DataSelf for Manufacturing with Natural Language

Alrighty. Hello everyone. Thanks for joining us. My name is Johnny Giorardi. I'm Data Sales founder and CEO. First of all, thank you for making time today. I hope you're staying safe, well and sane, which is my biggest problem, unfortunately. Today we're going to be talking about data cell for manufacturing companies. We'll talk about several things. Major key performance indicators for manufacturing companies, natural language, AI, machine learning and analytics. I have with me also Clark Walliser. Clark Walliser has been in the manufacturing industry for a long time and he's going to give a quick introduction about his career around manufacturing initiatives. And then throughout the webinar he's going to cover three different use cases that we have helped clients with manufacturing analytics. So Clark, would you please put the webcam on and introduce yourself and give us the first use case.

Greetings. My name is Clark Walliser. As you know, most of my career I've worked for about 30 years as a primarily a Sage consultant with a focus on manufacturing. About eight years ago I went to work for an end user who manufacturing company. And so I got to be on the other side of the coin, the end user implementing the manufacturing first with Sage, then with Epicor, then I went to another manufacturing company as an end user also and that was using Microsoft Dynamics 365 for manufacturing. So I've gotten a fair broad background on the manufacturing and the kinds of issues you deal with. Finally I ended up at dataself and with the business intelligence side of it and now pulling together the manufacturing data into a usable format and medium that you can use to help make decisions in a timely and accurate manner. That's pretty much it right there.

Perfect. Would you go for the use case please first one sure.

Use case one. We've got three use cases we're using here. These are actual customers. I can't use their names, we haven't gotten permission yet. But they are real customers. The first case, pretty general. They have an ERP system and it has all the manufacturing functionality, et cetera. And they were trying to use that data to build their own dashboards directly from the ERP system using Crystal Reports reporting services. But it was clunky, it would not run in a timely manner. They would be pulling data from multiple tables, pulling it all together and the report might take hours to process. And it was very complex to do all the table linking, etc. They chose to go with data self and here's the reasons why. It simplifies the whole process. The data warehouse pulls all the data together into one TABLE so they can pull everything from one table for their dashboards. They can use the data warehouse to refine their data to include only qualified data. Sometimes they have sales orders that are in the system that they don't want to include or purchase orders or work orders that they certain ones they include exclude. This gave them the ability to customize that and so they could clean their data before they reported on it. Then they would use, then they would use the user interface tools. You can use either Power bi, Excel or Tableau to pull this data out and they would get a clean report in moments instead of hours. So they're not doing it. There's nothing fancy there. It's simply getting the data they need when they need it. They had tried this before in crystal reports and reporting services as I said, but those are only report writers. They didn't have the ability to store to pre build the data source into one data set. And so this they have also taken the time to learn how to use their tools. So they are self maintaining. They're not dependent on us for any of the daily items. The result of that they get the data they need when they need it to make decisions to keep from. To keep from stopping operations.

All right Clark. Yep, thank you very much. I remember that at least one of the parts that why from taking hours it went down to seconds to run reports is I think the bill of materials was one of the components. They have a multi layer bill of materials and I think they were using either crystal or reporting services in running through all the possible ways to combine. The bill of materials was a major problem for their prior system. And when we use data warehousing and tools like Tableau we have the right dials to take complex data manipulation processes and turn them in seconds instead of taking so long. Anyhow, Clark appreciate for the first use case I'm going to go now into the more overview of how data self adds value to manufacturing companies. So the first good news is manufacturing companies, they have a lot of data being captured to run their business. Like Clark mentioned, if you are in this webinar probably is because you come from one of these different ERP and MRP systems. You know you use for your traditional ERP in your financials, your sales and whatnot. But also you might be using the MRP components or an add on to the system to run your manufacturing processes. But typically to run your business you also use other data sources like Excel. Very popular to use Excel to do a lot of the data analysis of your manufacturing processes or your Budgeting your production orders. There's a lot of things that people do in Excel, or maybe you're more like the Google sheets instead. You don't like Excel and Microsoft doesn't matter. You know, there are different tools people use to do their data