Bill Howell 0:00:04.7:
So today I have the pleasure of introducing Mia Radic. Mia is with Akamai Technologies. If you’re not familiar with Akamai, they are a $3.8 billion company with 10,000 employees that deliver the last mile of the internet to your doorstep, and a lot of in-between is what Mia tells me too. Mia lives here in Brooklyn. She is with her partner, who is in tech, and her dog, who is clearly not in tech, is what she tells me. She has a lot of great hobbies that were developed during COVID that she continues to pursue, which is awesome, including pottery and quilting, which is fantastic. Nearly seven years at Akamai, eleven years in supply chain. Mia was first a supply chain project manager, now leads a team that runs $200 million worth of planning and procurement spend at Akamai every year. So thank you for being here, Mia.

Mia Radic 0:00:51.5:
Thank you.

Bill Howell 0:00:51.5:
We really appreciate it. So, Mia, could you first describe Akamai’s business a little bit for everybody who doesn’t know it, and a little bit about what supply chain means to you at Akamai? A little bit different than, I think, some others.

Mia Radic 0:01:03.6:
Yes, for sure. So at Akamai, we do content delivery services, security services, and now cloud computing, and we do that through a network of about 350,000 servers globally. We’re in about 4000 locations and in 130 plus countries, so we have a vast network of infrastructure across the world. We like to say we make life better for billions of people, trillions of times a day, so our servers are getting pinged all the time, doing a lot of work across the world. So what supply chain really means to us is establishing that capacity of the network infrastructure and keeping it growing, replacing old capacity and so on all across the world to make sure that we can deliver those products and services to our customers. We think of the supply chain team as more of a sourcing and procurement planning function, but widely, the supply chain is figuring out, what are the capacity needs of the business? All of the different business units, all of the different products that need servers and need capacity. What are they doing, where do they need to be in the coming years, and then figuring out, how are we going to meet that demand? Can we give them physical infrastructure? Can we give them virtual machines?

Mia Radic 0:02:26.0:

Do we need to serve their needs in country or out of country, depending on where their customers are? So a lot of complexity to figure out on the demand side, and then the supply side of that is actually sourcing all of the materials we need. For my team, the big one is servers, but we also have all of the auxiliary infrastructure that’s related to that. Switches, routers, the actual racks that the servers go into, and then you have the colocation, so renting the space from a vast number of companies around the world to put those servers to have enough power to power them, to have enough people in those data centers to actually get all the work done at the right time. Then once the servers are there, the infrastructure is live, and it’s been QA’d, now we can actually use these machines to power a platform.

Bill Howell 0:03:15.3:
What’s beautiful about it is it’s super complex, and it’s evolved over time to encompass things like headcount planning and server planning, as well as time back into sales, which is pretty beautiful. A real connected planning story, I think, going on in Akamai that’s fun to watch. So supply chain, you said, has been up and running for about, what, four years now, five years?

Mia Radic 0:03:35.8:
It’s three, maybe, actually.

Bill Howell 0:03:36.8:
Okay, so maybe only three years. Could you describe what life looked like in supply chain planning before you had Anaplan, and then how it’s transitioned or evolved since?

Mia Radic 0:03:46.5:
Yes, we had a consultant tell us a few years back that we were running a $3 billion company back then with the resources of a $300 million company, so we needed to scale up. A lot of stuff was in spreadsheets, so we had a quarterly planning cycle. A whole lot of people would go and pull a whole bunch of data from different systems, so what’s going on with the capacity, what are our inventories like? We work with a lot of warehouses around the world, we have to be contacting them for all of the – how are stocks, what’s going on with our orders, what’s in progress and so on. So every month, we would do a whole round-up of information, even though we were planning still at the quarterly cycle, and then we would combine it in a bunch of spreadsheets and figure out, can we net out the server demand that we have against all of the inventory of servers we have globally and all of the stuff that’s in process all the time, to see, what do we need to buy next, and what does the forecast look like for the next year or 16 months? We are really giving our vendors a long-term forecast because so many components in our supply chain have long lead times.

