The purpose of this post is to provide a high-level comparison of IBM Planning Analytics and Anaplan. The comparison represents a Point of View (POV) of Application Consulting Group (ACG) based on our experience with both platforms and general understanding of the respective tools’ capabilities based on market feedback and experience. This comparison is not complete or exhaustive of all differentiating points, nor does it dive into significant detail on the individual differentiators. The objective is to outline the key differentiating concepts that, along with other due diligence, will help companies make an informed decision on which is the most appropriate platform to use for their specific needs.
This comparison consists of the following categories:
- Historical context
- Platform Scalability
- Modeling Flexibility
- User Interface
- Deployment Options
- Market Positioning
- Cost (Licensing, Implementation and Support)
- Future Outlook
First, a walk down the memory lane
The roots of both these platforms date back to the 1990s with Adaytum Software and Applix. Both were founded around that time and provided analytics and planning capabilities based on multi-dimensional technologies. Adaytum tried to create a product / brand for themselves, but it was ultimately acquired by Cognos Software in 2003 who leveraged the technology for their Cognos Enterprise Planning platform, which was pretty successful. Over time, however, Cognos realized that the EP software based on Adaytum has challenges with scalability and was not able to effectively handle and analyze increasingly large and complex financial performance data.
Cognos went out and purchased Applix in 2007 to replace EP with Applix’s 64-bit in-memory analytical engine called TM1 that provided a much more scalable platform. Until then, Applix was an independent company that created a product called Table Manager 1 (TM1) to streamline and automate the consolidation and analysis of financial information. As an in-memory system, its growth was constrained by the hardware (memory) limitations at the time.
Cognos was acquired by IBM in 2007 (shortly after Cognos bought TM1) to boost IBM’s Business Intelligence and Performance Management capabilities. A combination of IBM’s sizeable sales / marketing footprint and the increased availability (and lower cost) of memory drove a huge boost in adoption of the system, including with large enterprise customers. IBM made a significant investment into the system on both the front-end and the server and in 2017 renamed it IBM Planning Analytics, which is what the system is today.
In 2012, the original founders of Adaytum came out with a new venture called “Anaplan”. Anaplan was a business planning software that was fundamentally based on the Adaytum technology plus some additional integration components. They made the application user friendly with the idea to offer it as a subscription in the cloud. Both the idea and timing were genius, they were pretty much the first to come out with a cloud planning tool when the whole world started a shift to the cloud a rode a pretty steep wave of growth to become a well-established platform on the market. That is fundamentally where Anaplan still is today.
Key Differentiator - Scalability
Probably the biggest and most significant differentiator between the systems is platform scalability.
Anaplan, being fundamentally based on the Adaytum technology, is not able to scale at the individual cube level. The technology has a limitation in terms of cube size, which we believe is around 100 billion intersections in the Anaplan Hyperblock. That seems like a lot, but it is a fraction of what is required for a multidimensional system of even a modest size, especially if you do not actively manage sparsity (empty cells).
Therefore, to support a meaningful business application, Anaplan had to divide its solutions into many small cubes and string them together. The term “Connected Planning” is not a virtue, it is a necessity to handle the lack of core platform scalability. Anaplan did a very nice job packaging this all together in the offering. They created integration mechanism between the cubes and provided users and developers tools to manage the process. Fundamentally, however, the system can only handle so much, plus the fragmented design increases solution complexity. Anaplan would struggle to support a true enterprise performance management process at scale. It is typically deployed as a purpose-driven application or a department level solution that supports a specific process and it works well for that purpose.
IBM PA has no practical limitation in terms of cube size or dimensionality. TM1 was built as a memory-based product and with the Microsoft Windows 64 bit architecture (which has a theoretical limit of 16 million terabytes of memory) there is no real limit on the size of the system. IBM PA can hold extremely large and complex data models all in memory and maintain good usability and performance. Some of our customers have massive 15+ dimensional cubes with over 100k (some over a million) members in dimensions, they cube sizes often exceed 5TB. There are obviously considerations that need to be taken into account when deploying these large systems but, assuming appropriate design and build, the system will scale and perform just fine.
The above difference in scalability has a direct impact on the flexibility each tool provides for modeling.
Given the need to build Anaplan as a solution of multiple connected parts, the company provided some “guardrails” to do that in an effective manner. They separate the user / developer from the deeply technical tasks / requirements by using configuration options and tools. The result is a system that is pretty intuitive and easy to use for development and maintenance. The work is more configuration using pre-defined menus and options, many rules and calculations are created using a GUI. This user-friendly experience with not-too-steep of a learning curve has been a major factor in the popularity of the solution and its growth. It was created an opportunity for large integrators to scale up their consulting groups with Anaplan and push the platform as part of their business solutions, which also contributed to Anaplan’s growth.
This usability benefit of Anaplan is offset by limited modeling flexibility. Because you work with largely pre-defined menus and options, as well as the need to manage multiple connected solution elements, there is limited flexibility to build a truly customized solution that will flex to any process or requirements. Anaplan will satisfy the core needs of most customers based on typical requirements that are based on a standard business process. However, the platform will be constraining if trying to satisfy any reasonably complex and custom requirements at scale. When we talk to users of Anaplan of how the application is deployed, in most cases it comes down to them doing most of the work offline and just loading results into Anaplan for consolidation r just doing pretty simple input with very little complexity.
IBM Planning Analytics, on the other hand, does not have any constraints or limitations. The platform will support any permutation of configurations and will wrap itself around the customers’ process requirements, whatever they are. It supports any number of dimensions, elements, hierarchies, real time calculations or defined processes, transactional and operational metrics, no practical limitations. There are no pre-defined or required structures, it is truly free form design.
