Prices convey important information to guide resource allocation. But prices alone are not enough to finalize deals. For example, the touch and feel of non-standardized goods and services or the condition of standardized goods and services must also be known to evaluate whether the prices are reasonable.

Sometimes, it is difficult to evaluate what we see and what we are told because the seller knows more about the quality of the goods and services than the buyer (See Lemon Laundering). If such information asymmetry is not resolved, no transaction will take place because the buyer may offer too low a price to discount the uncertain quality. Or the transactions would result in unhappy buyers.

To get a better deal, buyers and sellers resort to different means to convey or conceal the truth. For example, sellers may use proxy indicators to signal quality (See May the Best Brand Win!). Buyers may pretend to be less interested or have more viable choices. Product warranty is another way to signal product quality because it is expensive to fake. But the proxy indicator loses its information value if every seller offers similar warranty terms which may not have to be honored immediately. Free trials are of course the ultimate tool to demonstrate product quality.

In general, reducing the cost of information will lead to greater economic efficiency because resources are allocated to higher-valued uses. For example, the low-cost agglomeration of buyers and sellers in eBay auctions ensures that both buyers and sellers get the best possible deals.* As the middleman, eBay also benefits from the transaction fees.

The Internet has done a good job of providing free information on expensive items where consumers' need for information is the greatest and where sellers could afford to support the sites with ad revenues. The economic principle behind the success of these information sites is quite simple. Namely, the marginal benefit of information for expensive items is high enough to justify the marginal cost of information search.

However, those information sites that are based on anonymous reviews are susceptible to adverse-selection bias as only people with extreme views are likely to submit reviews. Still, useful information could be inferred if readers take care to read between lines.

But more accurate information is not always helpful. When risk is involved, knowing the exact incidence of an outcome may destroy any voluntary insurance arrangement due to adverse selection. The risk is still there and needs to be spread, but those who are risk free this round will not be motivated to pay any insurance premiums.