Chris Platzer 0:00:09.4:  

My name is Chris Platzer. I'm manager of ACoE; stands for Automation Center of Excellence. I've been with AWS for a little bit over seven years, and primarily responsible for automating manual activities. So, within sales and sales operations we have a quite a few manual activities, even at AWS. While it's a very large organization, we grew rapidly so we had to adjust and ensure we can scale. Automation is one of the means to reaching that scale. 

 

Dana Therrien 0:00:41.1:  

Can you just give us an idea of the scope of the Anaplan deployment at AWS and what you're using us - the major problems that you're using us to help solve? 

 

Chris Platzer 0:00:52.1:  

Well, the first one that comes to mind is sales hierarchy. Originally, sales hierarchy was managed disparately across layers of our business. We had it managed in Excel. Then it went to Quip, and we used a home-grown solution in Honeycode for a brief period of time to understand our breadth of sales engagement through their management chain. So, essentially account manager, district manager, etc. 

 

Dana Therrien 0:01:18.3:  

So, the sales hierarchies, the organizational structure for AWS from AVP down to VP down to division, whatever it might be? 

 

Chris Platzer 0:01:27.0:  

Exactly, and Anaplan gave us an opportunity to put this in a single place, facilitating role-level security not only for our Anaplan models, but also our reporting suites, Tableau, etc. 

 

Dana Therrien 0:01:42.2:  

So, it's feeding all the other source systems that might be using the sales hierarchy to determine relationships and routing of, let's say, accounts for quota-setting purposes and things like that. Is it feeding into Salesforce? 

 

Chris Platzer 0:01:57.3:  

It is feeding into Salesforce. It's also feeding into various models we have in Anaplan, some being for sales forecasting, quota setting and the like. 

 

Dana Therrien 0:02:07.4:  

So, sales hierarchy also and territory and quota and forecasting, anything like that? 

 

Chris Platzer 0:02:14.3:  

Yes, and it's proven to be much more reliable than Excel, as you can imagine, or disparate systems. So, you also can connect that to when - let's say you have an account manager that's changing roles or you have someone joining the organization or leaving the organization. You're going to need the ability to have that visibility transferred through role-level security for what reports they'll see and how quota will change - as you previously mentioned. 

 

Dana Therrien 0:02:47.2:  

You said your responsibility is to go out there and automate manual processes. How do you do that? Do you just go out there and start hunting in the organization and looking for things that can potentially be automated? Or people bringing them to you? 

 

Chris Platzer 0:02:59.5:  

It's a mixture of both, to be honest. I think we have some scenarios where we've been able to identify areas of opportunity by reaching out to the field and saying, ‘Okay, what's happening?’ This is both within our field sales but also our sales operations, and understand where time is being spent. But we're also getting quite a bit of interest, I'll say, from various parties within the org. So, we may have - our annual planning team may reach out and say, ‘Hey, we've got a serious problem here to navigate.’ So, we'll have to balance that backlog and understand what's priority and then take steps to remediate. 

 

Dana Therrien 0:03:40.2:  

Do you work within a Center of Excellence? How is it structured? 

 

Chris Platzer 0:03:47.4:  

Great question. I'd say our Center of Excellence is building. We're starting to build that muscle so we're more collaborative and intentional with our efforts at scale. Right now we're, to be honest, somewhat disparate because we're building a lot of these mechanisms to a standard, a standardized format. But I think one of the benefits of Anaplan is that it enables us to facilitate mechanisms that are not completely standardized. 

 

Dana Therrien 0:04:18.4:  

When I think about the AWS product, it's largely a usage-based product. Is that is it fair to say that?  

 

Chris Platzer 0:04:24.5: 

Yes.  

 

Dana Therrien 0:04:25.4: 

So, if you think about the traditional sales motion, you've got sales opportunities that are getting managed somewhere. Then once those opportunities close, you get an order. Then you're uncertain as to what the usage could potentially be once you close a customer. Maybe there are ways that you could derive that by looking at the characteristics of that customer, certain things like the number of employees that they might have, the size of the organization. But eventually, that opportunity forecast has to get turned into a bookings forecast. That bookings forecast has to get turned into a revenue forecast. You need to provide some predictability to the leaders of AWS so that they know what's going to happen six, twelve, eighteen months down the road. When we're thinking about forecasting, what elements of that are you using to manage, or Anaplan to manage? 

