As far as I can see, the problems surrounding National Express’s operation of the East Coast rail franchise stem from two basic issues:
1. The bidding process will necessarily result in the most optimistic bidder obtaining the franchise, so there is always a possibility that the franchisee will have overstretched itself in making the bid and may be unable to fulfil the contract. That in itself isn't a major problem; the potential for failure is part of what makes markets more responsive than state monopoly control, but it does mean that, if the government wants a continual service in all circumstances, it needs to consider the possibility of failure at the outset and allow for it.
2. Perhaps more importantly, the contracts are static whereas the economy is dynamic. If the contracts don't allow for the possibility of a recession occurring during the term of the contract, there is always the possibility that a bid which seemed reasonable at the outset will turn out to be non-viable in a downturn.
One possible way around point 2 and possibly point 1 as well, would be to apply the same methodology I’ve outlined in the past as a potential system for self-assessed land value tax. In simple terms, the potential franchisees would bid at the outset for the right to run the service, knowing that while they hold the franchise, they will be required to pay an annual fee equal to a percentage of their own estimate of the resale value of the franchise (say 15%), knowing that, if another operator agreed to meet that valuation, the franchisee would be obliged to sell to them.
That way, in a downturn, the operator would be able to reduce the valuation, rather than going through the upheaval of giving up the franchise only for it to be re-auctioned at a value likely to be far lower than the original.
Obviously, there would be a lot of detail to work out, but I think the idea has potential.