yeah of course um you know and um you know but but one of the things is where where customers have particularly you know i see that type of thing where customers have have have said right well we're going to do this ourselves we've got a bunch of contractors in and uh you know we're going to build this ourself and typically

they get you know let's say an obie contractor but that you know he we're going to build this ourself. And typically they get, you know, let's say an OBIE contractor. But that, you know, he might be fantastic at building dashboards that look beautiful. But that doesn't necessarily mean he knows how to get the best of modeling the data in the physical layers. It doesn't even mean necessarily that he's particularly good at modeling it in in the rpd um you know if he doesn't know um you know you kind of have to know pretty well all of them um you know between relational between the you know the visualization end between the rpd um and and s base and only once you've got that oversight of all of them that's when you can make um you know highly valuable recommendations to a customer about how to give them you know performance consistency beautiful reports you know and

everything um you made a good question the point you made there actually just interrupt you i mean you talked about contractors there and i think one of the things that's interesting is to think about is what do you think is the state of knowledge of contractors and generally people working in the market now of the oracle products because there was a period you know probably about sort of five years ago when I think you could probably argue that OBI and S-Space in their standalone kind of like versions were fairly well known you know I had a part in that too I wrote a book on it at the time and and you were working in that area what's the state of skills like now in the market do you think and you know? Being frank and being honest sort of thing.

Yeah, yeah. Well, funnily enough, I was listening to Christian Berg's podcast that I think – when was that? You did that. That was your Christmas and New Year special, wasn't it? Yeah, yeah, yeah. Longest one ever, actually. Yeah, that was interesting. I know.

I did look at that one and think, wow, you guys had a good chat there. But to be honest, I think Christian christian made some really really good points because one of your questions to him is you know why have you always been a contractor christian and he was kind of positioning it as saying well you know the reason it's worked for him is because by being a contractor he can go in without an agenda um you know he he can make the recommendations of what he truly believes is right for the customer without having to necessarily follow the corporate playbook that that you know one of the big four might might uh you know might feel like they need to um and he was absolutely right in that but equally he was simultaneously absolutely wrong in in my opinion as well because whilst big four um

you know have to kind of typically follow the corporate playbook um we're cubics our entire reputation is actually off the back of um being able to make recommendations that are right customer and that work for you know that work for the customer but where we can you know where as a consultancy organization we can we can add more value is it's very hard for one person to know absolutely everything so as you know as a contractor or even as a you know let's say as argument for sake for as a team of contractors right you go into a project and there's let's say maybe four contractors on there they're not necessarily going to work as effectively as a team because they've all got their own remits they're all they've all got their own agenda i was a contractor for four years

myself right you're um you know it's the there's a lot of contractors where their remit is get a renewal get a three-month renewal get six-month renewal um you know and and therefore come become sometimes a little bit protective. Now, one of the things I really noticed with what Christian was saying was actually he's made his reputation by not being protective. He's made his reputation by doing a good job and actually having the difficult conversations with customers saying, I think there's a different way of doing it. And that's one of the things that as cubics we can do as well is because if we roll four people onto a project um if the project changes let's say for example you

know let's say the you know the customer came up with with some new requirements that required a different solution architecture through in the world of contractors everyone's going oh my god but but you know that the contractors are going to become defensive. We can very easily switch. We can very easily change and switch in other guys with different skills or so forth. So it gives us flexibility. we're not kind of bound by that uh you know corporate uh you know clipbook of uh approach we can liaise much more closely with the particular business lines that are trying to solve a problem um but you know communicate both up and down the chain as well so that's kind of where i

think yeah it's interesting interesting i think i think christian's angle as well a point as well was was that you know as a business there is there is that, you know, that you become very, we'll talk about this later on with the kind of business consulting conversation. But, you know, there's always a kind of a desire to maybe put less experienced people on the project to get them kind of experienced, and things like that. So there's, there's, there's things that go on from a business perspective. But, but like a business perspective, but equally, you know, I've met contractors who can be very, very earnest, but can be wrong as well. And I think it all depends. What about actually kind of knowledge of the products as well? Because there's been quite, I mean, the conversation you and I had, you know, in Athens was talking about, you know, every software,

there's a kind of a big, massive change in the kind of the platform, really. It went from, say, sort of Discoverer to siebel analytics and siebel analytics to whatever and and there's i suppose there's a natural kind of reluctance sometimes from people and i'm not putting christian in this in this kind of category but a natural reluctance to kind of go and learn it all again um but and cloud is interesting as well because cloud i think presents its own challenges about actually getting hold of the software as well yeah it's quite hard to for an extended period of time to get hold of oracle software what's the level of yeah what's the level of knowledge like now within the within the

industry of these products do you think actually before i answer that what i do need to say is i was actually completely agreeing i thought i thought christian um you know provided very very earnest and and very very insightful responses but i also thought he's atypical of contractors because what you often find with contractors is um you know uh you know particularly i've worked with many many contractors over the years there's often is yes yes yes i can do it um you know and then actually you kind of get them in and you can see they're furiously googling away and that that, you know, that's some, you know, some contractors. There are, of course, an awful lot of contractors that have been doing it for many, many years and are fantastically good.