Mia Radic 0:04:57.0:
During COVID, some of those lead times were 52 plus weeks if you can believe it, so having at least some sort of forecast was key for us to be able to get anything on time. So what Anaplan did for us is it brought us out of all of those spreadsheets that we were trying to copy over and roll forward every month, and it made it a lot easier to reconcile, what did we think last month versus what do we think now?

Bill Howell 0:05:22.3:
So really comparing past performance to future direction, and then how did you do against the plan that you put together? I’m curious, have you seen great results from that implementation and the performance you’re getting now? Has it improved, performance-wise?

Mia Radic 0:05:36.8:
For sure. One of the biggest things it improved is our visibility into what we have globally, and how useful it is, all of that inventory. Servers, as you can imagine, much like your home laptops, phones, the technology is changing all the time, things are getting better and better. So all of those servers I have in the warehouse from four years ago, they’re not quite as useful to me, even if, on our balance sheets, they have a lot of net book value left. We needed a way to be able to substitute or derive their usefulness today against the brand-new technology we’re buying now. So the really great thing that we built in Anaplan, we call it the substitution engine. It sounds really fancy, and it is. It was a way for us to pull all of those inventories and assign a bunch of attributes to it. So this was something that we kind of knew already, and certain people in the hardware roles knew. Like, this is a good machine, this is a bad machine, or this is something really old, no one is going to want to use that, versus every network, every use case can run on this really shiny new machine, but we could actually assign a whole bunch of attributes and say, ‘Here is how much memory we have, what is the disk space like?’

Mia Radic 0:06:49.9:
What is the CPU like, which networks have qualified it, so our big global network is actually divided into a lot of different networks that do different jobs. So we could take all of those attributes and we could group like types of servers. So we might have 30 different kinds of servers, we can group them into a few categories, and say, ‘These are the lightweight, these are the medium, and these are the larger, more expensive ones.’ Then we could say, ‘Here’s how much we have of the older versions, it should cover this much of our demand.’ We used to just say, ‘We’re going to try really hard to use up the old stuff,’ and now we can really put it on paper and say, ‘Well, 30 per cent of the ask should be covered with the existing inventory.’ We don’t have a case to go and buy more until we’ve used all of this up, or we’ve decided, as a company, this is no longer useful, we’re going to impair this equipment, or we’re going to repurpose it in some other way. So that visibility was really, really key, and what we got from the server supply planning model.

Bill Howell 0:07:50.1:
Yes, and these decisions are defining how fast the internet works for people globally, correct? So it’s, how fast we get information out depends upon the servers that you choose to put in, and when you replace those servers. Is that right?

Mia Radic 0:08:02.3:
Yes, it is, but if something is buffering at home, it’s not my fault [laughter]. Yes, but we have the world’s largest bank streamers, retail orgs, hospitals using our network to do what they need to do, and to get content to their customers globally, so making sure that we have the right servers in the right place is key. The internet does introduce latency with distance, so we want the server to be as close to you as possible. We want to have planned well enough in advance to put that server there.

[Aside conversation about the lights being dimmed]

Bill Howell 0:08:42.6:
All right, so thank you, Mia. I appreciate it. So when you first started using Anaplan, it sounds like you were completely unfamiliar with the product and the platform, so tell us a little bit about how you got familiar with the platform, how you learned to use Anaplan, and how quickly you got up to speed. Maybe a little bit about that background and that journey for you.

Mia Radic 0:08:59.7:
Yes, we were completely unfamiliar in the supply chain organization, but we knew that we had folks over in sales who were already using Anaplan, and it wasn’t new to Akamai at that time. That was really helpful. So we had some model builders already, we had some people who had done a sales implementation and were running some models. So it was helpful to have some familiarity of the company, but really, anybody in operations and supply chain was brand new to it. We started off watching some demos, looking at some interfaces, but then actually a bunch of people on my team did the model builder Level 1 training, and some people went on to actually do more than that. I’m halfway through Level 2, but it was really helpful to understand what’s under the hood. It’s helpful when you are the person making the spreadsheets and writing the formulas to know that you can trust what’s showing up in the model, that you know what’s – what are the lists, what are the modules that you’ve put in? When you look at a shiny app that’s displaying data, where is that actually coming from, and how is it being transformed into what you’re seeing?

