Anaplan doesn’t come up in many conversations about unicorns – you know, the Ubers, Airbnbs and SnapChats of the world. But if your company is struggling with strategic planning, Anaplan is one unicorn you’ll want to learn more about. San Francisco-headquartered Anaplan’s cloud-based platform replaces spreadsheet mania as a way to handle large-scale strategic planning in real time. Companies such as HP, Intel and Morgan Stanley are streamlining sales management, forecasting and a variety of other applications using Anaplan, which simplifies life for everyone from senior managers to model builders and end users.
[ Related: Cloud planning app rescues retailer from ‘Excel hell’ ]
Anaplan founder and CTO Michael Gould tells IDG Chief Content Officer John Gallant how the platform’s in-memory calculation engine can support dynamic planning for even the biggest companies, and why being built in the cloud gives Anaplan a leg up on traditional planning apps from rivals like IBM and Oracle. Gould also discussed Anaplan’s growing app marketplace where a variety of new tools are helping customers improve planning across the company, and how this unicorn is capitalizing on all the venture investing it has earned.
[ Related: CIOs embrace hybrid cloud and software-defined data centers ]
CIO.com: Why was Anaplan formed? What problems were you setting out to solve?
Anaplan founder and CTO Michael Gould: I started Anaplan 10 years ago. My background was with Cognos and IBM and I recognized the challenges of decision making in organizations. We’re trying to get away from the planning process being just an annual exercise and driving it towards decision making, a much more immediate course-correction process.
I’ve been in enterprise planning software for most of my working career and the products that were out there were pretty old. They date back to a mainframe system I’d worked on as an intern back at IBM in the mid-80s and the other products around had a similar vintage. I recognized the need for a new platform in large measure to address the problem that people were reverting to spreadsheets for almost all actual decision making in an organization. They would pull the data out of the existing planning systems, but do the real work in spreadsheets, send them around, collaborate over email and then, having made a decision, post the results back into the system as the means of recording, aggregating and reporting those decisions.
Anaplan was intended to address that problem for organizations, to give a platform for running the decision-making process.
[ Related: Anaplan ramps up in Australia ]
CIO.com: It’s a cloud-based solution today.
Gould: That is correct. We launched as a cloud-based platform.
CIO.com: Can you help readers understand the key elements of your Anaplan system and how they work together?
Gould: At the heart is an in-memory calculation engine designed to support the same kind of interaction and immediacy you get working with a spreadsheet; you can enter data and immediately see it recalculate. You can adapt and change. It’s very versatile in terms of different problem spaces that you can apply it to. On the one hand trying to keep the agility and versatility you get with spreadsheets but with enterprise scale. Enough structure, enough data volume handling, enough performance to be able to run the kinds of modeling and planning activities required for the largest companies in the world.
CIO.com: Can you describe Hyperblock?
Gould: It’s an approach to the way we handle the data in memory and it’s optimized to enable very fast changes to the data. If you think of BI systems, pulling data out of transactional systems and structuring it in a way that you can query and report on it, to a large extent the data isn’t changing in real time.
With the planning application, end users are interacting with it and wanting to see the impact of those changes. If you’re doing promotions planning, you want to create a new promotion, assign some products to that promotion, you need to see the rollup of that in terms of your planned expenditure across different customers, across different product lines and so on.
Hyperblock is the way the data is structured internally, in memory, in order to allow very fast propagation of changes through the system, both in terms of data entry, putting in new data values and seeing them recalculate, but also in terms of some of the structural changes, creating a new promotion, adding products to that promotion. Those are opening up new slots within your data structure that you then need to recalculate. The Hyperblock architecture is really how we optimize for those interactions, that very fast change, both in terms of data and in terms of structure.
[ Related: Tech Titans Talk: The IDG Enterprise Interview Series ]
CIO.com: You strive to make this accessible to virtually anyone and that’s always seemed to be a challenge for planning or analytics tools. What have you done from a user experience perspective to make this useful for lots of different people in the company?
Gould: First, we’ve aimed to keep the user experience unified for everyone from administration to model building to the actual use of applications, both in terms of contribution -- end users who are feeding data into it -- and executives who are reviewing data within the application. They all work with the same web-based interface but just have access to different parts.
For example, there will be some modeling areas that are only accessible if you’re a model builder. But the real point is to keep that a single, simple experience. In many ways that’s building on the success of the spreadsheet as a ubiquitous tool across organizations large and small. If you’re a senior manager or executive reviewing data, you go into it as if it were a simple dashboard and that’s all you expect to interact with within the platform. End users contributing to the planning process will interact with certain screens that guide them through the tasks that they need to contribute and then there are other parts of the platform that become available depending on what your access is.
CIO.com: Whom do you sell to and who brings it into the organization to start?