 

Chris Platzer 0:05:18.4:  

There's a lot there, but in short, yes, you're right. It's looked at regularly by our leadership. Some of the elements there, you mentioned opportunities. Sizing and understanding how much revenue we're going to recognize from those, is a little bit complicated because it's customer-driven. It's like a utility company so we're not certain whether or not you're going to turn on your lights and leave them on overnight. We have no visibility to that, but we do have historical trends. So, we are able to leverage machine learning, neural net models to push that forward and make some assumptions, so that's one element. Pipeline: we are able to discern to some accuracy what pipeline will deliver. It will depend on the volatility of the customer. You can imagine, our games industry, highly volatile, so in some cases we do leverage that. In some cases not, if that makes sense. 

 

Dana Therrien 0:06:22.3:  

In the forecasting cadence of AWS, how frequently are you rolling up a forecast across your organization?  

 

Chris Platzer 0:06:30.6:  

Continuously. Weekly, we have a forecast review. 

 

Dana Therrien 0:06:37.2:  

From the sales organization?  

 

Chris Platzer 0:06:37.9: 

From the sales organization. 

 

Dana Therrien 0:06:38.8:  

You're trying to discern which of those opportunities are going to close? Price? 

 

Chris Platzer 0:06:43.2:  

Yes, and it's aggregated up so we actually leverage our field teams to help us - help inform us. 

 

Dana Therrien 0:06:51.2:  

So, there's a bottoms-up aspect to it? 

 

Chris Platzer 0:06:53.5:  

There’s a bottoms up and a tops down; both, yes. 

 

Dana Therrien 0:06:55.5:  

Then once you can determine that, then you try to predict which of those opportunities are going to close, and then after that is, at what amount do we think that they're going to start billing, and then for how long, and you’re projecting it out that way? 

 

Chris Platzer 0:07:09.3:  

Yes, in short. Our bottoms up, so we leverage Anaplan to engage with our field teams, get feedback on our machine learning model. That helps guide that forecast as it's aggregated up that sales hierarchy. We can, ultimately, report to our executive leadership across AWS. 

 

Dana Therrien 0:07:32.0:  

As far as I know, I think AWS might be Anaplan’s largest customer that's using cross-organizational forecast, so from sales to bookings to revenue. Then are there any cost elements? Is there any tie-in to opex from that? Are they making decisions about where to invest based upon what they're seeing and projecting for revenue? 

 

Chris Platzer 0:07:53.2:  

I would say that's in the road to come; I'd say that's in the future, ideal-state future. I think there's a lot of opportunity to go there. 

 

Dana Therrien 0:08:02.4:  

Yes, so when I start to see the opportunity for Anaplan, it's a progression. I'm starting to see more companies start to think about that. I started just thinking about it as a cross-organizational forecast and reforecasts throughout the year, and then thinking about opex after that fact. 

 

Chris Platzer 0:08:19.3:  

Yes, it's a great opportunity. 

 

Dana Therrien 0:08:21.0:  

There've been a lot of questions just on a technical level on how you're managing Anaplan. What advice would you have for people who are just starting their journey now and thinking about how to construct their models and what types of individuals? 

 

Chris Platzer 0:08:39.5:  

I would mirror some of the guidance given earlier today, which is: don't try and do everything all at once, and start small. I think that's very good advice. Have focused objectives that you're meeting with the output of these models. Be set on purpose of the model and have stakeholder engagement throughout so you're keeping your tree trim, so to say. 

 

Dana Therrien 0:09:11.4:  

Do you ever spend… Are there any models that have been in operation for a long period of time and then you start to go in and reassess them and whether they're performing the way you wanted them to and what needs to - what could be improved? 

 

Chris Platzer 0:09:25.2:  

All the time. We were talking about forecasting earlier. We've been through three different rebuilds of forecasting. It wasn't required. Truthfully, we probably could’ve kept the same model in place. But part of the Amazonian culture is to iterate on mechanisms, so we do that. We iterate on the mechanism. We look at what inputs we're taking in, the data we're leveraging, what calculations we're actually applying within these modules and then the expected output. At times, yes, we see benefit from redesigning and launching something quickly. That redesign doesn't necessarily take a long time; we've redesigned models in two weeks. 