But the risk is they then become a little bit singular as well, because, you know, there's, you know, in fact, one of those, there's a guy I've been, you know, there's an S-base guy i've known for 20 years he still really only knows s-base bso because that's all he's been exposed to and it's hard as a contractor to make sure you stay at the at the forefront but you'll you know the sort of contra view that you shared there is you know and um you know it's a particularly well-known you know that the big four do this where um you know they spin out spin up the big guns to win the sale um you know, it's a particularly well-known, you know, the big four do this where, you know, they spin out, spin up the big guns to win the sale, you know, do a bit of solution architecture with the big guns. And then the graduates come in, you know, and the graduates, you know, you know, fantastically hardworking. they don't know. They just haven't got enough projects in their pocket to know that, okay, the documentation says that should work, but the truth is, here's a much better way of doing it.

And they have to learn the hard way. You have to learn somewhere, don't you? And also customers, they ask for blended rates typically. They want you to bring the rate down so that it's either that or send it to India. I mean, do you still see much um i suppose offshoring going on within the industry uh now or is it is that has that gone away now or what um actually i'd say it's um as full speed ahead as ever because funnily enough i was uh you know i had as uh

you know we were discussing this this kind of exact topic with my fellow directors at Cubix yesterday. Because we've got an India office. And what I see is there's kind of very clearly two different types of projects these days. There are the larger projects where you're, let's say for argument's sake your your build team are five plus and what we're seeing is that customers very clearly and very specifically want that uh you know local face who is the sme who is you know who is the solution architect who is the the go-to guy but at the same time they don don't want to spend on having that entire locally-based build team that they would have classically had. They want a lot of

that build capability then to go offshore. So we're seeing that's one of the things customers are particularly liking now. And to be honest, we're finding that works really, really well. I mean, one of the things we've tried to be very careful of is make sure that our India team are not just developers, but they're consultants. And, you know, we've got some fantastically good guys, you know, particularly some of the senior team in India, so that, you know, they're, you know, so that they're able to fulfill a broader spectrum of roles. There's that type of project, but there are still many, many instances where customers do still want the local footprint or UK-based footprint. And because it's part of a much, much larger program,

what they want is the faces where they can, you know, they can just, you know, kind of grab the guys and pull them in a meeting room and say, hey, we need to do this. One of the things I'm not seeing, and I'm going to say at the moment where the entirety of it is india-based unless the customer has a has a lot of you know has a strong india uh base themselves already um you know they still want that that sort of you know that that customer success manager or the solution architect or the relational relationship manager locally and then they want to sort of leverage the offshore resources to to manage cost and to increase capacity increase scale so it's kind of a hybrid uh you know thing that that's that we're saying is is sort of what works for customers at the moment so so are you finding then i mean other conversations i've had with people working in the kind of s space and obi space is that you know cloud has created an expectation now that projects will be kind of quicker to deliver um yeah and certainly with um certainly

with the infrastructure side and the wiring together side in theory that should be a bit shorter are you finding that are you finding that that um projects now are coming in cheaper in terms of time and days and so on or always other customers does their kind of does their kind of scope and expectation grow meaning that the project is still as big i mean what's the what's the size of a project like these days and what's the kind of budget like and how's it go well certainly getting an environment up and running for a customer is um from if you know in the old days, you would be, you'd be waiting a month or two for the, you know, for the infrastructure team to plug a server in,

then have the Windows SAs to, you know, to install Windows, you then have the DBAs to, you know, it could be a long lead time and, you know, certainly many tens of days in total to spin up an environment, even if the, you an environment even if the you know even if the bi piece or the s base piece itself was was was quick and easy um with um oracle cloud you can have a fully functional very large very powerful um database s base and BI server running in less than a couple of hours. I haven't timed it, if I'm brutally honest. It is still quite fiddly. It is still quite fiddly with OAC, though, isn't it?