Mia Radic 0:10:10.6:
I’m not saying that I could go into my model, and go and make adjustments, and do all of that. The model builders know way more than I do, but I am familiar enough that I can trust what’s happening, and I can make suggestions, and make adjustments in a slightly more knowledgeable way. So the training was really key. Also, we had a great partner during our implementation for supply chain. We worked with [?Servello 0:10:32.9] and they really held our hand through mapping out, what do our processes look like, what do we want them to look like, and walking us through, okay, you did this here. This is the process, these are the whys of what you’re doing, and here’s how it’s going to work now, and it made sense.

Bill Howell 0:10:51.4:
Yes, I know [?Brian] at Servello has added a ton of value to the project you’ve been working on, and there are other projects that have come out of that. Maybe describe what that working relationship looks like between you and Servello over time. How have you [?serviced 0:11:03.3] the use cases together? How have they helped you go down this path together?

Mia Radic 0:11:08.4:
Yes, I feel like there was a point where they knew what we were doing better than we knew what we were doing. We went through, we made a schema, and did all the DISCO diagrams, and I feel like, at some point, we were updating it multiple times a day. We would go through it, and every single day, map out different parts of the process during those initial stages and say, ‘Okay, here we have these inputs, we have these teams contributing their data.’ They’re doing this amount of preparation to that data. How can we simplify? How can we get this to work automatically or more seamlessly every single month? We really weren’t afraid to just rearrange the diagram, or make it work better, or look at – we had to learn how to think in terms of the dimensions in the model, so who needs to see which dimensions. At one point, does it seem nice to simplify, and at one point, do you need to explode it back out to a whole bunch of different parameters? So the Servello team, I feel like they really learned what we do.

Mia Radic 0:12:19.8:
I don’t think that there are people at Akamai who knew what we did as well as they did by the end of the process, and so they were able to go through those use cases with us, and then actually build everything out in a really nice, agile process. We did a really good UAT, made sure we caught as much stuff as we could, and then when we launched the model, we still kept updating it, and we still do today. We’ve had, with my model, about a year-and-change now of using it. I would say, every month, we find something else that we want to do differently, or some view that we want to change a little bit, or some additional data that we want to add, or something that we built that we just don’t use, and so say, like, ‘Hey, we don’t need that page anymore, we thought we did.’

Bill Howell 0:13:04.9:

Yes, that’s great. It sounds like they’ve been really the definition of a true partner in the business, which is great to see, and I love the – the use cases that continue to come out are actually pretty impressive. How have you quantified value? I know we’ve done some work with your team about business value. Have you done some work yourselves on quantifying the value you’ve gotten back from the supply chain projects?

Mia Radic 0:13:25.3:
Yes, in terms of just dollars saved, that substitution engine, and being able to make better use of our existing inventories was huge in terms of savings. We are now able to just be a lot more efficient with our monthly process. We expanded it from just servers that’s managed by my team, to our team that manages all of the other equipment, so there’s a lot more complexity. I’m lucky to work on just one – one complicated product, but still just one product line, whereas the other team has – works with a ton of different vendors, a lot of different manufacturers around the world. Their model has dozens and dozens of pages and sections that different people work in. So we can now speak the same language when we talk about supply chain planning. It’s not just that one person does it in a really sophisticated way, and then they hand it off to somebody who just kind of adds ten per cent and calls it their forecast. We have very different and similar-in-philosophy forecasting methods now across all of the different products. So I feel like it really brought us on the same page in that way, and so that’s been a huge benefit. The time savings for the whole team has been really great.

Mia Radic 0:14:48.8:
It’s enabled us to look at what we want to do next, where we want to make improvements to work closer with our vendors now that we have the time to. So in that area as well, it’s been a big benefit.

Bill Howell 0:15:00.0:
That’s awesome. So if you could describe the biggest wins that you’ve had, along with the biggest challenges that you’ve had along the way, I’d be curious to hear both sides of that coin.