 

Dana Therrien 0:10:14.5:  

When you redesign it, do you scrap the old model, or do you refine the existing model? 

 

Chris Platzer 0:10:19.4:  

Great question. It depends. In the example of doing one in two weeks - and this was specifically for EMEA as a geo - we built a completely new model for forecasting for their entire sales hierarchy within two weeks, so it depends. 

 

Dana Therrien 0:10:37.2:  

You mentioned EMEA. Is it completely different than the one that you would use in North America? 

 

Chris Platzer 0:10:43.3:  

At that time it was. This was 2021, I believe, but yes, at that time. Now, they're in the same model. 

 

Dana Therrien 0:10:56.1: 

They're in the same model, so why didn't you just lift-and-shift the one that you had existing and add people to that model? 

 

Chris Platzer 0:11:04.3:  

It was mostly due to data requirements. Within that business, they had different requirements than what we had within North America or APJ - APAC at the time. So, we decided to keep our model size small. It makes sense. We have the flexibility to launch something different to another geography, but still consolidate the output of these models into one outcome, to one output. So, we facilitated the needs of the business, kept our data size small, at the same time meeting our objectives. 

 

Dana Therrien 0:11:47.4:  

So, was the idea to stand them up fairly quickly, get them using Anaplan, let's solve this forecasting problem and then we'll bring them into the family after the fact? 

 

Chris Platzer 0:11:56.3:  

Yes, exactly. 

 

Dana Therrien 0:11:56.9:  

Is that typically how you iterate? 

 

Chris Platzer 0:12:01.3:  

Not always, but this time it really was useful because there are so many field sellers that we're engaging with as part of this mechanism. So, we wanted to meet the customer where they are, facilitate that need and then iterate towards standard solution. 

 

Dana Therrien 0:12:18.5:  

Amazon is just an incredible technology company. We all know what Amazon does and we know there are lots of different businesses within it. But based upon my experience with the company over the years - not just with Anaplan but with other companies I've worked with - there's a pride of ownership in building things themselves because they define an entirely new market through technologies that they developed. When you first started getting involved with Anaplan, was there an inclination for them, 'We don't need Anaplan; we're just going to build it on our own, and why would we need to go outside'? 

 

Chris Platzer 0:12:54.3:  

Well, I think there's always interest in being able to build within Amazon, but I don't see it as… I think it transitioned quickly from competitive to collaborative. I think when we were able to identify that we had the ability to put in place a scalable solution that's multi-dimensional, replacing the need for Excel immediately, I think the approach towards Anaplan shifted at that point in time. However, the short answer to that: yes. Amazonians like to invent and simplify and have the ability to build. But I think the focus has now shifted to what we can leverage Anaplan for. 

 

Dana Therrien 0:13:47.4:  

I know there are many different pieces of Amazon out there. Are there other businesses inside of Amazon that have an interest in doing what you guys are doing? Do they ever question you? How does that go? 

 

Chris Platzer 0:13:58.4:  

Yes, there are quite a few that reach out. Typically, how that goes: they'll want to understand how we developed our business case, what went into - what kind of information I provided to help bring Anaplan to AWS. I'll share that information. They'll walk through what training looks like. What does kicking off Anaplan look like? What is a good use case for that first step on the moon, so to say, to start that process? That's where we typically start conversating and building out plans and collaboration. 

 

Dana Therrien 0:14:35.2:  

Because it's a different business unit, it's not like you can, 'Hey, we already have Anaplan stood up. Why don't you just come and join us and use what we've already built?' It can't work that way. 

 

Chris Platzer 0:14:45.4:  

It typically does not scale to that degree. Typically, as an example, Eero was one that stood up Anaplan. They had to work through that somewhat on their own. While I was able to help and walk in their instance and talk to them about things that we would do, they had to build that muscle. 

 

Dana Therrien 0:15:10.1:  

So, the primary - the foundational use case was around the sales hierarchy and managing that and all the complexity around that. Then forecasting is obviously a huge use case for you as well. Are you also doing territory and quota planning? 

