Supernova Award Category

The Problem

Within the food manufacturing industry, increasing competitive pressure is causing companies to need a more accurate forecast of product demand in order to better understand future costs to maximize profit.  During the forecasting process, large amounts of data are produced that need detailed analysis to understand where the opportunities and risks exist.  Analyzing this data can be difficult and time consuming which poses a significant challenge when business decisions need to be made under a tight time line. Businesses need to understand the input costs, when the product needs to be made, and costs to move the product throughout the supply chain to its end destination.

To better understand these costs, ConAgra Foods needed a solution that allowed deeper analysis on its businesses, while shaping future outcomes by more accurately predicting future events. These modeling capabilities would allow general managers to see the impact of business drivers on the margin and profitability of a change prior to implementing. Not only would this help the general manager, but it also provides other business areas a better understanding of the total delivered costs. To allow for this, the end solution needed to be able to process large amounts of data at detailed and aggregated levels, while still allowing for visualization of the data and exception indicators when the unexpected happens.

The Solution

To accomplish this ConAgra partnered with SAP to co-innovate on a Total Margin Management (TMM) solution that provided costs at the lowest level of granularity, and the forecasting capabilities to model scenarios to better predict the future. This solution comprised of two components, TDCA and MMA powered by SAP HANA. Total Delivered Cost Analytics (TDCA) allows general managers to do detailed analysis on cost inputs and visualize the information to make informed, responsive decisions. Margin Management and Analytics (MMA), provides the ability to then take this broad cost decomposition, along with other detailed information for volume, trade and A&P, and run analytics to model future costs and revenues, ultimately allowing for the modeling of profitability at the product/customer level. With this information, general managers can generate a more accurate forecast in less time with less people. This enables ConAgra Foods to meet changing business needs in a flexible manner.

The results

- Improved processes and shortened forecast cycle times while reducing efforts needed.

- Leveraging technology for direct source data feeds, flexibility/usability through SAP HANA and analyzing Product & Customer profitability

- Increased speed of decision making through customized analytics and modeling business scenarios

- Significant Performance Improvement -- Prior to TMM, running a detailed forecast bridge was impossible without letting the overnight processes pre-calc all scenarios. - With TMM, we can now run detailed bridges dynamically in under 60 seconds.

Level of Detail Improved -- Prior to TMM, system generated P&L's were generated at a higher level of the product and customer hierarchies and needed manual analysis to get to low level details. With TMM, the system will allow general managers to quickly get to detailed P&L information in a matter of seconds.

- Understand the true cost and revenue drivers at product/customer level at the lowest level of granularity

- Respond faster to risks and changing market conditions

Metrics

Data Compression -- Prior to implementing HANA, data needed to support TMM was 6TB. With HANA and TMM, we have been able to reduce the data size to 800MB. 

Significant Performance Improvement  - Run detailed bridges dynamically in under 60 seconds.

The Technology

To analyze the huge amount of data & make decisions in near real time, we looked to SAP HANA as in-memory platform. Calculations that used to run in batch, can be returned instantly &  data analytics became faster. Users quickly perform iterative scenarios & “what if” forecasting on large data sets which  increases the reliability of the margin forecast &  produces a single forecast that is risk adjusted &  applicable to multiple business areas

Disruptive Factor

TMM will enable visibility & ease of analytics of product costing and product and customer P&Ls. The main consumers of this information are general managers & finance. TMM will have approximately 300 users which will have the ability to dive into the details of various lines of the P&L through seamless integration from the source system. The business will also have the ability to model customer profitability and predict future opportunities. Users want simple and powerful tools. The frond-end is very easy to use; almost like a pivot table in Excel allowing users to very quickly retrieve data which significantly cuts work out of finance processes. Instead of using paper reports with outdated data and users looking backward trying to digest and reconcile data, now the data is much more integrated and users have access to data to see specific customers as well as all the way down to the material components that go into making the food products.

Key to driving transformation is close collaboration between business & IT through open communication to address new challenges/opportunities. It’s important to understand the direction that each other wants to go and make sure we are making the partnership work. Aligned objectives and making a commitment to the project helps. The project enabled us to do more with the data we have and better understand the data we did not have. We see many new opportunities to leverage the power of SAP HANA and Big Data

Shining Moment

Within weeks ConAgra was able to save money & create a better forecast. In first forecasting cycle we identified 2 big wins. By leveraging greater detail in costs related to transportation, ConAgra was able to work with certain retailers to modify certain CPU rates saving thousands.  Also, within the first forecast cycle, ConAgra was able to create a better forecast by going to a lower level of detail on conversion rates which lead to a more accurate volume forecast