I mean, I think that... Well, it's fiddly the first time. Yeah. Talk us through kind of like... Talk us through what's involved in an install and a setup, really, and where it can be. I suppose some of the tricks you found or things you found that work well or whatever on that yeah i mean you're absolutely right it can't the the very first time

i mean i remember the first the first time i you know i thought okay let's you know let's uh let's make sure cubics are the guys that know how to do this ahead of anybody else so you know so a bunch of us uh you know when it when it very first came out we whipped out the credit card and uh you know signed up and on off we go so yeah that very very first time is a little bit um fiddly because there's there's a few unknowns but once you know what you're doing it's highly repeatable and in fact it's highly repeatable to the extent that you can actually automate it you could actually use the api to spin up an entire OAC environment with just a few standardized command line scripts, you know, using curl if you wanted. But some, I mean, certainly some of the gotchas is you need to design the cloud networking the right way if you want to use a VPN.

Tell us about that. Tell us about that because that's kind of non-obvious, really. We talked, I think it was about three points in, we discussing this on uh in athens and where does the complexity come in there and why why does it need to be dealt with and why is it there in the first place well the reason why is when you do um an install of all of the products basically you always end up with the ip address and the host name of the different nodes so that they know how to how to talk to each other so the complexity comes in is that it's very easy to spin up an environment and more or less hit next next next however once the machines are built and they've got their ip addresses and their host names sort

of embedded in the in the applications a little bit if you then want to integrate that into the customer's environment you need a vpn now one of the things that um you know whether it's bt or ibm or you know even the customer's own you know infrastructure team always want to do is they want to architect vpns a particular way they want to they want all of the machines that are in the cloud to have uh to have the vpn using public ip addresses inside the private network and you know and uh now that's not a technical requirement but it is always with with all of the you know with all of the you know with all of the sis we we you know we've been working with that's always a requirement.

Now, what you don't want to have to do is to then kind of have to go and do a bit of head scratching and thinking, oh, we have to go and re-architect our OAC environments because all of the private IP addresses are on a private range, you know, 10 dot something or 192, 168 something. You know, that would be the default way it gets built because it's the simple way to build it so that's one and a great example of something where an early question you want to ask the customer is you know is do you are you happy to run it as sort of standalone cloud or do you want to integrate it to your on-premises environment through a vpn um you know so that's uh you know that's certainly i mean mean, I don't mind saying I know we were the first guys in the world to do that because we, you know, because we actually had to get Oracle to do a little bit of a bug fix in there. And the day they made that release, we sort of had that environment built for one of our customers. So that's kind of one of the

gotchas. But there's a whole bunch of other gotchas as well that you just, once you know, you know, and they're easy, always use high performance disks. Actually, that's kind of one of the gotchas. But there's a whole bunch of other gotchas as well that you just, once you know you know and they're easy, always use high-performance disks, actually. That's a really big-ticket item. If you don't know it, you go through and you hit next, next, next. You get standard disks, and you think, oh, I thought this was going to be a bit faster. And instead, what you do is you add more cpus you add more memory truth is both in relational database and s base you want

the fastest possible disks and they're really low cost oracle have done really really nice job actually of making storage cheap so um for relational databases and s base less less important on the bi side unless you're going to turn caching on um you know always use the high performance disks but again it's one of those things if you miss if you've missed that off the order form at the beginning you know that could be a bit of head scratching on on how to how to do it so okay lots and lots of things you need to bear in mind so so one like one real last question let's face now it's quite a few obviously you've got to talk through here but um so when i last got my space training course by the way no i i only ask questions i know the answer to you see so i know i so so so actually there's one thing i'm i'm curious about so when

i last left uh obi-12c s space was in there as purely as a kind of query accelerator with a new API to do this kind of query acceleration job, which struck me as kind of a bit of a kind of a last century solution to a problem these days is solved by things like in-memory caches and so on. I mean, did that ever go anywhere? And is that being used? Or I mean, for anyone listening, maybe just kind of recap on what that is.

And is it there now? Or is I mean, for anyone listening, maybe just kind of recap on what that is. And, you know, is it there now or is it kind of like a blind alley? Well, if I'm honest, in terms of the way I look at solution architecture for customers, I've always found it to be a little bit of a blind alley. Because if you're going, you know, and this includes whether or not you're doing cloud, whether or not you're doing on-premises. If you are going to the effort of building an S-base cube, I've always felt that it's worth then putting all of the extras, all of the good stuff that S-base adds on as well to facilitate and enable all of the ad hoc reporting capability, the built-in financial intelligence, the built-in

time series intelligence. And therefore, if you're just using it as an accelerator, you're missing out a little bit, you know. So, you know, so that's why if S-Base is in the mix, I would always custom build the S-Base cube. It doesn't take much extra effort, actually. It's, you know, it takes, you know, because by the time you've built your data mart, you've got your dimensions, you know, you've got your fact table, your dimensions and that, it doesn't take much extra effort to have a, what I would call, optimized S-Base Cube design. But the upside for the customer is quite profound because it's not just an accelerant then it it gives you all of that ad hoc all of that you know you know ragged hierarchy capability time series financial

intelligence you know so it's a you know it's an and and and and you know in terms of the the capacity and the capability for the customer your other point then about in memory is um is one of the things I mostly see is if customers are struggling. So let's say, let's leave S-Base completely out of the loop here for now. I'm going to say 70% or 80% of the cases where I've seen customers struggling with BI performance, in almost all of those cases, well, the 70 or 80% of cases, what I actually find is that they've not modeled their data effectively in the physical layer.

