In Part 1 of this post, I discussed some of the shortcomings to rely on a Proof of Concept as the main decision point to select a particular suppliers Spend Visibility solution. If you decide a POC is required as part of your selection process, below are some suggestions as to how to most effectively conduct the POC as part of your overall selection process.
Some “Pros” of Conducting a Spend Data Classification Proof of Concept
- Allows sourcing people to throw data over the wall and see who responds.
- Enables companies to see their data reflected in vendor Spend tools, and companies can see how they would interact with their own data.
- Provides an idea of the vendor classification capabilities to a standard classification schema, like UNSPSC.
A Better Process
- Don’t do a POC too early in the selection process, and as a key selection event focus, or you may miss larger approaches to Spend Data Classification that better match your company sourcing programs. Not just UNSPSC classification, but also sourcing category classification that will support your organizations sourcing programs.
- When you provide company data, focus on areas that may provide new opportunity, which helps everyone focus on the best job possible (savings opportunities are good).
- Do a POC to see your data in the vendor tool and prove it out, but define a realistic (but meaningful) sized data set.
- Focus on the vendor classifying to your Sourcing category structure, not just UNSPSC or other.
- Allow for a hands-on workshop to review the classification process and see full blown vendor capabilities for the long haul. It is through these workshops that enable you to really see what goes on “behind the scenes” and separate out fact from fiction.
Hi,
Few of the customers have thier Sourcing Category Structure mapped to the UNSPSC codes. Other few have UNSPSC mapped to thier Custom Classifcation codes (Custom Commodity Codes) which in turn is mapped to Sourcing Category Structure. The Custom Classification Code count might be less than the 8 digit UNSPSC (Say could be 6 digit) or can be more (Say 10 digit). So in such cases if the vendor classifies into UNSPSC, the reporting could be generated as per internal Sourcing Category Structure.
Also what is your thought on the input data classification? Whether it should be the Invoice Data or PO data or say going further down one level whether it has to be the Master data? Which data needs to be classified in what scenario? as customers tend to toggle between one data set to another for classifying.
Also our experience suggests that Data Cleansing and enrichment exercise and the resultant accuracy would give us a more data classification accuracy.
Thanks.
Senthil.
I appreciate your feedback and input. One of the points that I make is that UNSPSC is NOT needed or necessary for many companies. We are finding that most companies we are working with prefer to classify directly to their business Sourcing Categories, and eliminate UNSPSC all-together. This is because they discover that UNSPSC does not map well to their business Sourcing strategy and structure, and that is ultimately what they need. So as UNSPSC is no longer needed to get there, companies now have an alternative and effective business classification capability direct to Sourcing Categories. A simple example of a “Sourcing Category” at 4 levels… Level 1)= Direct, 2)= Metals, 3)= Steel, 4)= Stainless Pipes (or steel pipes, or Stainless general or steel tubes, etc. UNSPSC does not have this taxonomy, but it is how the company wanted its sourcing analysis and structure. How this maps to UNSPSC we don’t care – this is the primary company business sourcing taxonomy in play. With the sourcing taxonomy example above, we can report on all direct, all metals, all steel, and all types of steel components as different dimensions and breakout.
As for input data to classify, we classify to the lowest level possible on ALL input data, including AP, PO, GL, Pcard, Travel, Expenses, Financial data, etc (no limits). All this source data is integrated and classified to the chosen taxonomies, which is always 1) Company Sourcing Categories, and then as options UNSPSC, or any other taxonomy desired (eClass, Achilles, internal custom codes, CAS chemical codes, etc).
I agree that if a company also “enriches” it’s supplier data (SIC/NAICS, parent-child, Diversity, Risk) there are more valuable dimensions for both analysis and management tracking on the backend as these classifications are integrated with ongoing Spend data refreshes.