Gould: What we’re looking for initially is a particular pain point that is a sufficiently high priority that there’s a lot of attention to it within the organization. That could be in financial planning, sales operations for quota planning or sales forecasting. It could be in supply chain for demand forecasting. We’re looking for is a process that’s either broken or inefficient and there’s a real need to address that. We work outwards from there as people grasp the potential applicability of the platform across multiple use cases within one function and then across multiple functions.
CIO.com: What typically is the role of the IT department in the deployment of this?
Gould: We aim to engage with IT from the outset. The design of models is often most appropriately done by the business analyst, people within the business who understand the domain and the problems they’re trying to solve. But at the same time you have to have IT involvement for data integration, for security signoff and for the governance around the quality of the data.
Typically, we’re working both with IT and the business -- getting collaboration where they understand the roles, where IT stops and the business takes over. We encourage IT to own the quality of the data going into the system because they have the background, the understanding of the data integration process and the governance around that. Also, tying the data back to the corporate data warehouse to make sure it all ties in before planning processes start, making sure you’ve got the right data before you get going.
CIO.com: Are IT departments using this for their own planning?
Gould: We have customers using it for IT costing, tax planning for expenditure, things like that. Typically, Anaplan is geared towards future decision making, as I said earlier. It’s not really designed to be a transactional system.
Actual management of what’s currently going on in your business is not really what we do. We interact with other systems when you’re at the transactional level, but we focus on forward-looking planning.
CIO.com: Can you talk a bit more about the predictive analytics capabilities of the product?
Gould: We have integration with optimization tools for predicting supply chain, for optimizing production across different factories. We have a range of statistical forecasting methods available within the platform. An area of we’re looking at a lot -- not live yet in the platform but a very hot topic -- is machine learning and the use of historical data. For example, customer buying patterns of telecom companies, then feeding that into your planning and forecasting models.
In a sense, it’s not at the heart of what we do in terms of the actual technology that we produce. We’re aiming to integrate with other tools. We’re not building our own machine learning algorithms in order to drive our models, but it’s very closely coupled and something that our customers are very interested in.
CIO.com: Can you talk a little bit along those lines about what are some of your key products and tools out there that you are integrating with?
Gould: We integrate with optimizers like CBC we use for linear programming optimization. At the back-end we would be integrating with Python, primarily to make use of the wide range of forecasting techniques that have been developed, as well as the main programming languages where we see the most active development. Apache Spark MLlib for the machine learning side.
CIO.com: Can you give me an example that shows how Anaplan is being used?
Gould: One of the customers that has been working with us for the longest [time] is Intel Security, formerly McAfee, which became part of Intel a couple of years back. They started with territory and quota management, looking at the problems they had with something approaching a couple thousand sales reps across the world.
Obviously, with a sales force of that size you have a lot of churn of sales reps leaving, new ones coming in and just keeping track of what quota they had allocated to the sales reps. This was a very Excel-driven process. We implemented a solution for them. It went live in just a few weeks. They then came back to us and asked if we could help with their commission calculation process, the crediting rules to take their existing bookings data, how to allocate that to the right sales reps and calculate commissions and then start to model out future commissions, which are a huge expense for any company that’s driven off an enterprise sales team.
From there they spread out across multiple applications in sales operations, into finance and starting to connect all those pieces together. They have close to 30 different applications up and running now.
CIO.com: Is there another use case that you can share with us?
Gould: Another great example is HP. They use us for quota planning. We’re actually working with two different customers there, but they implemented a global rollout of their quota planning, essentially the process of assigning customer accounts to their sales reps across the world.
It used to take them around four months to go through that process. They started a couple of months before the end of the fiscal year, but it wasn’t actually complete until a couple of months into the new year. This whole process was so delayed that they were basically struggling to understand where they were as they went into the new year. We implemented a solution for that with them. They went live within five months start to finish and now are able to have that quota plan in place right at the start of their new year.
CIO.com: You mentioned the old-school competition that was out there when you built this product, but who do you compete against today? Are you typically augmenting existing systems or replacing existing systems?
Gould: Our main competitors are still the big vendors; IBM, Oracle, SAP. To some extent we compete with the use of Excel, but we’re not trying to replace Excel. We’re not a spreadsheet.
Excel is a great tool for personal use. What we say is we’re trying to get away from inappropriate use of spreadsheets for running large-scale processes, planning applications within big companies. It’s primarily the planning applications from IBM, Oracle and SAP but not on the transactional side. We’re not competing against their GL systems or anything like that.
CIO.com: Are any of those offerings cloud based at this point or are you the only company with cloud-based capability for that?
Gould: Some of them have taken existing old systems and created a cloud version of them, but I think that’s a very different proposition to building something for the cloud from the ground up. You can take pretty well anything and put it in the cloud if you try hard enough, but basically you’re running single instances of software and just hosting it on people’s behalf.
Anaplan is fully multi-tenanted. We can run multiple instances within the same server for different smaller clients. Bigger clients typically require dedicated machines just for the same data processing that’s required. I don’t see any of those [rivals] having true, ground-up applications that are built for the cloud.