And inevitably what they've done is they've turned on caching or something like that to try to put a bandaid over the fact that what they haven't done is modeled the data correctly and efficiently in in the physical layers underly underneath so um you know in memory yes caching yes always as a time and place for both of those but what i actually usually find is that it's relatively straightforward to look at the way they've done the physical modeling, the way they're storing the data just in the underlying physical layers. Whatever that physical layer is, you can usually get a significant extra amount of performance out of it just by modeling the data more effectively.

And a classic error customers would make, You know, they've, you know, if their question is mostly, you know, I need to know a year to date, I need a year to date number, they've usually got the different months of the year in the rows. Classic thing, pivot it into the columns, 12 times, you know, 12 times less number of rows,

12 times faster queries. Such a big win for a customer. And then your caching and, you know, and sort of in-memory database, they can kind of become the icing. You know, you're not using them to underpin the performance. You're using them to turn the dial up from 10 to 11 rather than from one to, you know, five,

maybe if you know what I mean. Interesting. I mean, having worked in the startup world the last kind of year and a half, two years, with um i suppose data stores like bigquery you know the thought of data modeling it's like kind of discussing carpentry or kind of falconry or something it's it's something that is is kind of i suppose in a way not needed but he's not even thought about really um yeah but but but then you know but certainly i think like you say i mean you do even with services like bigquery in the end you have to kind of be aware that if you're going to join two big tables together, that's not going to work.

You know, you've got things like nested columns and so on. And I think it is, like you say, you know, you can throw, I always thought caching was the last desperate throw the dice on a project where you really hadn't done your homework in the first place, really. And I used to always laugh when we found, you know, OBI best practices things on the internet and it would always say on there turn off logging and enable caching and I think at that point that that is that that's kind of like that used to kind of help Mary yeah exactly so yeah so the other thing really I want to talk to you about was when we got talking about this again in Athens was was your take on consulting really so so you know tell us a bit about I suppose in a way let's start with recruiting okay so you're gonna imagine i've i hired you as my kind of head of delivery or

something in a company where we were going to uh build a consulting team what would you look for in people and how would you build start to build a team really and what kind of what kind of characteristics of people you'd be looking for really in terms of a good kind of mix of people in the consulting team the yeah it certainly is um different to what it used to be in you know let's say even 10 years ago it was very easy to have a significant proportion of the of the of the consulting team being more or less purely technical you know they could be heads down you you know, behind the keyboard, just just, you know, bashing bashing code out code out kind of guys. One of the things that's changed is because the products have more capacity and more capability immediately built into them. There's less of a need for, you know, for that, you know, coder type of person.

And it's much more now about understanding the business and about guiding the customer to do things the right way, because there's so much that comes out of the box. There's there's much less of that purely technical. I mean, it's still on the outcome for the customer. Very, very personable, but also very good at managing the expectations and the relationships of, you know, with the customer. And in fact, we're doing, you know, we're doing a project at the moment where one of our guys is in there doing some business analysis. And one of the things the customer has done is they've given them a sort of a spreadsheet or Excel workbook with roundabout 35 sheets of layering of all of the calcs and all of the analysis that they want to do.

And the message that we're sort of working through with the customer in that example is you could do that but these are what other customers in you know similar state to you that you know similar place to you trying to solve the same problem are doing they're doing it this way and you know and the you know and they're doing it therefore much more quickly much more cheaply much more cheaply, much more cost effectively. So that, you know, so the project is costing them less. And that's a lot more part of what the, you know, what a consultant needs to be able to do these days is not just bash code out, but actually, you know, guide and train and, you know, sort of really take the customer along for the journey of not just saying, yes, OK, we'll write you some code to

solve that problem but very much saying you know in our experience if you want to solve that problem here's how we would recommend doing it and in fact one of the things we we teach our team is you never say no to a customer because if you say no you know the conversation has ended and this is a you know this is a counterproductive you know that's a counterproductive conversation so the answer you know or the technique actually is yes but so yes we could do it that way but if we do do it that way here are the consequences of it so we recommend we would do it this way instead um now sometimes you know you drill into a bit more detail and you know you kind of you know you have to do it one way instead. Now, sometimes, you know, you drill into a bit more detail and, you know,