Mia Radic 0:15:09.4:
Yes, I think the wins are that our processes are – they’re reliable and they’re repeatable, so we can step in – we roll forward our model in the first week of the month, and then we can step right in the next day and say, ‘Okay, what has changed?’ What do we need to do now? Who do we need to go talk to if something is very different than we expected, or we know that new asks have come up and so on. We have more standardized expectations across the team. That’s been a challenge to get there because we had to not just learn what Anaplan does, but then move all of our processes into it, but now it’s much easier. It’s hard to imagine that we did it all in spreadsheets, like, a year-and-a-half ago for some things, three years ago for other things. So that’s been a huge benefit, is that everybody understands what’s happening in these models and why we use them. We also present cleaner, more consistent data across the business.

Mia Radic 0:16:17.5:
So when our CFO comes to us and asks, ‘What are you spending, what does your budget look like for the next year, how does your outlook compare to what we budgeted [?and told the street 0:16:27.0]?’ We can really easily tell them, and every month, we can produce something that looks exactly the same, which is… Easier said than done.

Bill Howell 0:16:38.0:
It looks like magic.

Mia Radic 0:16:38.8:
Yes, it does. It does feel a lot better to be able to consistently deliver that, and to do it in much less time. I think, what’s also been nice is that it feels like we control the quality of our model, and it feels like we can change it if we want to. We didn’t have supply chain apps that the team talked about this morning when we did our implementation, but it sure would be nice if we had even something to start from. We built everything from scratch. We did everything customized to exactly how we wanted it, but that also means that we can just keep changing it, and we can keep making improvements when we need to. Our, what we call that beyond server, all of the other stuff model that we launched only a few months ago, we’re constantly improving it. There are multiple tickets a week that are getting worked on, and a whole bunch of stuff being launched every single month into production to improve our models and make them better.

Bill Howell 0:17:35.2:
It’s not an army of people that work on it, is it. How many folks, total?

Mia Radic 0:17:38.8:
There are about a dozen model builders across the company, but really only three that work on supply chain, so it’s a small team.

Bill Howell 0:17:45.0:
Yes, that’s great, and as part of a CoE that’s covering sales for territory quota, also covering the other parts too, the CoE of model builders is the 12, and you’ve got three that are just supply chain.

Mia Radic 0:17:57.1:
Exactly, yes.

Bill Howell 0:17:57.8:
That’s good. I’m sure you’ve learned a ton during the implementations you’ve done that inform what the next implementation should look like. What are some of the most important learnings you’ve had?

Mia Radic 0:18:09.7:
Data is really tough. We’re now a 25-year-old company. Everybody has their own data source, everybody trusts their own data source, and we have ten different places that you could get an answer to the same question. So standardizing that, trying to get people to buy into, why should we do it this way, why should we have some data standardization practices when my stuff works just fine over here [chuckles], that was really challenging. We continue to see issues with data, and as we want to add more use cases, and add more stuff to our models. I wouldn’t say don’t do this until you’ve figured everything out because you’ll probably never get there, but that was one of the biggest issues that we had. Where does it all need to come from? What are the right sources of truth? How do we eliminate all the lag that we had before? We are a very technical company, we have a lot of people who just will – they’ll build their own tables, and they’ll use their own queries. We’ve also done some standardizing in that area to use more common dashboards and more common tools for reporting, but before we did that, it was every group for themselves doing things. So data was a big challenge.

Mia Radic 0:19:35.5:
Understanding the processes and mapping them did take a little while. Things were complex, and we didn’t really have it documented. We had maybe some specific stuff documented of, here’s how you do this, so that when you have a new analyst, you can point them to an SOP and they can hit the ground running on something, but we didn’t have a cohesive view of what does supply chain really mean, and how do we broadly put it into one document or one model that didn’t exist, and that took quite a while.

Ben Howell 0:20:05.5:
Got it. So I’m going to throw you a curveball. Not one we’ve talked about before this session. So on a personal level, how has your career changed since you started doing this project and working on these initiatives within Anaplan?

