>>Yeah, yeah. >>Hello and welcome back. You're just in time for our third session spot. I Q amplify your insights with AI in this session will explore how AI gets you to the why of your data capturing changes and trends in the moment they happen. >>You'll >>start to understand how you can transform your data culture by making it easier for analysts to enable business users to consume insights in real time. >>You >>might think this all sounds too good to be true. Well, since seeing is believing, we're joined by thought spots. Vika Scrotum, senior product manager. Anak Shaped Mirror, principal product manager to walk you through all of this on MAWR. Over to you actually, >>Thank you. Wanna Hello, everyone. Welcome to the session. I am Action Hera, together with my colleague because today we will talk to you about how spot I Q uses a. I to generate meaningful insights for the users Before we dwell into that. Let's see why this is becoming so important. Your business and your data is growing and moving faster than ever. Data is considered the new oil Howard. Only those will benefit who can extract value of it. The data used in most of your organization's is just the tip of the iceberg beneath the tip of the iceberg. What you don't see or what you don't know to ask. That makes the difference in this data driven world. Let's learn how one can extract maximum value of the data to make smarter business decisions. We believe that analytics should require less input while producing more output with higher quality in a traditional approach. To be honest, users generally depend on somebody else to create data models, complex data queries to get answers to their pre anticipated questions. But solution like hot spot business users already have a Google like experience where they can just go and get answers to their questions. Now, if you look at other consumer applications, there are multiple of recommendation engines which are out there, which keep recommending. Which article should I read next? Which product should I buy? Which movie should I watch in a way, helping me optimized? Where should I focus my time on in a Similarly in analytics, as your data is growing, solutions must help users uncovered insights to questions which they may not ask, we believe, and a I automated insights will help users unleash the full potential off their data Across the spectrum, we see a potential in a smart, AI driven solution toe autonomously. Monitor your data and feed in relevant insights when you need them, much like a self driving car navigates our users safely to their desired destination. With this, yeah, I'm happy to introduce you to spot like you are a driven insights engine at scale, which will help you get full potential off your data like you automatically discovers, personalize and drive insights hidden in your data. So whenever you search to create answers, spot that you continues to ask a lot more questions on your behalf as it keeps drilling and related date dimensions and measures employed insights which may be of interest to you. Now you as a user can continue to ask your questions or can dig deeper into the inside, provided by spotted you Spartak. You also provides a comprehensive set of insights, which helps user get answers to their advance business questions. In a few clicks, so spotted it. You can help you detect any outlier, for example, spot that you can not only tell you which seller has the highest returns than others, but also which product that sellers selling has higher returns than other products. Or, like you can quickly detect any trends in your data and help us answer questions like how my account sign ups are trending after my targeted campaign is over. I can quickly use for, like, toe get unanswered how my open pipeline is related to my bookings amount and what's the like there. What it means is that how much time a lead will take to convert into a deal I can use partake. You, too, create multiple clusters off my all my customer base and then get answers to questions that which customer segment is buying which particular brand and what are the attributes last and the most used feature Key drivers of change spotted you helps you get answer to a question. What factors lead to the change in sales off a store in 2020 as compared to 2019? We can do all this and simple fix. That's barbecue. What is so unique about Spartak? You how it works hand in hand with our search experience, the more you search, the smarter. The spot that you get as it keeps learning from your usage behavior on generates relevant insights for you for your users. Spartak. You ensures that users can trust every insights. A generator. It broadly does this and broadly, two ways. It keeps their insights relevant by learning the underlying data model on. By incorporating the users feedback that is, users can provide feedback to the spot I Q similar to any social media back from, they can like watching sites they find useful on dislike. What insights Do not find it useful based on users. Feedback Spot like you can downgrade any insight if the users have not find it useful. In addition to that, users can dig deep into any Spartak you insight on all calculations behind it are available for a user to look and understand. The transparency in these calculations not only increases the analytical trust among the users, but also help them learn how they can use the search bar to do much more. I'm super excited to announce Partake you is now available on embrace so our automated A insights engine can run queries life and in database on these datasets so you do not need to bring your data to thoughts about as you connect your data sources. Touch Part performs full indexing value to the data you have selected, not just the headers in the material and as you run sport in Q, it optimizes and run efficient queries on your data warehouse on. I am super pleased to introduce you. This new spot like you monitor the spot that you monitor will enable all your users