you kind of, you know, you have to do it one way or another, but it's those types of interpersonal skills, you know, those types of, you know, the poise and the presence, you know, that consultants need these days to be able to command a room, to be able to, you know, to collaborate and liaise with a customer and sometimes have those difficult conversations where you need to say to the customer there's a better way of doing it we know you've been doing it that way for 15 years but there's a better way of doing it now and uh you know persuading the customer that uh you know to do that so that's the type of sort of persona and type of you know profile of a person

that we need and funnily enough just two weeks ago um we we had a cubics vision day and and the majority of my presentation throughout the day was on tips and techniques to you know to uh you know to take customers on that journey yeah so do you tend to recruit do you tend to recruit kind of experienced people or i know you're going to say in the end it depends in the mix of them but do you is your strategy to employ people who are the finished objects or people who are very junior what what do you do what how do you what's your strategy around recruiting at the moment well when we spoke in athens i think i did say depends um because i was thinking about it and i thought yeah well

you know it's it does depend it's a mix isn't it really I mean okay put it a different way on a on a if you have a pod or a kind of a project team for example how do you try and sort of balance the mix of of kind of I suppose young people new people experienced people that sort of thing what we're finding and what what we're doing these days is when when I started um it was very easy for me to fly in on a project, keep my head under the radar, you know, bash out a bunch of code, you know, a bunch of bugs in it because I was learning on the job and, you know, senior who was kind of keeping me under their their wing a little bit because the product and because the tech and because the expectations of customers have moved on so far these days that's much much much harder to do so so actually at the moment um we in terms of a uk recruitment policy we pretty much exclusively go with guys that have got at least three more likely five plus years of experience um now it might be that they've got for argument's sake five plus

years experience of on-premises builds you know that that you know that's that's a very typical thing but if people have that project experience that customer facing experience they've done enough projects to know what works what doesn't enough projects to know what they're good at as well i mean that's knowing knowing yourself is one of the key key things know what you're good at know what you're less good at and either put the effort in to become better or learn how to you know learn how to to body swerve the things you're less good at so people who have that self-awareness um so yeah i'd probably say we recruited a more you know or not probably we in the uk we're certainly recruiting at a more senior level these days with the hardcore more

pure technical guys being you know the offshore resources these days so how do you how do you make money on that then i mean because it's it's yeah the thing that certainly i found used to find in my days in this area was that you know you you would the the margin you would make on on kind of new stuff coming in um was the margin that kind of ran the business and people who knew people who knew what they were doing people who experienced also knew the value of what they were doing and also wanted to be able to go to conferences they wanted to be able to kind of write blog posts they wanted to kind of get some time down so how do you balance that out really how do you how do you make money on those people and motivate them um and and make it all work really it's all about making sure you provide good customer value

um customers customers are happy to still you know to pay a good rate if what they're getting is top class, you know, consultancy expertise and recommendations. What they're not, what customers aren't prepared to pay for anymore is someone who sits there and turns up and says, tell me what to do. You know, they're the types of guys where you know customers go whoa uh you know but what they want is is you know guys that land that it can very quickly start saying to the customer well you know i understand your business have you looked at this we think you should explore this we you know being very very proactive so at that end of the scale now and that's not necessarily your principal consultants uh you know just your principal consultants you know you've got you know sort of practice

director principal consultants senior consultants everybody in the organization needs to operate in that way um you know and that's a sort of you know that's still a very broad spectrum of of of people but the very junior the graduate end that's where there's less of an opportunity in in in the uk now but it's very cyclical right i mean you've been in it long enough as well uh you know it's very very cyclical um you know and uh you know and to be honest that's why one of the reasons why we have our own sort of marketing uh uh department why we have our own sales teams as well, because because it is so cyclical. You know, we work with Oracle.

You know, we're exclusively an Oracle platinum Oracle partner, but we work with the Oracle sales team. We've also got our own initiatives, our own, you know, our own you know our own campaigns we run our own products as well uh to to take some of that uh you know cyclic nature out of it so that we can uh you know we can ride the highs but the lows aren't quite so bad for us either and to be fair i mean obviously you're not one to blow your own trumpet but cubics are pretty good pretty good in this area and in terms of quality and i've known roger is it roger the guy who founded it um roger cressy i think it's roger cressy isn't one of the founders yeah roger's yeah well exactly right roger the original founder of the company actually was a gentleman called julian owens um and he is you know he he still is the ceo he lives out in australia

now uh you know he he so he founded the original company when it – when that original company was – I'm going to call it upgraded, but, you know, was sort of wound down and replaced with Cubix International. It was Roger, Roger Cressy, who you know, Julian, and another gentleman called Ed, Ed Halsworth, who is also in the Australia office now as well. They were the sort of three three founding members of cubics and then myself Paul and Neil albeit you know particularly myself and Neil we joined the company 20 years ago junior guys worked our way up you know through through the years one more question for you how how has cloud affected projects then i mean obviously oracle were late we're somewhat late to the game with cloud really but but they have an offer which is very sort of differentiated um there's the you

