Spectrum auctions

To address increasing industry demand for electromagnetic spectrum and wireless broadband, governments utilize auction systems to sell licences to transmit signal over certain bands of spectrum.

With unparalleled expertise in spectrum auctions worldwide, the Smith Institute created a bid reconstruction tool for Vodafone in preparation for the 2015 Indian multi-band spectrum auction. This tool would enable the world-wide telecommunications provider to build a clear picture of bidder activity throughout the auction and strategically guide their own bidding behaviour.

The challenge

Vodafone wished to have a tool capable of:

  • generating a list of possible reconstructions for each circle;
  • ordering these reconstructions by plausibility; and
  • combining reconstructions across the different geographical regions

The tool needed to be constructed rapidly, with a practical run-time and scope for analysts to input proposed reconstructions.

The process

By combining information released during the auction with a key understanding of the objectives held by different bidders, inferences were drawn about bids from other buyers. These deductions informed valuable adjustment to Vodafone’s bidding strategy and thus a successful outcome.

The bid reconstruction tool was used by Vodafone’s analysts during the live auction and confirmed whether a proposed reconstruction was consistent with information provided by the auctioneer. It also gathered valuable knowledge about competition and market evaluations.

For 6 weeks our technical staff worked closely with Vodafone to create a system that was straightforward for analysts to use and reflected the complexity of the auction design. As a result of the collaboration, Vodafone could tackle the hidden challenges of the bidding process and secured spectrum in in all six of its existing 900MHz bands.

This was a large-scale problem, with unavoidable uncertainty in its definition. It would have been quite easy for the Smith Institute guys to object – how can we solve an optimisation problem if you can’t define it properly? – but they took it in their stride, seeing that the best approach would be to build a tool that would allow analysts to adapt the problem statement as they went along. I would certainly consider the Smith Institute for similar challenges in future.

Steve Babbage
Vodafone Distinguished Engineer, Security Research Manager and Group Chief Cryptographer at Vodafone Group R&D