Mia Radic 0:20:22.3:
I do think that Anaplan saved me time, and I think it gave me some confidence in what we’re doing. It enabled me to talk about it across our groups more broadly. So I’m in an interesting position where I’m a manager of a relatively small team, but every month, I go to our CFO and I say, ‘Hey, here’s what the model is saying, here are all of our inputs. This is how many more millions of dollars of servers we need. Please approve it [laughter].’ Usually he does, or he says, ‘No, go back to the drawing board and figure out’ – here are the metrics that we need to hit. Figure out, what are the inputs that are maybe more flexible, and we’ll go and do that, but it’s a lot easier and a lot faster. So it used to be that we would have to do a lot more work, have to do a lot more manual spreadsheet intervention to adjust, like, what does this forecast look like? Okay, maybe we’ll haircut all of this by ten per cent. It’s a lot faster and easier to do that in Anaplan, and what we’d like to do more of is some more scenario planning where we can say, ‘Okay, this is right in line with budget, but what if we do start to see more revenue in this area, and we do need to make slightly more investment?’

Mia Radic 0:21:41.5:
Could we do that? Would we have the parts to do it? Will we be able to get the servers there fast enough? That’s really key for us [unclear words 0:21:47.3]. So for me, careerwise, I think it has put me in a place where I am more confident about all this stuff that I’m talking about. There’s a lot less human error in the whole process, which makes it easier to be confident about what’s happening in our systems and tools. It’s given my team more visibility across the company. People know what we’re doing. They see the sophisticated tools that we use. They know it because they see it on the sales side. We hope that we’ll be able to do more integrated planning through our sales team and our quotas, and work with customers, and bring that all the way over into supply chain. So I do think it’s elevated the whole team in a way.

Ben Howell 0:22:31.5:
That’s great. The confidence is evident, for sure, and exposure to the CFO is good, as long as you have answers, right.

Mia Radic 0:22:37.2:
Yes, and if you don’t, you just sit there quietly and [laughter]…

Ben Howell 0:22:41.4:
Soak it in! A follow-up question from me, and then we’ll open it up to the audience, so if you have questions, feel free, and have them teed up. What advice would you give this group for folks that are new to Anaplan or new to initiating projects specifically on the supply chain side? What guidance would you give them that you wish you’d had when you started?

Mia Radic 0:22:59.9:
It really can do most stuff. It’s really hard to think of some kind of process or something that you’re doing in your supply chain that can’t function in Anaplan in some way. We’ve seen our model builders and our implementation partner be so creative and imaginative in what this tool can do. It’s not a spreadsheet, it’s not just a slightly different way to look at a spreadsheet or a pivot table. There’s so much capability there, and some stuff that we’ve done is – it’s just really cool, and it’s really interesting that we were able to get there. So keep a really open mind on, like, what the possibilities are. I don’t know what it looks like with the supply chain app now, but it sounds like there’s a really good compromise now of some out-of-the-box capability and then a lot of flexibility and customization to exactly what you do. Nobody out there wants the Akamai supply chain model. It is so specific to what we do, in dealing with servers and global vendors, and warehouses, and data centers where they need to go, but it’s really cool that we were able to make that, right. So think really broadly about what you would want your supply chain models to look like, and what do you need them to do, and what’s actually helpful.

Mia Radic 0:24:21.6:
What is the stuff that you just do today that maybe you wouldn’t actually do tomorrow if you didn’t have to? Get your data ducks in a row as much as possible. So it’s also a good opportunity to say, ‘We’re going to get a lot of value from this tool.’ We know that first, we need to do some data standardization. We know that we need to have some better governance over where things are, who can have access, which teams provide their inputs, and which teams just read everything that’s there. So spend a decent amount of time figuring out all the data because it will make things easier. Not that you can’t, of course, keep changing it, and keep updating it as you go along. Then, yes, think about how you’re making things repeatable and scalable. Maybe you are working with a hundred products today, but you should really be building this model to function with a thousand products. Can you categorize things more? Can you simplify views for yourself, or for your management and your leadership, so that you’re not bogged down at that line-item level or whatever it is years from now when things have scaled a lot.

