5 Steps to Develop Procurement Data Analytics
How to master the procurement analytics
Why we failed in analytics.
Our predictive and prescriptive capabilities are barely there.
Perhaps, we are trying to make short-term commodity market predictions, i.e., we are not buying coal in winter and not placing critical orders during the New Year holidays. Our ammunition is limited - Platts, macro-economic forecasts, CPI, Industry Price Index, exchange rates, news headlines.
Our foresight is primarily based on past failures, which we try to not replicate by avoiding former faulty practices (even that is not entirely successful.)
5 practical steps we took to develop the procurement data analytics
Upon accepting the problem, we introduced a couple of improvements, some of which worked quite fine:
- Established commodity frame agreements with a pool of qualified suppliers who bid on regular e-auctions. The quantity and diversity of those suppliers ensure their appropriate behavior in line with the general market trends. Plus, we can constantly adjust the sourcing levers, e.g., e-auction frequency, size of lots, order timing, rectifying price deviations, or forming cartels.
- Formula pricing is our lifeline. We blended the global market intel data of a commodity index provider with our predictions on the margin fluctuations (differential,) which is the only element of the formula we manage (the rest are commodity index, exchange rates, and duties.) Already using this approach for fuels and metal and intending to extend it onto coal, electricity, natural gas, and copper cables.
- Consistent SRM activities with strategic suppliers enable access to the market intel of your partners. Their insight is golden, as long as your relationship is mutually fair and transparent. To gain the trust of our partners, we had to improve a lot on our payment discipline.
- Data cost money, so you would have to invest in new sources of it. Industry analysts, commodity exchanges, trade chambers sell their analytical data and market predictions, which would pay off nicely, as long as you have a working model to feed this data into.
- You need to raise a new breed of experts - data analysts. They are 33% computer geeks, 33% data crunching magicians, but still procurement people! The one we got created the BI suite from his smartphone, developed a dozen statistical models, tapped onto Platts and local market intel. Yet, he's a great expert on public procurement leading strategic sourcing projects and drafting contracts.
"The first step toward change is awareness. The second step is acceptance."
(c) Nathaniel Branden
Comments
Post a Comment