to keep track of their key metrics. Spartak, you monitor will not only provide them regular updates off their key metrics, but we also analyze all the underlying data on related dimensions to help them explain. What is leading to the change of a particular metric monitor will also be available on your mobile app so that you can keep track of your metrics whenever and wherever you go, because will talk for further detail about this during the demo. So now let's see Spartak in action. But before we go there, let's meet any. Amy is an analyst at a global retail about form. Amy is preparing for her quarterly sales review meeting with the management, so Amy has to report how the sales has meat performing how, what, what factors lead to the change in the sales? And if there are any other impressing insights, which everyone should off tell to the management? So but this Let's see how immigrant use part like you to prepare for the meeting. So Amy goes to that spot, chooses the sales data set for her company. But before we see how many users what I Q to prepare for the meeting. I just wanted to highlight that all this data which we're going to talk about is residing in Snowflake. >>So >>Touch Part is going to do a life query on the snowflake database on even spot. A Q analysis will run on the Snowflake databases, so we'll go back and see how you can use it. So Amy is preparing for the sales meeting for 2019. We just ended. So images right Sales 2019 on here. She has the graph of the Continent tickets, >>so >>what she does is immediately pence it >>for >>the report. She's creating Andi now. This graph is available >>there now. >>Any Monnet observed >>that >>the Q four sales is significantly higher than Q >>three, so >>you she wants to deep dive into this. So she just select these two data points and does the right click and runs particularities. So now, as we talked earlier, Spartak, you recommends which columns Spartak Things Will best explains this change >>on. >>Not only that, you can look that Spartacus automatically understood that Amy is trying toe identify what led to this change. So the change analysis we selected So now with this, >>Amy >>has a bit more business context when he realizes that she doesn't want to add these columns. So she's been using because she thinks this is too granular for the management right now. >>If >>she wants, she can add even more columns. All columns are available for her, and she can reduce columns. So now she runs 42 analysis. So while this product Unisys is running, what the system will do with the background, this part I Q will drill across all the dimensions, which any is selected and try to explain the difference, which is approximately $10 million in sales. So let's see if Amy's report is ready. Yeah, so with this, what's product you has done is protect you has drilled across all dimensions. Amy has selected and presented how the different values in these dimensions have changed. So it's product. You will not only tell you which values in these dimensions have changed the most, but also does an attribution that how much of this change has led to the overall change scenes. So here in the first inside sport accuse telling that 10 products have the largest change out of the 3 45 values and the account for 39% increase. Overall, there has been look by the prototype category. It's saying that five product types of the largest change out of the 15 values, and they account for 98.6% of total increase. And they're not saying the sailors increased their also demonstrating that in some categories the sales has actually decreased to ensure the sales has decreased. Amy finds this inside should be super useful so immediately pins this on the same pain, but she was preparing for and she's getting ready with that. Amy also wants to dig deeper into this inside. My name goes here. She sees that spot. I Q has not only calculated the change across these product types, but has also calculated person did change. So Amy immediately sorts this by wasn't did change. And then she notices that even though Sweater as a category as a prototype, was not appearing in the change analysis but has the most significant change in terms of percentage in comparison to Q two vs Q four. So she also wants to do this so she can just quickly change the title. And she can pin this insight as well under spin board for the management to look at with this done. Now, Amy, just want to go back to this sales and see if she can find anything else interesting. So now Amy has already figured out the possible causes. What led to the increase in sales? So now, for the whole of 2019, as this is also your closing, Amy looks, uh, the monthly figures for 2019, and she gets this craft now. If Amy has to understand, if there is an interesting insight, she can dig into different dimensions and figure out on her own or immigrant, just click on this product analysis. That's product immediately suggest all the dimensions and measures immigrant analyze sales by Andi many. We will run this What will happen is this barbecue system will try to identify outliers. The different trend analysis Onda cross correlation across different measures. So Amy again realizes that this is a bit too much for her. So she reduces some of these insights, which she thinks are not required for the management right now from the business context and the business meeting. And then she just immediately runs this analysis. So now, with this, Amy is hoping to get some interesting insights from Spartak, which immigrant present to her management meeting. Let's see what sport gets for her. So now the Alice is run within 10 seconds, so spot taken started analyzing. So these are the six anomaly sport like you found across different products, where their total sales are higher than the rest. He also founded Spot. I just found eight insights off different product types which has tired total sales and look across these enemy sees that oh jackets have against the highest sales across