 

Chris Platzer 0:15:24.1:  

So, there is some quota planning going on. In-year planning is definitely being talked about, how we manage some of those motions, continuous planning. On the roadmap, going to connected planning, so we are going through elements of those - but I would say not overall, if that's fair. 

 

Dana Therrien 0:15:49.1:  

I've heard it called always-on planning or continuous planning. What does that look like? What are the features, the dynamics of that? What are some of the requirements? 

 

Chris Platzer 0:15:57.2:  

Some of the requirements are, I think I mentioned earlier about people coming on board and leaving and people in transition. So, any kind of change to sales operations, the business from a sales lens, has to be captured. Then we have to adjust quotas, territories, comp plans - you name it. So, having an environment where it's connected so we're not having siloed operations. For example, you may - let's say someone leaves: a sales manager has to put in a request, 'Okay, I've got somebody leaving, I'm going to tell HR.' Now they have to go through a separate motion, 'Okay, I'm going to have to adjust quotas. I may have to set up air cover.' Separate motion. Comp plans. Those are all really one event, so having an event to an intended outcome as a connected environment is the ideal state, instead of disparate motions through siloed operations, with hopes of efficiency in execution. 

 

Dana Therrien 0:17:02.5:  

I've never heard it called - I've never thought of this. That's an interesting way to think about it; as an event that's got multiple repercussions throughout the entire ecosystem. Is that how you guys are thinking about it now? That's how you're scoping this problem out?  

 

Chris Platzer 0:17:15.7: 

Yes.  

 

Dana Therrien 0:17:16.7: 

Yes, so what does that process look like? 

 

Chris Platzer 0:17:20.0:  

It's a lot of research, I'll say, gathering information and understanding how these siloed operations are actually functioning. Understanding the requirements of building a connected solution and ensuring the stakeholders are aligned with the next steps. That's how I'm thinking about it currently. 

 

Dana Therrien 0:17:38.2:  

Are there multiple organizations involved in these, let's say, sub-events that happen? Like HR is involved and sales operation is involved, maybe finance is involved, comp is involved - and even sub-organizations within organizations. 

 

Chris Platzer 0:17:52.2:  

Exactly, so there are a lot of stakeholders to gather requirements from and understand expectations both from a timing and an execution perspective. 

 

Dana Therrien 0:18:01.5:  

In my career, I was always amazed that the sales operations team was often the champion of something like a CPQ project, because they didn't really own the product. In most cases, they didn't own pricing. But it was in their best interests to streamline that process because it created sales velocity. It increased bookings and it increased revenue for the company, and the product team didn't really want to take it on. It was hugely complicated, so I found that sales and revenue operations were often taking on projects that weren't… If it didn't get solved, it was going to be a problem for sales. They basically said, 'We'll solve it because nobody else wants to do it.' So, now you've got these disparate organizations that you're dealing with. Who's the one that comes forward and says, 'I want to solve this problem,' even though there are all these different suborganizations that are part of it? Who's the guy that raises his hands and says, 'Just let me do it'? 

 

Chris Platzer 0:19:05.1:  

I don't think anyone's raising their hand eagerly, I would say it's… 

 

Dana Therrien 0:19:09.4:  

Yes, it's like… 

 

Chris Platzer 0:19:11.2:  

First, the one that comes to mind is comp. They're on the hook to create those comp plans, have alignment and ensure that we're not at risk as an organization when we have shift. So, they've got a deadline they have to meet. They are raising their hand. They're interested. I would say sales themselves, enabling our sellers to be able to understand where they sit immediately, that's ideal state. They're our customers so we want to make sure they're happy. 

 

Dana Therrien 0:19:41.5:  

So, it comes from a need from a sales manager to quickly on-board somebody, put them in their seat, help them understand what their account assignments are, what their territory is going to be, how they're going to get paid, and then hit the ground running so that they don't have a two-year ramp period; instead, they have a sixty-day ramp period? 

 

Chris Platzer 0:19:57.3: 

Exactly.  

 

Dana Therrien 0:19:59.2: 

Yes, so it still holds true that sales is often the champion for some of these problems because they're the ones that are really held most accountable for the results, I would say. It's easier to measure sales than most other organizations in a company because they're all carrying quotas and you can see whether they're hitting it or not. 