silo manipulation, but likely you may also use CRM like you know, Salesforce, HubSpot, other systems. You might have E Commerce, you might have Google Analytics, you might have payroll. Overall we have, we have seen that companies usually use more than 20 different data silos to manage their business. More than 20 on average. Data self we are a fairly small business and we use more than 20. So Google Analytics, Excel sheets, project management is all over the place. We call this the data house divided. You know, if you're trying to understand what's happening with your business, what your major KPIs are, typically the information is all over the place. Typically you have to know where to go, you have to know the credentials, you know to, you have to know how to run reports. If you need to customize reports, it becomes complex. So we hear over and over and over from people coming to us that they are flying blind. You know, management don't have, doesn't have visibility into what's going on in a timely fashion. Sometimes when they get to know about something, it's been days after weeks and even sometimes months later. So today in this crazy world we live in, flying blind is something that people are trying to get out of and make better, more informed decisions in a timely fashion. Typically what companies do is to use what we call Excel. Hell, Excel is a great tool. I'm not talking specific about Excel, but people use tools like Excel reporting services, crystal or even power, bi and tableau directly connected to their data sources and eventually becomes a little monster. You know, all of these different places, nobody knows exactly when the data was refreshed. Is that the way you're calculating your metrics in one report? The same way as calculating another report built by someone else? It becomes very complex if you don't have a centralized way to manage your KP or Q performance indicators, to have a single version of the truth. So everyone will be using the same standards. So it's very hard. And when you're using these tools and using this direct approach like most people do, it becomes a labor intensive process. Many people, they have hours and hours and hours every day or week or month to get the reports properly calculated, properly vetted and distributed. So it's hard, it's labor intensive, is in error prone. So these flying blind Excel hell and labor intensive is a very high level problems. We hear all the time in the way we solve or the way we recommend you to solve. Even if you don't do use data self is by using data warehousing technology. So what I'm showing you now is going to be showing you the data self way to fix these problems. But you do not need to use data Self. These concepts as a whole can help you if you have the resources to build your own data warehouse from scratch, you know, absolutely you should assess, should you build yourself, should you get someone that to build you data warehouse, should you go with data self, Those are your questions to solve. But with a data warehouse, it's an easy and established platform platform to bring initially maybe your MRP data. You only want to tackle your manufacturing process at first because it's painfully right now to do analysis. But eventually you want to go beyond manufacturing. You want to go to core financials, you want to go to sales, you want to go to payroll, you want to go to Google Analytics. No matter what data warehousing is, an established technology can bring all of your data silos together. You can glue the pieces in the data warehouse. Building and maintaining a data warehouse requires proper expertise. You know, if you have people, good people have experience with data source data warehousing, they can, you know, automate a lot of these tasks that usually people do manually in, I don't know, Excel increase and whatnot. And then once you do the data warehousing consolidation, then you can plug, plug different reporting tools, Tableau, Power, bi, Excel, reporting services, Crystal, whatever it is. And now reporting would be easier, faster, everything coming from a single place where you define the truth. How do you calculate your overtime, how you calculate your cost of goods, how you calculate your gross profit? You define those metrics in your data warehouse. Now when you pull the reports and the metrics, all of them will follow the same version of the truth. So an architecture using a data warehousing is usually a big component to tackle a lot of complex reporting needs. And if you use tools like Tableau and Power bi, which we embrace and we support, these tools are like the gold standard when it comes to reports and dashboards. Easier, faster, more empowering for your end users. In anytime, anywhere, you can find the reports and KPIs using desktop applications, web browsers and mobile devices. Very empowering. Finally, from a data setup standpoint, and we're going to be showing some to you, we have built in our 15 years of experience more than 5,000 report dashboard and KPIs templates. So if you decide the Data self is the right fit for your data warehousing and analytics, you don't start with a blank sheet. You start with a very rich library, reports and dashboards to help you get to focus on data analysis quickly and easily. And then you can customize it as you wish. So this is an architecture that is usually very helpful to tackle complex manufacturing reporting needs. With that in mind, I'm going to open another section for Clark to give the second use case. Clark?

