This post is the first in a series of postings that discuss the issue of bundling or lotting in B2B Reverse Auctions. Insights reported are the results of a large-scale study that I conducted for my dissertation at Indiana University. In the process of this study, in-depth interviews were held with over 30 companies utilizing reverse auctions. To confirm results and test hypotheses, an extensive survey was successfully administered, leading to many valuable conclusions, insights, and current best practices.
Bundling within an RFQ.
One of the more important components of the request for quote (RFQ) process is the determination of the appropriate order lot (bundle) that a vendor will evaluate and quote upon. The RFQ order lot may consist of a single stock keeping unit (SKU), but will more often be comprised of a set or bundle of different items that is attractive to one or more potential bidders. The composition of a bundle determines the set of products / services to be delivered by a vendor and the value of the transaction. The way a bundle is configured can also influence purchasing performance and the competitive interest between buyers and suppliers, especially in an online setting.
Bundling defined.
Within the context of B2B procurement, bundling is defined as the aggregation of two or more products (SKUs) and / or services by the buyer into a bundle that is put up for bid to potential suppliers as part of a single RFQ. In a very constrained setting, a single overall price needs to be quoted on the RFQ (a per-SKU price table may be required in addition), and all items in the RFQ need to be bid on. Deviations from the RFQ, for example in terms of content, conditions or specifications, are not allowed. However, more flexible forms of bundling are also possible; for example, instances in which suppliers are merely encouraged to bid on all items in a bundle.
The lack of research into Bundling for Sourcing.
While bundling has received some academic attention on the sell-side of the supply chain, hardly any research on the buy-side has investigated bundling approaches and associated issues. Moreover, while purchasing managers and buyers have performed bundling (a.k.a., aggregating, lotting and combining) for centuries, little systematic discussion or evaluation has taken place concerning how to make this important decision, and what variables may come into play. This is surprising, since bundling can be an important element in the general procurement process, particularly in online bidding events, generating a more competitive environment and purchase price reductions.
The importance of Bundling for Reverse Auctions.
While bundling is often done in offline purchase negotiations, its criticality is heightened in reverse auctions due to their usual short duration and constrained environment. In contrast, in an offline setting there is much more flexibility present, and RFQs can easily be modified, even once they have been sent out, to accommodate issues previously neglected or not considered. This decreased significance may be the reason why bundling has received so little attention in purchasing research in the past. For online bidding events, however, all potential concerns must be addressed beforehand due to the inability to change the bundle during these auctions. As such, bundle configuration is a crucial element for success in these; bundles that are not set up right may not lead to the desired result. Despite this significance, no published research was found that provides a systematic investigation in bundling for B2B online procurement auctions. It was therefore the goal of my dissertation to fill this gap. Results and insights will be reported in the weeks to come.
As with this and all my future posts, I would greatly appreciate any comments, concerns, suggestions, or questions that you may have.
Bundling is typically used to deliver economies of scope – such that the buyer deals with the fewest number of vendors commensurate with the product and service requirement. Equally, bundling offers economies of scale by delivering greater (more attractive) volumes to vendors, and ideally results in lower pricing for the buyer.
In my reverse auction events I seek to bundle to the point where I gain my economies, but am mindful of the possibilities for diseconomies, particularly on international events. I feel it important to bundle only so much as to optimise competition; With a large product/service bundle spread over multiple geographies I find this can limit greatly the number of capable bidders, and so always factor this in.
Thank you so very much for your very valuable comments Ian! You mention a very important point in bundling for reverse auctions, namely that there is a fine line between bundling too much and bundling too little. This was also confirmed in my research.
There is a tradeoff not only between bundling too much vs. too little volume, but also in regards to the homogeneity or heterogeneity of the bundle. The bundle should not be too diverse in terms of requirements, capabilities and capacities needed (i.e., suppliers should be able to competitively quote on the bundle), but a bundle should also not be too simple (i.e., should not only include very popular items for suppliers). It is again a fine line that has to be followed: combining attractive (popular) and unattractive (unpopular) items in the same bundle, to globally optimize the result for the entire set of items. The bundle cannot be too complex (include only unattractive items), otherwise no suppliers would be able to competitively bid on the business; it however also cannot be too simple (include only attractive items), since then the unattractive items would most probably not get any competitive bids.
I think Ian’s point has particular relevance in Europe where often bundles need to be in geographical lot groups. However how complex the bundles are will also depend on the supply base and their knowledge of live auctions. For example take stationary. It is an obvious auction commodity and has been frequently auctioned. Bidders in the past have been requested to bid on a basket of items which may only account for 80% of the spend. However a year into the contract further analysis will often show the core products will account for only 50% of the spend. The benefit is slowly eroded.
In order to combat this erosion I recently completed an event with 2,500 stationary items core and non core. We had excellent results and although there was initially a lot of work for the bidders they were well able to adjust bids. To help them we allowed plenty of time for pre bids and extra long extensions. There were no offline spreadsheets. The suppliers in this market are well trained at participating in events.
On last point we still split the European spend out separately.
If you want to find some of the earliest literature on overcoming “the lotting problem” for large market baskets (i.e., overcoming the tradeoff between no feasibility with large ‘winner take all’ lots or sub-optimized winners when too many small lots because prize too small and supplier doesn’t know how much business will be won, so, can’t put lowest total cost bid out there), check out Chris Caplice’s groundbreaking work in mid 1990′s (which built on first combinatorial auctions done by Sears in early 1990′s) on combinatorial optimization in Transportation which laid the groundwork for all the more generalized “expressive bidding” RFx tools from Emptoris, CombineNet, etc. This link is a good summary from Chris and Yossi Sheffi…
http://chainalytics.com/news/article_200302-TransportationBiddingArticle.pdf AMR first covered this in 1997 and I did a deep dive redux field report on it in 2003 when I was at AMR… http://www.amrresearch.com/Content/view.asp?pmillid=15925&docid=11086 This required an AMR subscription, but I also talked about these advanced analytic tools in a Purchasing magazine article…. http://www.purchasing.com/article/CA407148.html
Feel free to contact me on this. Hope this helps,
P
Dr. Schoenherr:
I found your bundling article interesting and enlightening, but I am not sure if you have defined what the primary goal of bundling is. http://www.esourcingforum.com/?p=23 and http://www.procureiq.com/2007/01/15/combinenet-v-expressive-bidding/
If the primary goal of bundling items together (by the buyer) is to get better pricing, then you want the supplier to bundle the items together. The reason is that they are the only ones that will know how to optimize their unique supply chain in order to decrease their costs. If they can decrease their costs, they can sell the items at a higher margin for themselves, but still at a lower cost for the buyer. That is, they are lowering the cost of sale for themselves and can pass on the savings to get business.
If the primary goal of bundling is to guarantee a buyer business rule, then this can be accomplished via constraints on the outcome of the awarded allocation. In this manner, let the suppliers bid on line-items or let the suppliers bundle items together, but when you find an optimal awarded allocation, make sure that solution guarantees your buyer business rule constraints are satisfied.
In either case, bundling by the buyer is not needed and in fact disadvantageous. Having the buyer put the bundles together assumes that 1) all suppliers have the same business processes, same infrastructure, and same delivery mechanisms and 2) that the buyer has perfect knowledge of the supplier business processes, infrastructure, and delivery mechanisms. These assumptions are almost always invalid.
I am sure there are special cases when the buyer wants to and should bundle items together and then this should be allowed, but in general letting the supplier bundle the items together to reduce the cost to deliver and guaranteeing that the allocation abides by the buyers business rules will al most always yields a lower cost of allocation.
Sincerely,
Jason Brown, PhD