 

Chris Platzer 0:20:18.3:  

Absolutely, they are the champion for us, yes, without question. 

 

Dana Therrien 0:20:22.3:  

As the leader of this group, how do you facilitate those conversations and speak to your customers? How do you not get inundated with all these types of requests? Well, I saw some pretty good models in the last presentation that showed how they were assessing it in terms of ease of use and impact. Are you guys doing anything like that? 

 

Chris Platzer 0:20:41.4:  

We do have a matrix, so to say, level of effort impact and urgency. So, we do have some idea that we rate our backlog and go through a prioritization matrix. But we also have a partner team that helps us identify these requirements and helps us solidify the mechanism in areas where we can standardize and implement tooling. We do have some partners in our - on our team to help focus on stakeholder management, communication, enablement and also capturing requirements. It enables us to focus on building. 

 

Dana Therrien 0:21:21.1:  

All the models inside of AWS, are they locked down tight so nobody can touch them? I've spoken to some other individuals where they had a federated model, where there were certain ones that were common to the entire organization and they were managed centrally. But then the organizations out inside the field had the ability to go out there and create some of their own. If they worked quite well, they would be brought back in and then they would get federated throughout the entire organization and dispersed. Is that still the case? 

 

Chris Platzer 0:21:52.0:  

I would say it was once the case. So, for example we had a quota model that was transitioned to the central team. We had a sales hierarchy model that was transitioned, and that became sales hierarchy. In many cases, that transition has changed now. So, now the federated teams are owning the direction of these models but also collaborating to create a centralized model structure. 

 

Dana Therrien 0:22:23.3:  

How do you manage that need for central control and administration versus allowing for flexibility and ingenuity at the field level, where the people are experiencing the real problems? 

 

Chris Platzer 0:22:36.3:  

In my mind, it's output. Output will drive the requirements that each of the federated teams would be aligned to. As long as you have a standard output and outcome, it drives you towards standard operations within those models to facilitate it. 

 

Dana Therrien 0:22:55.1:  

So, explain that to me a little bit. The output, let's drill into that for a second: what is that? 

 

Chris Platzer 0:23:01.4:  

So, for example, forecasting may be a great example right now. We have different means of managing the forecast across our different granularities of business. So, we have some incredibly large customers where we may forecast very differently from our very small customers that are high touch versus low touch. Some that are self-serving and a complete technical aptitude to develop a cloud infrastructure. 

 

Dana Therrien 0:23:29.4:  

So, a big customer, it might be more of a collaborative forecast where there's some bottoms up and top down and some heads that are put on top of those things; is that a fair assessment? 

 

Chris Platzer 0:23:41.4:  

Yes, exactly, and so these might be different models, so to say, different mechanisms in a sense, but they each deliver the same outcome when we need to share it with executive leadership. 

 

Dana Therrien 0:23:53.7:  

I see, yes. In the end, we're all just looking at a forecast, but how you arrived at that number could be a little bit different depending upon the suite of accounts that you're responsible for. 

 

Chris Platzer 0:24:03.4: 

Right, so it's not forcing a standardized approach exactly, but we're enforcing a standardized output - which leads to an understanding. 

 

Dana Therrien 0:24:14.2:  

Yes, I remember 15 years ago when I was at Akamai. We had developed a forecast model like you're talking about, and it was largely a usage-based business as well. They're just a highway for the internet to bypass all the tolls and things like that. We were able to create a forecast that was generally right. But you had to take out those biggest customers. We called them big bit pushers at that time because they could create wild swings in what was going to happen. But everything below a certain level, it was just pretty mathematical. Some were bursting, some were not. Is that sort of like that with AWS, where a lot of it's predictable but some of it can go wild swings based upon what's happening in those businesses? 

 

Chris Platzer 0:24:58.1:  

Yes, there are macro events. 

 

Dana Therrien 0:25:00.3:  

Macro events? That's a great idea. 

 

Chris Platzer 0:25:02.4:  

Without question, there are a lot of things happen in this world and in the internet and cloud usage. We see it, so there are big swings in certain parts of the world, certain businesses, yes. 

 

Dana Therrien 0:25:19.2:  

So, what do you think you've learned? Well, share some learnings with other people who are embarking on this journey now. What advice do you have to other people who are about to do, or are currently doing, what you're doing now? Just some honest, frank advice. 