Yes, in this case the customer, the company, their needs were a little bit more than just simply getting the data out of the ERP and displaying it the way they wanted. Their, their production was driven by sales order demand and sales projections. And so that was a, that was a driver to make sure that the raw material availability and sub assembly availability was there. They wanted to do just in time, inventory and eliminate any shortages or overstocks. Their MRP system could do that but it wasn't always exactly, the data wasn't always exactly how they wanted it. So that became a difficulty and time it was recreating it each time took time and then sometimes that was too late, it was no longer valid data. So what data self allowed them to do with having a data warehouse? They could customize which sales orders qualified for demand, which purchase orders qualified for supply, which work purchase orders or work orders, et cetera could be factored in. And so they got a clean single source data set without having to do any extra work. And then they could do the reporting off of that cleaned and validated data source with all of its built in calculations. Again they didn't have to recreate the data set with links and et cetera is right there and is refreshed every day. The buyer had his pre made reports and so he could run those standard reports but he also had the ability to do has the ability to do ad hoc queries pulling from the same data and the ad hoc queries have their have built into them all of the formulas and calculations they would normally use so that they have consistency there. This enabled them to select only their qualified data and get the information quickly in moments instead of hours, which was important to them given the timeliness of it.

All right, good stuff Clark. I think also one of the messages from this story is a very common scenario with mid size manufacturing companies. You know they design the manufacturing process, you know the MRP process around a certain production flow. But so many times things change, they need to tweak things and then the MRP doesn't provide a lot of flexibility to change things on the fly. But if you have a flexible reporting tool that you can quickly play with the numbers and see if you change this, what happens. So we're seeing that BI is a good complement for manufacturing companies because it gives you extra flexibility to analyze your things, even though if your manufacturing processes in the MRP system would not allow those tweaks. So I think this was one of the values that this particular manufacturer was using Data Cell for flexibility of data analysis. The other one was more around performance, this one more about flexibility of making tweaks in their processes. All right, thank you, Clark. Moving along. So now we're going to be jumping into the Data Self demo. The idea of this demo is to give you some insights into KPIs reports and dashboards that can be valuable to you regardless of Data Self. So if you want to build your own, please. But we'll also show the Data Self framework as a whole. I'm going to be giving an idea of how we do the Data House United instead of Divided. Then I'll show some of the KPIs out of the box for manufacturing. I'll show natural language, how we're using natural language, or how we're empowering manufacturing companies in the natural language framework, AI and machine learning. I'll show some examples of customizations on the front end, on the back end, and also how people usually use this framework for report distribution. Having said that, now I'm switching to our Data Self portal and all the things I'm going to be showing the next 15 minutes or so can be embedded inside of your own applications. So if you have an MRP, if you have an ERP. Typically today most ERP and EMRPs, you can embed third party reports and dashboards in them. And this could be completely immersed in those applications. But the first thing is the Data House Divided versus the Datahouse United. Once you start bringing your data together in the data warehouse, you can use tools. And right now I'm showing Tableau as a tool and we have Tableau as an OEM component in Data Self, where when the user logs to the portal, in this case a tableau portal, depending on the user's credentials, you have access to all of your data silos. In this example, I'm showing an example where a user will have access to the ERP information. So there is sales information, there is inventory information. This user would also have access to the MRP component. So manufacturing is here. CRM could be from Salesforce could be from Sage CRM or could be dynamic CRM Akumatica CRM doesn't matter. E commerce, marketing tools, Google Analytics, more financial information and paychecks. So in this example I have access to all of these things. I just click on these buttons and I go and I can analyze those individual sections together in a single place. Think about the CEO of your company, the coo, the cfo, the managers. Giving them the ability to go to all of these data silos that they care for, reporting mostly not for operation. A single login on their mobile device and quickly seeing information is very powerful. We have many clients that once they saw this they got the vision to say where do I find the check? Usually it's not that easy. But getting a datahouse United is a big value for many manufacturing companies. Now because this webinar is focused on manufacturing, I'm going to be focusing only on this section of the data house United. And later I'm going to show you if you want to also learn some of these other things where to go on our website. You know we have last week or month we had a webinar around distribution. In distribution we focus on shipped sales, inventory management, financials and I think we also covered CRM. So if you're also interested in better analytics out of your distributions, out of your business as a manufacturer, watch that webinar to learn about all these other components. Today we're going to be focusing on the manufacturing only section and to get that started, typically what happens is manufacturer, they are selling goods, right? So typically this section here is to see what I have today on hand and what is available to ship. So all the rest is whatever you can ship today you'll be analyzing on this section. Well, what if you cannot