all the categories in December as well. Amy wants toe been this to the PIN board on M. It moves further now. Amy's is that it has also shown Total Country purchased their product a me thinks this is not a useful insights. Amy can get this feedback. The system and system asked, Why are you saying you don't find this useful so the system can remember? So you can also say that anomalies are obvious right now and give this feedback and the system will remember. In addition, Amy finds that the system has automatically correlated the total sales in total contrary purchase. Amy Pence this as well to the pin board. Andi. She loves this inside where she she is that not only the total sales have increased, but total quantity purchases have increased a lot more on their training, opposed as well. So she also opens this now anything. She is ready for her meeting with the management. So she just goes and shares the PIN board, which she just created with the management. And you know what happens immediately? The jacket sales category Manager Mr Tom replies back to Amy and says in the request, Any d really like this? So now we will see how Spartak you can help any educators as request doesn't mean really need to create these kind of reports every month to cater toe Tom's request. So with this, I will handle it because to take us walk us through How spot that you can cater this request. Hi, >>everyone. So analysts like Amy are always flooded with such requests from the business users and with Spot and you monitor. Amy can set up everyone who needs updates on a on a metric in just a few simple steps and enable them to drag these metrics whenever and wherever they want. And north of the metrics, they also get the corresponding change analysis on the device off their choice with hot Spot. What I give money being available on both Web and the mobile labs. So let's get started with the demo will be set up a meet and go to the search tab and creator times we start for the metrics you want to monitor, right? And please know if the charges already created is already created. All is available is, um, usually a section in a PIN board. Also dancer. Then there's no need to create a new child. She can simply then uh, right click on the chart and select moisture from the menu, which then shows, which then shows the breakdown off the metric he's going to monitor, including the measure. What it's been grouped by on what it is filtered on. Okay, and also as this is a weekly metric, all the subscribers are going to get a weekly notification for this metric had been a monthly metric. Then the notifications would have been delivered on a monthly cadence. Next she can click on, continue and go to the configure dimensions called on Page. Here A is recommending what all dimensions could best being the change in this metric, she can go ahead with default recommendation, or she can change the columns as she seems very she can click, she conflict, continue and go to the next page, which is the subscriber stage. It is added by default to the subscriber, but she can search everyone who needs update on this metric and add them on this metric by clicking confirmed, she'll see a toast message on the bottom of the page, taking on which will take a me to this page, which is a metric detail page On the top of this page, we can see the movement of the metric and how it is changing over time, 92 you can see that the Mets jacket, since number has increased by 2.5% in the week off 23rd of December has compared toa the week off 16th of December and just below e a has invaded the man is generated in sites which are readily available for consumption. Okay to discharge. Right here says that pain products have the largest change out of all the 28 values and contributes to the 88% of the total increase in the same. And this one right here is that Midwest is the larger Midwest has the largest change and accounts for 55.66% off the total increase. Now, all this goodness is also available on the mobile lab. Right? So let me just show you how business users are going to get notified on the based. On this metric, all the business users who are subscribed to this metric are going to get a regular email as well as push notifications on the mobile lab. And when the click on this, they line on a metric detail page which has all the starts, which I just showed you on the on the bed version, okay. And one cyclic on back burden. They land on this page, which is a monitor tab, and it summarizes all the metrics Which opportunity monitoring and gives them a whole gave you to stay all I want to stay on top of their businesses. Okay. Eso that folks was monitor. Now I'll search back to slaves and cover. Summarize the key takeaways. From what? That she and I just don't know. So it's part of you wanted, uh, Summit Spartak you. It automatically discovers insights and helps you unless the full potential of your data and that's what I do is comprehensive set off analysis. You can answer your advanced business question in just a few simple steps and the end speed of your time. Bring state. And with a new support for embrace, you can run sport like you on your data in your data warehouse and with spotted you monitor, you can monitor all the business metrics and not just died. We can also understand that teaching teaching drivers on those metrics on the platform of your choice. So with that, I'll hand over toe, you know. >>Thank you so much. Both of you That was fantastic. Um, I just love spot like, because it makes me look like much more of a rock star with data than I really am. So thank you guys for that fantastic presentation. Um, so we've got a couple of minutes for a couple of questions for you. The first one is for action. Um, once spot I Q generates a number of insights. Can you run spot I Q again on one of those insights? >>Yeah, As a philosophy off Spiric, you sport like you never takes the user to the dead end Spartak. You also transparently shares the calculation. So user can not