 

Chris Platzer 0:25:32.1:  

Honest advice? 

 

Dana Therrien 0:25:34.4:  

Not that I wanted dishonest advice. Unfiltered. 

 

Chris Platzer 0:25:39.4:  

You're recording this, right? Okay, I'll take the gloves off for a little bit. I'd say engage with your stakeholders. Ensure you know what they want and try to identify - work backwards from the customer's need but also try to identify where you can improve their experience in ways that they might not see. They won't know what technical capabilities or functionality you can drive through Anaplan. So, if you can see around corners and deliver added value, you've got a great opportunity. Your stakeholders may not know what are the possible, so I would say that's a massive opportunity to deliver. 

 

Dana Therrien 0:26:25.3:  

I'm happy you didn't say, 'Don't do it.' You see the value of it and it's really about educating the people that you serve and helping them understand how they can leverage Anaplan to solve some of the problems. I know one time when we were speaking, you mentioned just the time to value, like the time from identifying a problem to developing a solution and then executing on that. What's that like? 

 

Chris Platzer 0:26:51.4: 

I'd say that's the best experience I've had with Anaplan and my stakeholder community to date. We've had teams engage with us, 'Oh no, we have a problem. We're going to have to do partner - our partner team needs targets.' We've had to redesign how we're operating. We were able to deliver them a model in a week and launch with enablement. We've launched - I don't have the number offhand - but countless models within one-week and two-week timeframes, consistently. 

 

Dana Therrien 0:27:30.3:  

What's the reaction from the customer when you're able to do that? You must look like a miracle worker. 

 

Chris Platzer 0:27:36.4:  

Yes, setting expectations too high, probably, for continuous development is one thought, but it's a bit of surprise but also a lot more asks for iteration because a lot of these one-to-two-week timeframe solutions, we are scoping for an MVP - in total honesty. We're scoping for a minimally viable product and ensuring they're able to handle their operations. Then building out a backlog where we can iterate and improve and look around those corners, as I mentioned earlier. 

 

Dana Therrien 0:28:13.3:  

Well, Denise Oliver mentioned that at ServiceNow; they had model owners, people who are responsible for almost product lifecycle managers. Are you guys doing something similar like that? 

 

Chris Platzer 0:28:24.4:  

In a sense. I'd say our team is engaged across models, one for redundancy. We probably have a primary owner. You could say that; there's a primary owner. We iterate, along with our partner team that I did mention handles communication, stakeholder management, feedback loops. There are owners on that side that handle some of that product management that we collaborate with. So, we get both sides. Feature enhancement from our side, and then the stakeholder feedback from theirs. 

 

Dana Therrien 0:28:59.4:  

Lots of conversation about, well, what's the ideal profile for a model builder? Me personally, having managed this, I personally like somebody who's managed the business before, who understands the business, who understands things from the inside out and has an aptitude and a willingness to learn Anaplan as well. Because to me, they're bringing that experience to that tool and then they can help to shape the tool. What's your philosophy on it? 

 

Chris Platzer 0:29:25.2:  

Ideal state, a little bit of both. Having the ability to understand some data engineering elements is always nice. 

 

Dana Therrien 0:29:33.2:  

Yes, that's the truth. 

 

Chris Platzer 0:29:35.3:  

Typically, I mentioned this partner team that we have. Quite a few members of my current team came from that partner team and picked up Anaplan. So, they had that business knowledge, understanding of how revenue operates, understanding how our stakeholders view requirements, what's important to them. Transitioning them to Anaplan gave them the power they need to actually influence those mechanisms in the tooling. 

 

Dana Therrien 0:30:04.4:  

Well, Chris, we have a few minutes left for some questions and answers. I know that you've got other members from Amazon out here, too, and maybe they'll use this as an opportunity to ask you some tough questions.  

 

Chris Platzer 0:30:13.9: 

Oh boy. 