ship because you don't have enough ready to be shipped? Well then you go into the manufacturing component and usually the first step is planning your manufacturing processes. So when I click the inventory planning button, it's going to go to our pre configured dashboard to get started to plan for manufacturing. This dashboard can be heavily customized and usually it's heavily customized. But let me show typically how we help with the inventory planning component. Assuming this would be a good fit for your needs, you have for every of your products you have a major block here and then you have for your period, you know, this would be the current period, you know, right now I'm looking on a weekly basis, this would be my current week. I can see how many of each of these items I have on hand today. Right. So all on hand Today then I can see what's on work order. So I don't have a lot of sample database here. It shows on work order coming inbound to my work order a bunch of items by week. Then I see what's already being sold and what is the promise day. So you can put here by week what's going out of the door. Then we plug sales projection. How do you do your sales projection? We usually check how you do it and we usually can automate the process or we can help you make it more effective. So once the sales projection is calculated, then it's automatically plugged for a week and then we do what is the own hand projected by week. So this week, at the end of the week, what's going to happen when I take my own hand plus what's coming in minus what's going out. This is my projection on hand per week. And then when you look at the minimum hand, then you have the excess or short if it's red is negative. So you have less than your quantity on hand. Minimum, you have black, you are above. And then you can see if you're behind or not. Your shipments we use here typically waste so people can easily control some of the dials. This is again highly customized. How you want to do your projection of sales, how you do your own hand projection calculation parameters, are you going to look at all the items or on your a kind of items, what warehouse you want to consider your analysis. So this is usually if the client doesn't have a framework for projection that they only just want to automate, sometimes an Excel file, sometimes they use R the language or to do projection we can incorporate those as well. If they don't have those or if they see this as a better either we go the way that they are doing and automating or we use some template like this to help with inventory planning. So first step manufacturing process is inventory planning. Then typically once you have your inventory planning, you have to see okay, do I have my bill of materials? So now I click the bill of materials tab and again, this is an example of a bill of materials dashboard where the top left shows me the finished goods, how many I have available to ship today and how many are required in my orders. And if I short meaning I have to produce more. Now in manufacturing, almost every single KPI or key performance indicator might be a discussion. For instance, what exactly is available to ship in some companies is what they have sitting as a finished good in their inventory. So we use that definition to define what goes in this column. In some businesses, what's on water is coming from overseas and if it's two weeks away from getting to the port, we already consider quantity available. No matter how you calculate your KPIs, we can customize the data warehouse and the templates to be sure these metrics are specific to use case. Once you define the bill of materials finished good and you had your quantity short, then we can go to the next section where you have the finished goods exploded into their bill of material levels. Out of the box we have support five level BOMs. So out of the box, if you have from one to five, we pretty much have to see how they are structured and the templates pretty much get the data and allows you to analyze the quantity required by each level based on the total order that you need. If you have more bill of material levels, we can customize or train new people on how you can create 6, 10, 15, you know as far as you want to go down in your bill of materials level. So overall this will be the session where you can take the finished goods and explode into each of the bill of material levels and see the quantities required. Then we have what we call the raw materials by by bom. So for every bom, we have what exactly we have available and what's required and eventually is the list of all raw materials. So if you need to buy raw materials, if you need to order them, this gives you the whole picture based on the initial finished good in all of these dashboards we have if you were to use it asl, they're highly interactive. If you come over here and see where do you use these three particular raw materials. When you select them, these two sections are going to change to show where those three products are being used in which bombs and how they are broken down by bill of material levels. So very interactive to quickly see the high level picture and go and slice and dice your data accordingly. So that's an example. And again you can choose everything you can choose by warehouse on water, which is fairly common for manufacturing companies to see things that are coming from overseas. So that's the bill of materials. Once you know what you need to buy so you can produce things in case you need to buy more things, then you go to purchase more things if you need to buy more things. And this is an example of a purchasing dashboard. A little out of the box you can see by product the quantities that you have been open to purchase order. Whatever is in green means you're supposed to be receiving today. Yellow is lightly late one day late Orange is more late, took seven days, red is very late and these are future po. And again, if you want to see what's your behind and you click the red, you only focus on the vendors and the products that are behind and all the details associated. So just this a quick way to see how you can quickly CPO and focus on areas that might be of a problem moving along. So once you know the purchase order, you might want to actually go into the work order and see how this is going to be distributed in your production