only the keeper that on edit Understand how this product you inside has been calculated, but user can also run us for like you analysts is honest for data analysis as well. Which music? And continue to do not on the first level. Second level in the third level as well. >>That's cool. Thank you. Actually on then The next one is for because for spot ik monitor is it possible to edit the dimensions used for explaining the factors to change that was detected? >>Yes. It's an owner of the metric you can change the dimensions whenever you want and save them for everyone else. >>Okay, well, I think that's about all we've got time for in this session. So all that remains is for me to say a huge thank you to Because an Akshay Andi, we've got the last session of this track coming up in a few minutes. So grab a snack. Come right back and listen to an amazing customer story with Snowflake on Western Union, they're up next.

everybody welcome to this digital coverage of the verdict of big data conference you're watching the cube and my name is Dave Galante it's my pleasure to invite in Ben white who's the senior database engineer at Domo been great to see you man thanks for coming on great to be here and here you know as I said you know earlier when we were off camera I really was hoping I could meet you face to face and in Boston this year but hey I'll take it and you know our community really wants to hear from experts like yourself but let's start with with domo is the company share with us what Domo does and what your role is there well if Parker can go straight to the official what Domo does is we provide we process data at bi to scale with we provide VI leverage a cloud scale in record time and so what that means is that you know we are a business operating system where we provide a number of analytical abilities to companies of all sizes but we do that at cloud scale and so I think that difference is quite a bit so a lot of your work if I understand it and just in terms of understanding with Domo does--is there's a lot of pressure in terms of being real-time it's not like you sometimes don't know what's coming at you so it's AD Hoch I wonder if you could sort of talk about that confirm that and maybe add a little color to it yeah absolutely absolutely that's probably the biggest challenge it is to being the operating Domo is that it is an ad hoc environment and certainly what that means is that you've got analysts and executives that are able to submit their own queries without very with very few limitations so from an engineering standpoint the challenge in that of course is that you don't have this predictable dashboard to plan for when it comes to performance planning and so it definitely presents some challenges for us that we've done some pretty unique things I think to address those right sounds like your background fits well with that I understand here if people have called you a database whisperer and an envelope pusher what does that mean to do a DBA in this in this day and age well the whisperer part is probably a lost art in the sense that it's not really sustainable right the idea that you know whatever it is I'm able to do with the database it has to be repeatable and so that's really what analytics comes in right and that's where pushing the envelope comes in in a little right away that's what vertical comes in with this open architecture and so as a person who has a reputation for saying I understand this is what our limitations should be but I think we can do more having a platform like vertical is such an open architecture kinda lets you push those limits by the bit I mean I've always felt like you know vertical when I first saw the Stonebreaker architecture and doctors some of the early founders I always felt like it was the Ferrari of databases certainly at the time and it sounds like you guys use it in that in that regard but talk a little bit more about how you use Vertica why in a ym ppy Vertica you know why why can't you do this with our DBMS educate us a little bit on some of the basics but for us it was part of what I mentioned when we start and we talked about the very nature of the demo platform where there's a an incredible amount of resiliency required and so Vertica the NPP platform of course allows us to build individual database clusters that can perform best for the workload that may be assigned to them so the the open the expandable the the the ability to grow vertically as your base grow those are all important factors when you're losing early on right without a real idea of how growth would be or what it would look like if you were kind of doing that something to the dark you looked at the vertical platforming you can see well as I grow I can kind of feel with this right I can do some some unique things with the platform in terms of this poking architecture that will allow me to not have to make all my decisions today right about Harlem so you're using Vertica I know at least in part you you working with AWS as well can you describe sort of your environment that you give anything on Prem is everything in the cloud what's your setup sure we have a hybrid cloud environment where we have a significant presence in public files in our own private cloud and so yeah having said that we certainly have a really an extensive presence I will say an AWS and so they're definitely the partner of our when it comes to providing the databases the server power that we need to operator but from the standpoint of engineering and architecting a database what was some of the challenges that you faced when you had to create that hybrid architecture what did you face and how did you overcome that well you know some of the there are some things we need faced in terms of wine and made it easy that Vertica and AWS have their own they play well together we'll say that and so vertical is designed to reprise I'm gonna AWS and so that part of it the care of itself not our own private cloud and being able to connect