know there's the there's there's the changes in the way you can deliver things in the cloud and there's and there's different products and so on how has cloud affected the the business really of cubics over the last kind of couple of years it's turned it on its head to be honest because as you note oracle were well in in in the business analytics space oracle were were very late to the game they dropped off the gardener quadrant you know there's quite a lot of uh you know certainly a lot of a lot of contractors a lot of customers and a lot of everybody going whoa whoa what's going on here um you know so where they were um you know i would say on target with with the game was on the on the pbcs side one of the things we one of the things we clocked very very early on was that cloud was the future

we saw the we you know we saw the massive um you know massive impact that aws was was having across the board we realized very very early on that that cloud was absolutely going to be the future so we've made sure that um you know that we you know to to be right at the forefront of everything cloud and that's one of the reasons why i'm so evangelical about about uh planalytics cloud is because you know we made sure we were right at the forefront as soon as it came out we were you know we we were we were you know bashing the credit card hitting it hard testing it making sure that we knew i'm going to say more than everybody else i'm sure there's you know the different people you know to focus different areas but making sure that we were right right at the front of the curve.

Because what it means now, when we're looking at sort of customers want to do an upgrade, they want to go from on-premises. In fact, no, customers generally just want to do the right thing and they kind of almost don't mind whether it's on-premises or cloud. They want to do what's right for them. We need to be able to look customers in the eye and say, cloud is right for you and it'll save you money and it gives you more features and it gives you more capabilities. So we made sure we were right,

you know, very, very, you know, very much at the front of cloud and continue to be, I hope, you know, very, very much at the front of cloud and continue to be, I hope, you know, very, very much at the front of the cloud. More than 50% of our business now is cloud, you know, which given how recent Oracle Analytics Cloud is, you know, we've got, you know, quite a lot of customers already on Oracle Analytics Cloud, you know, and, you know, the rest of the Oracle Cloud products as well. You know, the HFM space, planning, those sides as well.

I'd say it's rare that it's been on the premises project now. I'd say the first port of call on all of those is cloud first. So back in my days of OBI, the thing that drove a lot of OBI business was sales of e-business suite. So that would drive that. Is there still something external to our EC&S space that's driving sales of those products? Or is that falling away now?

I mean, I know Fusion was there as an app suite at the time. Does that still exist as an outside driver? Yeah, very much so. And in fact, Oracle sort know regularly do a bit of a bit of a reorg as well and and you know technically you know basically to realign things i mean in fact there's just been another there's just been another reorg now um you know so yeah very much so and you know certainly what we saw was um you know well what what we did was we sort of engaged in in some partnerships with you know with sort of e-business suite uh you know well what what we did was we sort of engaged in in some partnerships with you know with

sort of e-business suite uh you know vendors and so forth that didn't have the s base or the you know or the or the business analytics or the you know visualization experience and uh you know and they were kind of saying oh well you know we know what we're doing over here but how are we going to do that bit as well and and you know conversely we were saying well you know okay we're you know at least at that time we were like well we're probably not going to make that investment into e-business suite um you know because if we're going to do it we want to make damn sure we can do it well um let's partner with somebody you know so that was what we did you know at that time what we're finding now is that um you know with with again particularly with the Oracle Analytics Cloud platform and all of the

other cloud projects products sorry I should say the projects are um are still very standalone they're perhaps more more standalone uh than we expected them be, which is to our benefit. There's been the occasional opportunity that we've not won because we couldn't do the EBS end, you know, but I'd say that's relatively rare. You know, I'd say that the segregation of, you know, different companies doing different, you know, or implementing different sort of systems is uh you know it's certainly still there are you finding there are you find sorry are you finding there are sales going on to customers who are not already kind of like heavily embedded with oracle so you know would you find people were going out and buying oac with their space and they're using say sap or something or they're using salesforce i mean or is it is it selling purely to people who are already heavily invested in oracle

at the moment um maybe not heavily invested but they are mostly customers who already know the good things about oracle um and i think that's actually one of the one of the challenges at the moment i was actually over in uh dublin a week before last i did a presentation to the you know to the to the um you know to the oracle digital team out there because one of the things that has become quite obvious is and this loops back to the point you made at the beginning of the conversation of what is oac the you know the tech reps were looking at OAC and saying, well, what is it? How do I position it with a customer? Because there's so many things it could be and you, you know,

and that's actually been, been particularly one of the challenges. That's exactly why we are, you know, you know, why I went over to Dublin to do, you know, to, to, to do a presentation to the reps over there, to help them understand what to position with customers, how to move it forwards. Because it is really good. It is very powerful.