The downside of this is that it requires a fair amount of technical knowledge and business acumen to build an effective solution. Because it is so open, it is critical to get the design right and implement using best practices for the solution to work. A well designed solution will facilitate a truly real time interactive solution that will very effectively facilitate a distributed business performance management process at scale. A poor design or build will quickly make the system sub-optimal and not user friendly.
User Interface and use of Excel
Both system have a pretty robust web based interface. They work slightly differently and I will not get into the nitty-gritty of the clicks, features and functions. Both offer capabilities for visualization, dashboarding, input capabilities, workflow etc.
The major difference here is the lack of live and interactive Excel interface by Anaplan. I think this is where all the new Cloud vendors in this space got it wrong. I am sure there is a cost befit to not having to support an Excel add-on and deal with the various versions of the SW and periodic updates. But the assumption that people will not need to use Excel is simply incorrect. We all listened to statements declaring the end of Excel since the early 2000s and as of today the tool is still a critical enabler of analytics and reporting. Excel is here to stay, there is nothing that will replace the flexibility of Excel for the finance or business analytics user.
IBM PA is one of only a few vendors that provides a fully functional live Excel interface with a live connection to the tool and all the capabilities of Excel. The Excel add-on keeps evolving with a new release about every 1 month and it is an important item on the IBM PA roadmap. While Anaplan does provide an Excel capability, it is not live / interactive and it is not part of the ground-up design of the application. I think this is where IBM PA and the other tools with an Excel option have a huge advantage.
IBM PA is the only mainstream vendor on the market that provides both a Cloud and an on-prem deployment capability. Each customer usually has a clear idea where they want to be so not a big difference to the individual user, however the dual option presents an advantage from a go-to-market perspective.
Given the “application” focus and the notion of pre-built models, one would expect that Anaplan is cheaper or faster to implement and operate. We do not see that to be the case, necessarily. We break down the Total Cost of Ownership (TCO) into the following three components:
Based on everything we see in the market, IBM PA is the more cost effective system to license. IBM PA offers a single license that includes all capabilities and tools, there are no extra fees for integration, applications etc. The licensing models can get complicated so it is hard to do an easy side-by-side, however in every single situation we have been involved over the last few years, Anaplan ended up being more expensive, or at least at-par with IBM PA, from a pure licensing perspective.
This cost element is highly variable and depends on many items such as data quality, number of integration points, complexity of calculation and methodology etc. What we find is that the “application” level readiness of Anaplan does not really create an advantage over IBM PA. You still have to get the data ready, connect everything, build your core structures and outlines, hierarchies, configure rules etc – all that work has to be done regardless of the system, the difference between the platforms is more in how you go about it. We had IBM PA applications that were deployed in weeks and some that took years to get going, I know the same is true for Anaplan. The only way to properly evaluate the implementation costs is to dive into the detail and design. Doing a pilot or Proof of Concept (POC) is a good practice to flush out the specific capabilities and assess the approach and the ability of the platforms to handle them. However, we do not find the pre-built nature of Anaplan models to be an obvious advantage in terms of implementation cost and effort.
The maintenance overhead is a bit of a function of how the application is deployed and delivered. Both systems are generally deployed and managed by business / finance, (or business aligned IT) vs traditional IT groups. A well-built IBM PA system will include automation and administrative tools that make it easy to manage the system. The user facing tools are self-serve for the most part so the ability to manage the system by users is more function of training. The nature of Anaplan’s core design does provide out of the box tools and routines that are pretty intuitive to the user so assuming proper deployment the two platforms are very similar.
Anaplan is clearly focused and dedicated to business performance management with a huge emphasis on FP&A as well as other related processes. The company has a defined marketing strategy and sends a clear message to its target audience. It has a dedicated salesforce that is generally knowledgeable and specialized in the space. Anaplan invested into applications that sit on top of the core technology across fiancé and other business performance management areas. This allows them to appeal to CFOs and other business buyers with a clear and articulate message. Customers know exactly what they are getting and that provides comfort during the buying process.
IBM, unfortunately, does not have nearly the same level of focus and dedication from a sales and marketing perspective. IBM sales reps cover a portfolio of tools that are not clearly related to each-other. Most reps are inexperienced and do not understand the EPM space. As a technology company, IBM sells to IT vs the business where these solutions are driven from. IBM did not make the same investments into the application layer that Anaplan did (and is not planning to as of now), it positions IBM PA as a platform that can be applied to any use case. As a result, they are not able to control the content and message. It leaves the outcome of any sales cycle at the mercy of who does the talking, demoing etc. The company is not sending a clear and consistent message to its buyers via marketing, trade shows or other means. This causes IBM PA to often be misunderstood, the buyers often do not know that IBM PA plays in their space, they are unclear what IBM PA does and are unsure if it will meet their needs. This leads to a huge opportunity cost for IBM and the platform.
To be clear, the above is just a matter of positioning and marketing of the products by the vendors. It does not have anything to do with the respective platform capabilities and the value that either provides to the user once deployed.
Future Outlook / Commitment
Despite the lack of IBM investment in applications, which is a conscious decision, the company continues to commit significant resources to develop the platform and improve the back end. The rate of investment and innovation has been very steady over the last 5 years and there is a robust product development roadmap in place. The IBM PA install base includes some huge global enterprise customers and IBM does not have the luxury of slowing down progress even if they wanted to. They have to invest, and they are.
Same has been true for Anaplan, clearly there has been significant ongoing investment into the system by the company, regardless of the ownership scenarios. Things may get a little more constrained after the 2022 acquisition of Anaplan by Thoma Bravo and the recently reported layoffs. This is an evolving situation that will have to be watched.