Optimizing Sales Compensation using Big Data

Optimizing Sales Compensation using Big Data

Sales compensation is a world swamped with data. ICM applications must accommodate a wide variety of data from upstream systems, facilitate processing, and deliver results through multiple channels and formats in order to support activities such as territory management, quota setting, modeling, forecasting, commission calculation, and on-demand reporting and analysis.

Today’s sales compensation can benefit from underlying Big Data technology, which can address these demands to run efficiently and reveal valuable insights quickly and efficiently.

Why Big Data?

Big data offers two processing techniques to achieve the desired performance and timeliness - parallel processing and the use of a non-traditional database.

Parallel processing is a way of spreading processing across multiple computers, which drastically improves performance.

Traditionally, processing of large data volumes of structured data is done using a relational database, and most compensation systems are built this way. But this approach can be problematic for commission processing, since it involves many manipulations around the same dataset, inserting data to create credits and then aggregating those credits all at the same time. This can lead to contention and limited performance.

Big Data solutions often use a non-traditional database, sacrificing flexibility for performance. Relational databases work on the premise that data must be assembled in many ways. For example, customer contact details, which may be used for many different purposes, are stored in a separate table that can be joined with other data as required. Non-traditional noSQL databases tend to store all the data required for a particular action in a single record, which is very fast for some actions but slower for others. Given the complexity of sales compensation data, the relational model still has the advantage.

Data timeliness

One of the goals of a sales compensation solution is motivating employees or partners by providing prompt results and payments to reinforce the link between actions and rewards. Yet, processing data and delivering fast results is challenging due to potentially huge ICM data volumes, combined with the complexity of the data.

Using Big Data technology, processing time may be reduced from hours to minutes, affecting the expectations, motivation and efficiency of users. Compensation analysts can perform better planning by modeling, simulating and comparing multiple compensation plans, and viewing results within seconds. Sales reps can be more motivated with daily earning reports and real-time visibility into their compensation. And sales executives have more analytical flexibility with the ability to drill down and examine data from any desired angle.

Data timeliness: With processing time dramatically reduced, sales reps can view their quota attainment in real-time.

Transparency & trust

A sales compensation solution relies on other systems for its source data, such as Sales Ops, Finance, CRM and ERP systems. Often, the data points needed to accurately run compensation is spread across systems and requires cleansing, validation and pre-processing before it can be used.

When data volumes are large, ensuring data accuracy becomes a process-intensive and time-consuming task. It must be run repeatedly to pick up back-dated changes to upstream systems. Getting this done in realistic time scales is often a limiting factor for plan design and the timeliness of corrections.

With the fast processing offered by Big Data, inquiries raised can be rectified in core systems, and corrected results can be made available to payees the next day. Plan design is driven more by results than by data constraints, and results can be more granular to reinforce the link between pay and performance.


Users of Sales Performance Management (SPM) systems constantly need to adjust results retrospectively, in order to rectify incorrect results in a timely manner. Plans that include a quota, ranking or a roll up hierarchy can be problematic. For example, a single back-dated transaction may impact results across the system. Such data adjustments can seriously affect the ability to deliver timely and accurate outputs.

One solution is intelligent recalculation. While not based on big data principals, it is a powerful technique that reduces the impact of backdated changes. If a single transaction from 3 months back is updated, intelligent recalculation would work out which payees and plans are impacted by the change and only submit these for recalculation. This can dramatically decrease data processing requirements and make it feasible to recalculate the system intraday to true up results after a manual update.