But you've got to understand the customer that you're trying to sell to and then know what you're up against as well. It's easy to misfire. well it's easy to misfire it's very easy to misfire so you know and again that's where i you know i hope we can uh you know continue to do well last product question for you now you're probably the only other person i could ask this question to and who would know the answer so um all right so you saved the hardest one for no so so when so at um at the event we were at they announced uh oracle autonomous analytics cloud uh which which i i thought was a practical i thought it was an april fools joke

but actually is actually a real product and i noticed it came out recently and it's actually documented and actually available um so have you seen it and what's your take on that really and what just explain what it is maybe to people without understand if if i'm brutally honest i haven't looked at it myself. I've got some of our team looking at it at the moment, and they're kind of currently working on it and reviewing it at the moment. I have to be brutally honest. I haven't actually looked at autonomous.

I can actually go into that then, actually, interestingly. I had a look yesterday, actually, and I was looking through the stuff on the OAC data lake. What it is, it's interesting because it's basically more more so you said earlier on about you could actually do an install of an OAC cluster with a couple of kind of curl scripts and so on curl calls and that's probably what it is you know you you install it through it's called oracle stack it's like I think it's templates effectively so it's a bit like the old days when you used to have those oracle vm templates when it'd have a whole bunch of stuff you could you could install e-business suite for example using a template it's like that so

it's it's it's it's basically OAC but it's with an easy install and I think a lot of the stuff that a lot of the controls and things that are exposed in OAC are again sort of taken away it's not got the same kind of I think machine learning approach to kind of tuning itself in the background but I think it's a bit more Oracle managed so it's a bit like an oracle managed version of oac which is a bit of a kind of contradiction in terms but yeah it's interesting i certainly do see um aligned to that a lot more things sort of coming coming down the pipeline i was talking to uh some of the oracle field reps yesterday because right now oac runs bait or underneath what's called oracle uh oci classic i you know sort of classic infrastructure and what there is is a shift to move everything

away from from classic into you know pure oci to to to simplify it to make things more scalable to make replication from you know a production to a DR site much more tight, much more integrated. So I certainly see that as a direction. And principally, anything that reduces the cost of ownership increases the value and reduces the time. I it's it's pretty profound what's happened so far um you know and uh you know in fact did you i'm not sure if you saw did you see the ai and machine learning presentation in athens i think i did yes i think i did which one of them yeah yeah yeah yeah a gentleman called matt matt stillwall oh yeah yeah yeah it's very yeah i mean i'm a huge fan of matt's because i really i really love his presentation. He really knows all of the products fantastically well.

And, you know, he's happy to share everything he knows. I actually got him to do a playback of that Athens presentation to our team a couple of weeks ago. Because I think that's where the really, really interesting stuff is going to start to happen. Plumbing is just going to happen, you know, but where's the more, you know, where's the increased value to the customer for that? But some of the things that are just around the corner, both in the DVCS world, you know, you can run a neural network directly in data visualization cloud services. For example, that was one of the things Matt showed us, you know, or the AI and machine learning cloud services as well. And the particularly nice thing is they've not architected the AI and machine learning cloud service to be standalone. It can do lots of things and run very standalone,

but there is an absolutely fantastic API in there as well. So when you build your machine learning or AI algorithm, let's say you've spun up a neural network. Once you've trained that neural network, it actually creates an API endpoint for you so you can embed it in almost any other product. That is fantastically powerful. So, you know, so that end of things, that's where I'm seeing, you know, I'm getting really, really excited, as you can probably tell.

By the way, my degree in the 90s, I was doing AI back then. I remember having to run a genetic algorithm for a week for my final year project and the only reason i didn't do a neural network you know that was optimized with the genetic algorithm was because i just could not get the computing capacity so that's the power these days right cpu you know in fact gpu capacity the power the capacity the capability now that's where you know there there is going to be such a massive change in in our industry now because it can try so many different so many variations you know training a neural network used to be very very expensive you can train a neural network you know of of a let's call it

mid-size in a couple of hours now you know and once it's trained it's lightning fast um you know so um yeah that's interesting it's an interesting thing you say that i mean you're talking there about the future technology yeah and you're talking about things that motivate you i mean you you've been doing this for a long time now and you sound massively sort of enthusiastic and motivated and so on what what what what motivates you and what drives you on to kind of keep doing this? Because consulting is an interesting game in that, you know, you have a lot of failures in projects. You know, you see a lot of kind of things repeating themselves. And, you know, it's a hard game to be in.