that because our public clouds has been a part of our own engineering ability and again I don't want to make a little light of it it's certainly not impossible and so we've some of the challenges though this pertains to the database really were in their early days that you mentioned when we talked a little bit earlier about marathas most recent Eon mode and I'm sure you'll get to that but when I think of our early challenges some of the early challenges were the architecture of enterprise mode when I talk about all of these this idea that we could have unique databases or database clusters of different sizes so this elasticity that's really if you know that the enterprise architecture that's not necessarily dandified architecture so we added this Munich things I think to overcome that right early to get around the rigidness though enterprise yeah I mean I hear you right Enterprise is complex and and you like when things are hardened and fossilized but in your ad hoc environment that's not what you needed so talking more about Aeon mode what what is e on mode for you and how do you apply it what are some of the challenges and opportunities there that you found um so the opportunities were certainly in its elastic architecture the ability to separate the storage immediately meant that for some of the unique data paths that we wanted to take right we could do that fairly quickly certainly we could expand databases right quickly but more importantly now you could reduce because previously in the past right when I mention the Enterprise Architect with the idea of growing a database in itself has its pain right as far as the time it takes to speed the data in that but to read to then think about taking that database back down no Innova though all of us under the eon right you had this elasticity where you could kind of start to think about auto scaling where you go up and down and maybe used to save some money or maybe you could improve performance or maybe in needham and at a time when the customers needed most in a real way right so it was definitely a game in that regard I always have to talk to the customers because I get to you know I hear from the vendor what they say and I think they sort of validate it so you know Vertica talks a lot about separating compute and storage they're not the only one from an architectural standpoint to do that but Vertica stresses that they're the only one that does that with a hybrid architecture they can do it off ram they can do it in the cloud from your experience well first of all is it true you may or may not know it is that advantageous to you and if so why well first of all it's certainly true earlier in some of the original beta ethnic for the arm prim GI mode stuff we I was able to participate in it and be aware of it so it's certainly a reality day I'm it's actually supported on pure spirit with flash played and it's time quite impressive you know for who who that who that will be for tough one a Spartacus question that they're probably still answering but I think obviously some enterprise users that probably have some hybrid cloud right they have some architecture they have some hardware that their sales want to make you so we certainly would probably fit into one of their you know their market segments that they would say we might be the wants to look at on pram er mo begin the the beauty of it is the elasticity right that the idea that you could have this and so a lot of times so I want to go back real quick to separating them and you know we start by separating it and I like to think of it maybe more as like decoupling because a new in a true way it's not necessary separated there's ultimately you bring the compute and the doors back together but to be able to typically couple it quickly replace knows bring in those that's certainly fits I think what we were trying to do in building this Emma I'll me let the ecosystem that could respond to a unknown or of a customer demand I see thank you for that clarification because you're right it's really not separating its decoupling in it that's important because you can scale them independently but you still need compute and you still need storage to run you your workloads but from a cost standpoint you're not to buy it in in chunks you can you can't buy granular segments for whatever your workload requires is that is that the correct understanding yeah and to be able to the ability to be able to reuse compute throw it in a scenario of AWS or even in the scenario your on-prem solution you've got this data that's safest here and ask for your in your storage but then the compute that you have you can reuse that right you could have a scenario that you have some query that needs more analytic more firepower more memory more what have you that you haven't so you can kind of move to the next important right that's maybe more important then and I grow them separately can I can I borrow it can I borrow that computer use for my perfect give it back type of thing and you can do that when you're so easily a couple different ooh all right and likewise if you have a down period where customers aren't using it you'd like to be able to not use that if you no longer require if you'd like to give it back go in it open the door to a lot of those things that allow performance and cross the spark to meet up we're going to ask you a question winsome pure a couple times are you using pure flash blade on-prem is that correct that is the solution that is supported that is supported by Vertica for the on print so at this point we were we have been discuss with them about some our own PLC's for that time before again we back to the idea of how do we see ourselves using it and so we've certainly discussed the feasibility of bringing it in and give it a job but that's not something we're Oh happily all right now then what is Domo for Domo tell us about that we really started this this idea even in the company where we say you know we should be using Domo