processes, your production centers and whatnot. I'm not going to go into that dashboard, I want to show another one so we can cover other topics as well. I'm going to show now the manufacturing dashboard. Typically this dashboard is highly customized, so likely you would not have something like this specifically, even though some of the widgets could be of value for your manufacturing processes. But the idea of the manufacturing dashboard is to give upper management a view of what they care when it comes to the manufacturing processes. So in this case, in this dashboard we have units produced as the bars and we have the percentage loss by each product. So in a quick glance I can see some items, I produce a lot more others and which ones I'm having a lot of percentage losses. And maybe as a decision maker of the company, I want to further investigate those things. Then here we have in this section, by every operator, for every product, you see the average hours to produce, whatever they have to produce. You can see the numbers and you can see the bubbles with the same colors as below associated to each person. And the idea is to find out why some people are more efficient than others. Can these people teach the others on how to be more efficient? Or you know, trying to find weaknesses in your process to enhance them. So these are ways to analyze that. You can see a summary of what you have to be shipping for the next days. If defects is something critical, you can put some widgets to see defects based on defect types, based on locations and hopefully find trends. And again to find why some places are having less defects than others. So upper management can keep tap of Those high level KPIs again in data self highly interactive dashboards. If I were to select only two products, what about all of these KPIs only for one product? Let me choose only one product right now. I chose a single product and it shows me what's going on with this product. If I choose three products, the whole dashboard changes to show all the subcomponents influencing those three products. All right, so these are the out of the box KPIs we encourage you to think about if you're doing manufacturing and out of the box, if some of them would be of value will be up and running fairly quickly for you. What I want to show you next is some of the ways to customize data self. And by the way, when I go through my different slides here, sometimes if you look at the products we're creating, the people associated when they go from tab to tab, they're not the same is because this is a sample database and in our sample database we haven't normalized all of them. So sometimes people wow. Why? I saw one product in one tab and I saw completely different things in another tab. My sample database is not yet beautiful to show that consistency. We're still working on it. Anyhow, what I want to show you next is wow, beautiful dashboards. But how easy it is to build my own dashboards, how easy it is to customize things by myself if I am a manufacturing company. That's what I'm going to show you next is customizing. I'm going to show first the report and dashboard editor, which is usually for people who spend some time learning it, even though it's fairly straightforward. After, I'm going to show how natural language can also get people who do not have time to learn report editor and how they can get their own answers, their own questions answered by themselves. So now I'm going to go to the report editor and today I'm going to be using a sales sample database. We're hoping that the next quarter when we do other webinars for manufacturing, when I have a production sample database to show how you can do production new reports today I'll have to use the sample sales database. But the mechanics are about the same once we bring your data to the data warehouse. And then in this case for the tableau framework as a user you go and you see a data tab that shows that particular data set. In this case sales. But if you're in a production system, this could be your manufacturing data, your bill of materials, your purchasing. You can blend the data together. You can do a lot of things. Right now I'm going to be focusing specifically on sales and I'm going to build a new dashboard from scratch showing three different major metrics or key performance indicators. One is I want to see sales by customer, by product for the past two years. Then I want to see sales and gross profit trends over time. And finally I want to see sales by territory or by state in the usa. I Want to put them, all of these three widgets together in an interactive dashboard that will be sharing with other people. And eventually I'm going to show how to schedule the refresh and email of this dashboard to my users. That's what I'm doing in the next five minutes, all of these things. So I'm going to show that it's not too difficult to do it. So my first KPI of this dashboard is sales by sales, sales by customer by product for the past two years. So I need to find my customer is right here. I put in my rows. So these are my customers in my database. I want to see sales, so I take sales I drag into my report. So sales by customer in the whole company history, I only want to see for the past two years. Well, let me take document, date, input in my columns and by default the tool give me my whole company history, everything. Well, I only want the last two years. So I'm going to tell that my year here becomes a filter in my dashboard. So I click that and now I have a filter and I say only show me the past two years. And there it is, sales by customer for the past two years. All right, I also want to see sales by customer by product. Well, I can find product over here and drag it to my report. And there it is. I have sales by customer, by product in the past two years. If you know how to drag and drop, you can do a lot of these reports and it's all fluid. Maybe you want to see sales by product by customer. Sure, you take customer, you put after your product and now you have by product by customer last two years. Maybe you want to see quantity sold. Sure, you find the metric you drag and drop. And now you have by product by customer, quantity sold and sales for those two years. I'm going to remove all these things because I want to keep it simple for the example customer, product. But as long as people.