 

Dana Therrien 0:30:14.0: 

I'll open it up from the audience to see if anyone… 

 

Audience 0:30:23.3:  

Hey. Alex from Amazon. Good to see you again. I think we've met once before. I'll start with apologizing. I think I'm one of the people that are causing ripple effects that are hitting you up left and right. I know our finance partners are talking to you - I think in real time. Are they calling you regularly? As you probably already know, a lot of what we do on the Game Studios side - so I'm head of business operations on Amazon Game Studios - is interfacing with finance. Finance at Amazon, at least in some capacity, leverages Cognos. Cognos works for all of Amazon so it works for no individual business; it sort of works for everybody - sort of. We are attempting to take away some of the high-effort, low-value work, redundancies, etc. But we've got to do that in partnership with finance. We do not want to steamroll them and have something that exists or work in a silo that doesn't leverage the end-to-end connected planning. 

 

Audience 0:31:37.0: 

So, you talked about outputs. How do you ensure, knowing that finance ultimately is going to be the recipient of the plans that you receive, but they also own the chart of accounts and the formal hierarchies and cost centers and project codes? You want to keep up to date with that. I'm sure there's a line drawn somewhere where there's not a two-way flow. If I know anything about Amazon - I don't know your particular situation - but there are a lot of controllership in place. So, how do you ensure that there's not that disconnect between the lines of controllership, where they want a human being to be that intermediary, versus all of a sudden now you have apples and oranges? 

 

Chris Platzer 0:32:26.1:  

Apples and oranges. I wouldn't be transparent if I didn't say we have scenarios that we deal with that problem currently. For forecasting, we transitioned it to finance. However, what we've done to remediate some of that is have shared datasets to facilitate both in essentially what you said is a two-way door, and identifying a source of truth and who will own that source of truth to complete the cycle between these, essentially, models between Anaplan and the data warehouse that finance is leveraging for this official - what we call a WBR; weekly business review. Did I answer your question? It's a singular source of truth that we use to facilitate it. 

 

Audience 0:33:13.3:  

Yes, just to regurgitate so I follow, or make sure I understood. You have reached an agreement with your finance partners to store data outside of, let's say, Cognos in some sort of staging area that we have all mutually agreed to keep up to date. We're all on our best behavior and we have documentation to support that. Then that is the source that you can pull through S3 or Redshift or whatnot.  

 

Chris Platzer 0:33:42.6: 

Exactly.  

 

Audience 0:33:43.2: 

I have one more question. I realize I have talked a lot. SOX compliance. We're getting into it as our business is scaling and there are a lot of interesting features in Anaplan that, I don't know, could be pretty supportive. So, pushing for approval, the change logs, all those things I think would be really helpful for that. Do you use it at all for that purpose? 

 

Chris Platzer 0:34:10.6:  

For approval work? 

 

Audience 0:34:13.4:  

Well, for just SOX compliance in general. So, I change the forecast; what are the assumptions? Who did it? Who approved it? All of that being archived, so to speak. In our world, I guess there are revenue recognition implications because of effective rate amortization - which I'm not going to talk about in detail - but it does impact. So, it is something that we need documentation around. It may be different in your world, but curious if SOX ever came up as a use case. 

 

Chris Platzer 0:34:47.3:  

It hasn't come up as a use case because this is essentially, what we're sharing to finance on sales perspective of the forecast. However, to your question: we are able to log every entry, every adjustment. We are able to log why judgment was applied in some scenarios. Ultimately, landing on everything in aggregate to what we provide to finance. 

 

Dana Therrien 0:35:17.1:  

I think I might be remembering this correctly - and Denise, I don't want to put you on the spot - but I think with ServiceNow, years ago you guys created a SOX-compliant territory and quota process. Am I remembering that correctly?  

 

Denise Oliver 0:35:29.0: 

Yes.  

 

Dana Therrien 0:35:29.5: 

Yes, so that is one of the only times I've seen a company do that, because the territory and code of process is generally a little bit arbitrary for companies. There are not a lot of controls in there, but they created a fully auditable territory and code of process that was so tight that it would that it would pass SOX compliance. So, Denise is right back there, too. I don't want to put you back on the spot in front of everybody, Denise, but I was impressed enough where I remember it four years later. I think we have time for one more question from the audience. I know it's been a long day. More Amazon. 

 

Chris Platzer 0:36:11.9:  

More Amazon. 