Bill Howell 0:25:38.2:
I love it. Think out of the box, think big, think ahead. That’s what I’m taking from it. That’s good. Awesome. Thank you, Mia. Questions from the audience. Who has got a question for Mia? I’m sure there are some. Yes, right here.

[Aside conversation about getting a microphone]

Audience 0:26:00.3:
I guess this is for you, Mia, but maybe, Bill, you can kick in on this one too. What role, if any, does the Anaplan platform play in your security strategies that you use to protect [?and ID 0:26:12.5], say, unauthorized intrusions, ransomware, etc.?

Bill Howell 0:26:17.3:
Interesting. You’re up [laughter].

Mia Radic 0:26:19.0:
Yes, so we have a lot of really sophisticated security products at Akamai. They run off of servers. They run off of either bare metal machines that we’ve put somewhere, or they run off of virtual machines that are spun up off of bare metal machines that we’ve put somewhere, right. Serverless just means there’s a server somewhere. For us, it means getting those inputs from all of those business units and the products that they build. That’s one of our fastest growing areas as well, so getting accurate forecasts of where those customers are, and where the products need to have the capacity, and then getting – whatever the capacity needs are. Maybe they need a certain amount of CPU, central processing units. Maybe they need a lot of disk space for some particular uses, so translating those into a number of servers that we need to buy, and then that becomes the Anaplan driving forecast for server supply planning. So we have, not just security but all of the different products that run on the Akamai network giving us demand signal into Anaplan, into our demand planning model, and then that demand planning model, which considers the global needs pushes data into the supply planning models.

Mia Radic 0:27:37.9:

The supply of servers, the supply of all of the other hardware, the supply of the colo. The supply of the field tech engineers that we need six months from now, a year from now, in which countries, to be able to do more deployments, so that we have the servers where we need them, and those servers can power the products that do things like ransomware stopping, DDoS stopping, and so on.

Ben Howell 0:28:05.0:
I’ll add a point that’s completely not related to security, but really interesting to me. One of the interesting initiatives that Akamai is pursuing now is the harmonization between capacity to deliver and sales. Capacity to sell to clients, right, so what they’ve found is – they’ve looked at their margins. Sometimes, they’re overbuilding capacity where they don’t have a market for that, and on the other side, they’re overselling where they don’t have the capacity to deliver, and so part of what we’re doing – because we work with sales and supply chain, we’re harmonizing that, and working on a project right now to kind of – it’s, I think, a bit like IBP or S&OP where we’re pulling those two pieces together. Pretty interesting and, I think, pretty powerful for the business, and something else Mia can talk to the CFO about.

Audience 0:28:55.3:
So does this tool help you to manage the volatility of the energy usage and energy price, right, which is probably the most vital factor for the data center business?

Mia Radic 0:29:02.5:
So we do have a colo and power management model too. So when I talked about building out all of the different variables related to servers, one of them is power use. Over the years, the servers that we’re buying, they’re only getting bigger and beefier, and they need more power, which also means that – you used to be able to put 40 of them in a rack. Now you can only put ten in a lot of data centers that aren’t actually able to support that much power. So our colocation model does take a lot of those factors into account, so we are trying to see – let’s say we need to add a hundred servers worth of capacity in Munich. What are the data centers who we work with? Do we have the space, but then also, do we have the power, and can we be forecasting power out for the next year, two years, and so on. So we’re trying really hard to do that, especially in places where we have a big presence, and we know we’re going to need to be changing out the servers that we installed maybe, like, seven or eight years ago with the new modern ones. Those power consumption demands are going to really change, right, so it does help us through that. That’s specifically happening in our colocation model.

Ben Howell 0:30:19.6:
Great questions.

Audience 0:30:25.9:
Hello. You mentioned about the different data sources before implementation of Anaplan, so probably, you need to cleanse a lot of data. Have you done this in Anaplan, and if yes, how did Anaplan help you to do that, to run that process?

