$300 Million and a Nobel Prize

Michael Watson and Arne Andersson
Michael Watson and Arne Andersson

Mike Watson heads the AI COE at Coupa, he has a PhD in Industrial Engineering and 20+ years of experience leading global supply chain business teams. Mike co-founded Opex Analytics and grew the company to success in the AI market prior to integrating with LLamasoft and now Coupa.  He is also an adjunct professor at two masters programs at Northwestern University and the co-author of two books (Managerial Analytics and Supply Chain Network Design).

Arne Andersson (Ph.D. in Computer Science) is VP of engineering at Coupa Software. He was one of the founders of Trade Extensions, now Coupa Sourcing Optimization (CSO). Apart from 20+ years in advanced eSourcing, he has an academic background within algorithms and optimization. He has been Chair Professor and Head of Department of Information Technology at Uppsala University. Arne is a member of The Royal Swedish Academy of Engineering Sciences, and he serves in the program committee for Information, Communication & Systems Technologies in the Swedish Foundation for Strategic Research.

Read time: 3 mins
$300 Million and a Nobel Prize

Donna Wilczek recently posted about a Coupa customer who saved $300M (versus their current spend) on a single sourcing event. This was a large event with 6,000 items and 300+ suppliers bidding. The buyer also had 400 different business rules as part of the process. The product behind this savings is Coupa Sourcing Optimization (CSO).

Interestingly, it is the science embedded within CSO that makes this kind of savings possible. It is the same science behind the winners of the 2020 Nobel Prize in Economics for their work in auctions. Specifically, the science is combinatorial optimization (Ed Rothberg, one of the world’s leading experts in optimization, wrote a nice explanation).

Combinatorial optimization is what allows the savings in these auctions. In short, CSO allows the suppliers to enter bids for combinations of items. This allows suppliers to bid on packages of items that allow for capturing economies of scale. Because of the high number of possible combinations of bids and the fact that there can only be one winner for a single item, it is actually impossible to determine the winners without combinatorial optimization.

In an example that we like to use to teach the material, a government agency in Chile switched bidding for school meal programs from a typical auction to a combinatorial auction. (Here is a link to the abstract of the article.) This new auction allowed suppliers to bid on multiple districts where they had economies of scale. This saved the government $40M a year out of a spend of $180M. They credit the combinatorial optimization with 40% of the savings. The rest of the savings came from changing the process.

That process is an important part of CSO too. In 2000, one of the co-authors, Arne Andersson, co-founded Trade Extensions, which would later become CSO. The product was created to combine combinatorial optimization with the better process possibilities offered by the evolving internet.

CSO ran the world's first web-based iterative combinatorial auction in early 2001. It was for Volvo. It was a combinatorial auction for wooden packaging material, such as boxes and pallets.

This first auction was carefully designed. The bidders placed their bids and received their feedback in a web interface. Bids could be placed on individual items and on arbitrary combinations of items. Bidding was done in rounds and used bidding rules (such as minimal increments) and activity rules (if you are not leading, you have to place a bid) to make sure that the auction moved forward at each round. In order to give the participants enough time for considering their bids, each round lasted for one hour. Savings was much better than expected, and therefore the auction lasted for more than a week. In summary, the auction was a great success.

Today, CSO has expanded beyond plain combinatorial auctions. Instead of just letting bidders place bids on combinations, suppliers can describe capacity constraints, various kinds of volume discounts, and other side constraints.

Furthermore, there are many buyer-side constraints to address as well. Buyer-side constraints include cost, quality parameters, CO2 emissions, supply chain modeling, supplier switching costs, different forecasting assumptions, various stakeholder needs, etc. This allows for the buyer to get much more value than just cost savings — the buyer can reduce risk and meet many other specified business requirements.

CSO is a great story of how Coupa is committed to innovation around both the process and the science to bring value to our customers.

Optimize your strategic sourcing and enable your organization to utilize multi-factor evaluation criteria (not just price) for supplier offers with Coupa Sourcing Optimization (CSO). 

For a more detailed discussion on the science behind CSO, here is an article in an academic magazine.