in our everyday business the sales folks the marketing folks right everybody we're gonna use Domo it's a business platform for us in the engineering team it was kind of like well if we use Domo say for instance to be better at the database engineers now we've pointed Domo edits tell fried verdict is running Domo in the background for some degree and then we turn around and say hey Domo how can we better at running you and so it became this kind of cool thing we played with where we're now able to put some dumb methods together where we can actually do their eye we can monitor using our platform it's really good at processing large amounts of data and spitting out useful analytics right we take those analytics out make recommendation changes that the day so now you've got still more for Domo happening it allows us to sit at home and and work now even when we have to even before we had to well you know look look at us here right it couldn't mean in Boston physically we're now meeting remote you're you're on a hot spot because you got some weather and your satellite internet and in Atlanta and we're having a great conversation so so we're here with with Ben white who's the senior database engineer at Domo I want to ask you about some of the envelope-pushing that you've done around autonomous you hear that that word thrown around a lot means a lot of things to a lot of different people how do you look at autonomous and how does it fit with Eon and some of the other things that you're doing you know I'm a tall amidst the idea of economy is something that I don't even know that I'm I have already ready to define and so even in my discussion I often mention it as a road to it exactly where it is it's hard to pin down because there's always this idea how much trust do you give right to the system or how much how much is truly autonomous how much authority is being intervened by us the engineers so I do hate on using it but on this road towards autonomy when we look at what would how we're using Domo and even what that really means to vertical because in a lot of my examples and a lot of the things that we've engineered a demo work designs maybe over something I thought was a limitation day and so many times Oh as we've done that verdict is kind of met us like right after we've kind of engineered our architecture stuff than we thought it felt on our side Vertica has some released it kinda addresses it so the autonomy idea and the idea that we could analyzed metadata make recommendations and then execute those recommendations without intervention is that road to autonomy and once the databases start able to do that you can see in our ad-hoc environment how that would be pretty pretty useful where with literally millions of queries every hour trying to figure out what's the best you know probably for years I felt like I I T folks sometimes we really did not want that automation they wanted the knobs to turn but but I wonder if you comment I feel as though the level of complexity now with cloud with on-prem with you know hybrid multi clouds the scale the speed the real-time it just gets the pace is just too much for for humans and so it's almost like you know the industries is gonna have to capitulate to the Machine and then really trust the machine but I'm sitting I'm still sensing from you a little bit of hesitation there but light at the end of the tunnel I wonder if you could comment sure I think that in the light of the tunnel is even in recent months in recent we've really began incorporating more machine learning in artificial intelligence to the model right and back to where we're saying it so I do feel they were getting close for too finding conditions that we don't know about because right now our system is kind of a rule rules based system where we've said well these are the things that we should be looking for these are the things that we think are a problem to mature to the point where the database is recognized and anomalies and taken on at imagining saying these are problems you didn't know happen and that's kind of the next step right identifying the things you didn't know and that's where that's the path we're on now and that's probably more exciting even then kind of nailing down all the things you think you know and to figure out what we don't know yet so I want to close with I know you're a prominent member of the respected member of the Vertica a customer advisory board you know without divulging anything confidential to me what are the kinds of things that you want Vertica to do going forward I think some of the end a in database autonomy the ability to take some of the recommendations that we know we can derive from the metadata that already exists in the platform and start to execute some of the recommendation another thing we talk about and I'm gonna pretty open about talking to it is talking about it is the new version of the database designer I think it's something that I'm sure they're working on lightweight something that can give us that's database design without the overhead those are two things I think as they nail or particularly the database designer as they respect that they'll really have all the components in place to do in based economy and I think that's just some victory where they're headed yeah nice well Ben listen I really appreciate you coming on your a thought leader be very open open-minded verdict is you know really open community I mean they've always been quite transparent in terms of where they're going it's just awesome to have guys like you on the cube to share with our community so thank you so much and hopefully we can meet face to face currently absolutely will you stay safe in Boston I'm one of my favorite towns and so no doubt when this when the doors get back open I'll be from coming down or coming I'm gonna do work take care all right and thank you for watching everybody Villante with a cube we're here covering the virtual Vertica of big data conference you [Music]