Mia Radic 0:30:40.7:
Yes, for the most part, we don’t do that in Anaplan. We get as much clean and good data as we can into our data hub. So the trickiest part for us was figuring out, like, what are the key pieces of data? We don’t want to upload everything there possibly is. We want just the right amount that enables us to do what we need to do today, and maybe tomorrow, but not the vast, vast, vast amount of data that Akamai has. So that was part of the original mapping, and deciding what’s going to go into the data hub, and then the data hub also exports out, at the end of our monthly planning cycle, all the key parameters that other tools need, or other groups need, or that we just want to use for reporting. So we do a lot of reporting through Tableau, for example. We don’t do most of the cleaning and organizing in Anaplan directly.

Ben Howell 0:31:44.6:
Other questions?

Audience 0:31:49.4:
Hi, Mia. Do you do any kind of forecast sharing or collaboration with your suppliers, and if so, how has that worked? Have they been able to be more supportive of you?

Mia Radic 0:32:00.4:
Absolutely, we do. We wouldn’t be able to buy anything if we didn’t [chuckles]. So we share a forecast. I’d say about 18 months out right now is what we feel like is still pretty reliable. For servers specifically, our vendors rely on that to be bringing in all the components that they need to bring in. We try not to actually give an order for servers until relatively just in time for us, that’s about six weeks out, but all of the parts really have to be there for us to build six weeks out, except for the little commodities and things that are available really quickly. So our vendors tell us that our forecast has improved. We produce a forecast and then we add an upside to it for the parts buying, and then anything we don’t consume, we roll forward, but we’re also able to monitor the inventory of parts in Anaplan, and say, ‘Okay, well, we have all of this demand coming in, we had forecasted 1000 servers for next month, but Anaplan is really saying we need 1500; can we do it?’ Are there enough parts in that forecast that we gave the suppliers to make that happen, and if not, okay, great, we’ll still order the 1500. Hopefully, we’ll get the thousand on schedule, and then when do we see more supply coming in?

Mia Radic 0:33:19.4:
So part of the monthly process is we tie off everything internally, and then it all just goes to the suppliers to say, ‘Here’s what we think we need, here are the different types of components,’ and for some suppliers, we even break it out. We have the bill of materials for particular servers in Anaplan, and so if it’s this 1000 servers, it’ll give them actually, like, hey, go buy a thousand chassis and motherboards, and 8000 drives, and so forth. So we give that signal.

Ben Howell 0:33:53.1:
All great questions. Time for one more. Going once. Sold.

Audience 0:34:06.9:
Hi, Mia. I wanted to ask you about how the regionality changed your models. Different parameters that might come into play based on the geography, or based on the market or segmentation, and how do you apply that within Anaplan?

Mia Radic 0:34:26.1:
Yes, good question. For the most part, we try to use Anaplan to standardize as much as we can. So the types of equipment that we buy for North America are the same that we buy for EMEA, APJ, and so on. We do have to keep track of some differences, like particular certifications, particular import restrictions. There are countries where we can’t deploy used equipment, for example, so those types of things, we’re able to build into our model to say, ‘Hey, you might actually have a bunch of useful used equipment that you could redeploy here, but you can’t do that because this country will not accept those types of imports,’ right. So that’s a little bit of a complication that we’ve been able to work in. In the countries where we know the forecast is uncertain, we try to maybe add a little bit more buffer. We have particular regional warehouses globally, where we also – we do their planning isolated. So for Brazil, for example, we run Brazil’s model independently of the demand for North America or somewhere like that. So we’ve been able to isolate the different levels of geographic split for our demand planners.

Mia Radic 0:35:47.2:
Most of them plan at the country level, or for really big countries, kind of region of that country, but for supply planning, we care about the warehouses in which regions globally, those warehouses serve. So the Brazil warehouse only serves Brazil. Really easy case there. The same with India, for example, but then we have a really large North America warehouse that actually serves all the rest of the Americas besides Brazil, and maybe a lot of APJ as well, so that needs to be planned in a different way. It needs to encompass more buffer, it needs to encompass requirements of a whole bunch of different countries.

Ben Howell 0:36:23.1:
That’s awesome. Mia, thank you so much. Great story. Really appreciate you sharing it.

Mia Radic 0:36:27.3:
Thank you.