 

Audience 0:36:18.0:  

I don't want you to feel lonely. Keegan, from Amazon Games. I've heard through a lot of examples that Anaplan helps to cut down on data wrangling costs. But there is some initial wrangling when it comes to actually corralling your data pipeline so that you can put stuff into Anaplan, and then taking the results out again. So, what does that initial opportunity cost that you have to pay, like in engineering and, I guess, business transition to get people on there for the first use case? How much do you have to pay before you start getting the benefit - roughly? 

 

Chris Platzer 0:36:57.1:  

Great question. I'd say for our first use cases, we had the benefit of having a mechanism already in place. So, there was a current state mechanism, for example, where we already had some data developed for the use case to be managed currently. That's where we started. I think we had a similar example earlier where I said, 'Hey, we took the current state and move forward.' That's what we did as well so we leveraged the same queries, the same tables in Redshift, pushed them to S3. Sorry for all the AWS lingo. For those that are familiar, we pushed it to S3 and then fed that into our Anaplan instance through an API Airflow [?job 0:37:40.9]. As that went into Anaplan, we had the structure in place, but the opportunity cost was not high because those tables existed already, thankfully - apologies again for the AWS lingo - for the infrastructure, yes. 

 

Dana Therrien 0:37:58.1: 

Yes. Well, I know I work with - I get a chance to work with a lot of companies. On the annual planning process, especially in sales, when they start to go through it, one of… Even though they do it every single year, oftentimes you have new players involved and assumptions change. So, the first question they ask is: where are the data? It's often contained in multiple places, and then how do I get the data? So, then who do I call and how do I negotiate all this stuff happening? Then how do I analyze the data and then how do I keep it updated? Then how do I share it with all the participants involved in the process? Then they'll go off and build models and spreadsheets, the same ones that they built last year, but they'll go in and nobody can find those spreadsheets. They'll go through the process of doing it all over again. I think of it as just building scaffolding on a building that you're going to renovate. Let's say planning is always under renovation.  

 

Audience 0:38:49.4: 

When you're not using Anaplan, you're tearing down that scaffolding and putting it back up every single year - and there's an opportunity cost in that. So, there's a start-up cost in that. When you created an always-on planning platform and that scaffolding remains in place because the building is always under renovation, you're saving months' worth of work on consolidating all this information and then deciding what the process is going to be and how are you going to share it. 

 

Chris Platzer 0:39:15.2:  

On top of security… 

 

Dana Therrien 0:39:17.2: 

Security. 

 

Chris Platzer 0:39:17.7: 

Secure environment. Version control. You've got all of the disparate versions. I have last year's version; I'm going to use that. Just version control alone can create chaos - and I've encountered it. Not a fun experience. Yes, I agree. 

 

Dana Therrien 0:39:38.3:  

One of the objections that I see coming from Anaplan, like prospective customers, is like, 'Wow, this looks amazing. We really want to do it, but we don't have the time to do it now.' The question is, 'Well, when will you have the time to do it?' I always think back to that Chinese proverb: when's the best time to plant a tree? It was 20 years ago, and then what's the second-best time to plant a tree? Today. At some point in time, you need to get over that obstacle of not wanting to go through that repetitive process of reinvesting every single year in that start-up cost. That's one of the things that's impressed me about AWS, is that you guys have now built this ecosystem of always-on planning that's staying there. Now you just continue to prune it constantly and add on to it and then expand it. Now you've got other business units inside of Amazon that are seeing what you're doing and are very interested in doing the same. 

 

Chris Platzer 0:40:27.2:  

Yes. There's a lot of opportunity to scale our - what we're doing now to additional use cases. So, I'm really looking forward to that. 

 

Dana Therrien 0:40:34.3:  

Chris, what departing words of wisdom do you have for this wonderful crowd that's joining us today? 

 

Chris Platzer 0:40:39.3:  

This is a wonderful crowd. I would say parting words of wisdom are: definitely, again, look around corners. I'm going to assume that many of you have looked into Anaplan, into leveraging Anaplan within your organizations. Having an environment that's multi-dimensional, that's secure, that you can engage with your customers to drive efficiency, should be your focus. I think that's where you should start. Engage with your stakeholders, share your vision and move forward. 

 

Dana Therrien 0:41:13.3: 

Great. Well, thanks so much, Chris, for volunteering to do this. We really appreciate it. We loved hearing your story, and thank you all.