Helpful Techniques to Launch Your Spend Analysis Project

This is the second in a multiple part series on Spend Analytics

Last week we introduced some next practice thinking around Spend Analysis.  This week we will talk about some of the techniques that can help launch a successful Spend Analysis project.

Almost every sourcing and purchasing professional appreciates the benefits of good clean data that can be readily manipulated.    However many organizations are intimidated by the anticipated time and resources needed to pull together a successful Spend Analytics (SA) program.  This week we will talk about some of the techniques that “Mpower” a successful spend analysis.

Model the Complexity Before Dealing With Data

Spend Analytics is a complex beast.  Information literally tumbles in from all directions -finance, purchasing, sales, transportation, manufacturing, suppliers, and customers.  It is important to prioritize the data flow before data is extracted, enriched, and stored.  Available data needs to be objectively graded for value, complexity, and validity.  Data that is difficult to manipulate, dirty, low value or unreliable should be put on the back burner until “easier” more valuable data is fully utilized.

Below is an example that profiles two very different companies.  One is a mid-sized service organization with a limited geographic footprint.  The other is a multi-national manufacturer.

Company A doesn’t capture data at the PO/item level.  Since there is no transaction level data they will have limited SA reporting.  More manual systems and reporting should suffice for data analysis.  Structured interviews can backfill missing purchasing data and round out the opportunity assessment.

Company B operates in a more complicated global (multiple languages, currencies, GL’s, etc.) environment.  Global organizations have difficulty getting comprehensive Spend Analysis projects launched due to independent “fiefdoms” with different priorities, organizational structures and objectives.    Ensuring a consistent version of the truth requires a significant investment in upfront governance.  Company B’s desire to systematize the process also means tight collaboration with IT and finance to create a plan establishing how supply chain data “foots”, who will “do” the work, and who “owns” the program.

Eat the Elephant One Bite at a Time

Many Spend Analysis systems companies sell the sizzle of fully-automated SA systems. Having spend data auto-classified and enriched into canned and customized reports is a powerful feature.  We would caution anyone contemplating this level of investment to consider how often the information will be accessed and how it will be used to improve results.  There are many spend reports that are simple to generate yet provide tremendous insight:

• Spend by supplier
• Spend by location
• Spend by item/commodity/material
• Spend by transaction
• Combinations of the above

Each of these reports involves a limited amount of data integration to support the report.  Before companies focus on adding new reports and associated data it is important to invest in training users in how and when to utilize the basics.

Understand the Limitations of Spend Analysis Solutions.

Spend Analysis helps us understand how money was spent in the organization.  Before significant dollars get invested in mining the past, equal dollars should be allocated for understanding how suppliers are performing in the present.  Some areas to explore:

• Investing in basic supplier scorecards and KPI tracking
• Enforcing service level agreement (SLA) performance
• Tracking and monitoring “realized” vs. “booked” savings
• Capturing the value from supplier-led initiatives

For organizations that are resource constrained, customer and stakeholder surveys provide a cheap way to understand the “bang for the buck” that suppliers provide.  As we’ve pointed out in previous articles, low-cost, poor service providers can be more damaging to an organization than higher cost suppliers with stellar service.

Check back next week for part three.