What motivates you to do it and how do you still keep on top of it like you seem to do at the moment? Yeah, well, one of the key ways I stay on top of it is you seem to do at the moment yeah well one of the key ways i stay on top of it is i drink a lot of coffee yeah that that you know i need my morning coffee um but aside from that um to be honest one of the things i realized and and when i when i was talking to you earlier about what what what is a perfect consultant you know one of the things that i think is is very important for a consultant to be is self-aware and one of the things i learned about myself um you know many many years ago now is um i don't like things being steady state because nothing is more dull doing the same thing again and again and again and again that's a factory worker right that's you know that that's you know the factory worker pulling

a lever you know that or you know that that's um you know, the factory worker pulling a lever, you know, or, you know, that's, you know, that's not mind food at all. But you've been doing S-based projects for 20 years now. I mean, is that not the same sort of thing? I mean, how do you motivate yourself to do another S-based project? There's very few pure S-based projects now. The motivating thing is, you know, it's almost never, you know, there's never going to be a PRS-based project anymore. It's always a business analytics project now. It's always where you're putting visualizations on top. You've got an EDW or a data mark behind it. You're always trying to push the limits of capacity, of and um you know looping back to my sort of point about machine learning and ai i think there's a lot of hype in the industry as there was about big data there's a lot of hype in the industry about what machine learning and ai can do um but the truth is

a lot of it's hype, but underneath the hype, there is, you know, whereas people used to say in the, you know, in the Hadoop world, okay, you know, brilliant. Yeah, we've got Hadoop. Right, what should we do with it? And everyone kind of scratched their heads a bit.

In the machine learning and AI space, everybody's getting it and the things it can do are actually, if it's implemented the right way are quite profound um the you know some of the prediction capability some of the some of the things you can do because the computer is faster and because the apis are now there because the interfaces are there because the visualizations are there some of the things you are actually going to be able to start to do are really profound those repeatable jobs

let's look at finance right finance running month end a lot of what is done through a month end process is the same thing month on month and the finance guys are kind of becoming data entry clerks because they're just running the same process again and again and automating um you know and this is kind of then jumping into the you know robotic process automation field as well automating those tasks so that the finance guy the finance guy who has spent you know five years training to be an accountant right very smart guy very highly skilled they don't want to be a data entry clerk they want to be analyzing the business understanding the business making recommendations of you know of how to innovate and how to improve the business and

restructure it to improve performance and improve profit profitability and reduce costs so let the you know let the machine learning and the AI do the repetitive stuff and the business, therefore, do the innovation and do the analysis that then provokes innovation as well. And that's where the interest is, innovation. I know, yeah. And I think, probably like yourself, I just love this technology, really. I mean, it's something that I still get massively excited by,

by bringing some data in, analysing it, stretching things to the maximum and so on, and to the point, probably like you, that you do it in your spare time and in your weekends and evenings as well. It's just something you find very interesting, really. So it's been great speaking to you. I've noticed my children are back now from being out. I can hear music playing downstairs and uh that sort of thing so i think it probably says to me it's time to uh yeah my twins are in bed yeah yours are in bed yeah so it's been great speaking

to you how do people just very quickly how do people find out about cubics and yourself on twitter and that sort of thing just a few kind of like pointers to that well our um the the first and easiest place to to find us is on our is actually on our website cubics.com um there is a large large library um on there i mean it's one of the things we've we've really invested heavily in is making sure that everything all of our offerings in terms of what we do how we do things uh you know how how we operate um are on our website um i'm also um very active on linkedin as well um so you'll always you'll always find me on there i've got my own personal blog on matthewark.com as well um personal blog because you know sometimes i i have a view on things that is a little bit less corporate but i want i want to share uh you know um you know roughly you know kind of the same as

same same as you're you're doing here um and uh active on twitter as well and uh you know i think in fact i've just logged in what's my twitter handle cubix matt excellent excellent and and quite a few people from cubix will be at k scope i think quite shortly as well i think quite a few of your team are going over there. Yeah, yeah, that's right. We've got a team going over there. We've got a, we're sort of just hoping to get one of our customers over there as well to co-present with us on Oracle Analytics Cloud

and, you know, hopefully give us a, you know, talk to everybody there about uh about the journey they've been to been through going from uh sort of you know bunch of legacy on-premises stuff uh into Oracle Analytics cloud so uh yeah we're pretty excited about that excellent and I think having to persuade my wife that I was going to uh to Disneyland purely for a conference was that was a tough one really but it should mean that your customers will be quite keen to go there. So,

uh, so Matt, it's been excellent speaking to you. Um, thank you very much for doing the, uh, the interview and, um,

yeah, hopefully see you soon. Another event. Brilliant. Yeah. Look forward to seeing you again soon, Mark. And,

uh, thank you for, uh, thank you for having me on your podcast. Thank you.