Anaplan vs. TensorFlow

I've implemented a number of projects for Anaplan for Sales Performance Management use cases. It is obviously built for financial planning, but it allows for a lot of flexibility for territory and quota, ICM, sales forecasting, and other important use cases. Territory and Quota is very powerful in the tool as it organize complex assignment structures into hierarchies for easier analysis and reporting.

TensorFlow is great for most deep learning purposes. This is especially true in two domains: 1. Computer vision: image classification, object detection and image generation via generative adversarial networks 2. Natural language processing: text classification and generation. The good community support often means that a lot of off-the-shelf models can be used to prove a concept or test an idea quickly. That, and Google's promotion of Colab means that ideas can be shared quite freely. Training, visualizing and debugging models is very easy in TensorFlow, compared to other platforms (especially the good old Caffe days). In terms of productionizing, it's a bit of a mixed bag. In our case, most of our feature building is performed via Apache Spark. This means having to convert Parquet (columnar optimized) files to a TensorFlow friendly format i.e., protobufs. The lack of good JVM bindings mean that our projects end up being a mix of Python and Scala. This makes it hard to reuse some of the tooling and support we wrote in Scala. This is where MXNet shines better (though its Scala API could do with more work).

Incentivized

• Anaplan removes the time consuming process of integrating the results of individual spreadsheets.
• Anaplan facilitates the standardization of assumptions across all sub-processes
• Anaplan provides full transparency of the calculations and source inputs
• Anaplan allows us to automate certain planning processes that would have been impossible when relying on the computational capabilities of an individual computer.

• A vast library of functions for all kinds of tasks - Text, Images, Tabular, Video etc.
• Amazing community helps developers obtain knowledge faster and get unblocked in this active development space.
• Integration of high-level libraries like Keras and Estimators make it really simple for a beginner to get started with neural network based models.

Incentivized

• Some end users don't spend the required amount of time to familiar themselves with Anaplan.
• Auditability, while good in theory, can be lacking and clunky when needed.
• Model building can become convoluted over time, leading to confusion over the specific use of line items.

• RNNs are still a bit lacking, compared to Theano.
• Cannot handle sequence inputs
• Theano is perhaps a bit faster and eats up less memory than TensorFlow on a given GPU, perhaps due to element-wise ops. Tensorflow wins for multi-GPU and “compilation” time.

Anaplan is a very strong multi-dimensional modeling tool that provides a calculation engine to empower a complex planning process. It is fairly easy to learn for those with experience in similar tools, or excel. It forces structure and auditability that spread sheets do not have, along with extensive security capabilities

As a user it is a very simple tool, but at the same time with a very mature and powerful calculation engine. It is very easy to switch from excel or traditional tools with added capabilities of multi dimensionality and real time calculation engine to see quick insights needed to create plans and scenarios

Support of multiple components and ease of development.

Incentivized

There are very few outages. Maintenance is scheduled on two or three Saturdays per month, so as not to affect businesses. When there is an outage, users are kept informed of progress to restore the platform and typically this takes no more than an hour. Anaplan customer support is very responsive if we ever have questions about platform issues

Everything is calculated in memory in the cloud. It's nearly instantaneous updates when you make changes. The only time things get a little slow is when you have a massive model with very intricate calculations...but "slow" for Anaplan is not what I would call "slow" for something like Hyperion. We used to have Hyperion calcs that ran for 60 mins before you could use data. The equivalent would be 60 seconds in Anaplan.

Incentivized

Support quality has dropped since Thoma Bravo has taken over. I think some serious re-focus needs to happen here -- part of the beauty of being in the Anaplan community was how involved you felt in it before. Before I didn't dread sending a support ticket, now I am starting to.

Community support for TensorFlow is great. There's a huge community that truly loves the platform and there are many examples of development in TensorFlow. Often, when a new good technique is published, there will be a TensorFlow implementation not long after. This makes it quick to ally the latest techniques from academia straight to production-grade systems. Tooling around TensorFlow is also good. TensorBoard has been such a useful tool, I can't imagine how hard it would be to debug a deep neural network gone wrong without TensorBoard.

Incentivized

In my opinion, in-person training is always the best if you have the option to do so. This allows real-time interactions with the instructions, whereas the online training I took required me to write-down questions, email them, and wait for responses. This slows down the process, as you can imagine. That said, in-person training is an extra cost and it likely isn't needed for everyone. I would suggest selecting a small number of people to take in-person training and then having them act as mentors to the rest of your team. That way, as the rest of the team takes the online training, they have a resource to help them in real time.

Anaplan training materials are clear, simple, easy to understand and to follow. Visuals are excellent. The vendor is good at updating training materials in a timely manner and encouraging users and administrators to keep coming back to Academy site for refresher courses or new feature courses. I really like their interactive diagrams

One key insight from implementing Anaplan is that success comes from focusing on designing the process, not just building the model. Anaplan is extremely flexible—there are very few planning scenarios it cannot support—but that flexibility means the project needs strong governance, clear ownership of requirements, and a well-defined data model. When those foundations are in place, implementations are fast, iterations are easy, and users can quickly see value.
In our projects, both Financial Planning and Integrated Business Planning models were adopted smoothly because we involved business users early, kept the model design intuitive, and leveraged Anaplan’s Excel-like syntax and user-friendly dashboards. The result was more efficient day-to-day work, reduced manual tasks, and increased collaboration across teams. In short: when you combine Anaplan’s flexibility with a structured implementation approach, adoption and value realization happen quickly.

Use of cloud for better execution power is recommended.

Incentivized

Anaplan is more powerful than Pigment considering that it is an Enterprise class system and is able to manage bigger data sets. Anaplan allows for advanced scenario modeling and formula capabilities along with custom reporting functionalities. Anaplan has proven its capabilities and stability across various use cases and across bigger enterprises when compared to Pigment which is still in earlier phases of its development

Keras is built on top of TensorFlow, but it is much simpler to use and more Python style friendly, so if you don't want to focus on too many details or control and not focus on some advanced features, Keras is one of the best options, but as far as if you want to dig into more, for sure TensorFlow is the right choice

Incentivized

We have managed to leverage Anaplan for financial planning and forecasting across the business. It is now used by almost every department, with more than 50 users (but I know of companies that have hundreds of users) and still the platform is quick and reliable. It is easy to make changes to divisions and departments or add users and apply different user settings - the core part of the model is not affected and end users can continue their work without any disruption

• Anaplan's implementation led to a significant reduction in planning cycle errors and bugs, streamlining processes and improving overall accuracy in data inputs
• Standardizing the planning process and enabling cross-functional collaboration through Anaplan enhanced our ability to adapt swiftly to changing business needs, resulting in improved agility in decision-making
• The platform's capabilities, especially in Demand Planning and Supply Chain, positively impacted our ROI by optimizing resource allocation and solving complex business problems efficiently across multiple functions

• Learning is s bit difficult takes lot of time.
• Developing or implementing the whole neural network is time consuming with this, as you have to write everything.
• Once you have learned this, it make your job very easy of getting the good result.

Incentivized