The answer - it depends on who you ask.
My definition is - a payment of $1.00 or less.
I would like to propose a new term (just what we need, right!) - Minipayment.
What is a minipayment? My definition is - a payment of between $1.01 and $5.00.
Then we would have a plain old "payment" which I would define as anything above $5.00.
There has been a lot of attempts to solve the micropayment problem. They go by the names; Digicash, CyberCash, MagnaCash, PepperCoin (all dead and gone) and the latest incarnations; SpareChange, PayByCash, etc. PayPal has toyed around in this area as well but without much success.
The problem is that if the consumer pays by Credit or Signature Debit (see earlier post), the cost of the transaction includes a flat fee of ~.20-.30 plus some % of ~2%. On a $5 transaction run through PayPal's (& now Google Checkout's) standard pricing, the fee works out to be 9%.
The "solution" has been to have the customer fund a larger amount, say $20 (works out to a 4.5% fee if a typical SMB merchant is using PayPal/Google Checkout). The problem with this are; a) the % of consumers willing to commit that larger amount on the hope that they will use it at one or more merchants that support it, b) the economics for the provider and the merchant of providing customer service around these very low value payments and c) for the credit card companies, the % of these transactions that result in disputes at a very high cost. Other solutions are individual game currency, pre-paid cards, mobile payments but each of these have a variety of issues whether it is; utility, accessibility, cost, fraud, etc.
I do not have a magic bullet for this one, I know too much about how all this works and therefore the challenges. I am always intrigued to observe the volume of noise on this issue and the amount of VC money that finds it way to the latest attempt.
p.s. The biggest challenge is not any particular technical issue. If the credit card companies or PayPal wanted to offer a viable solution, they could. It is the Incumbent's Dilemma in spades, however, since in order to price a micropayment service in the realm of what the merchant would desire, it would undermine the pricing models for the higher value payments since the fact is, it costs about the same to process a $1.00 transaction as a $100.00 transaction and at the volumes the big guys are processing, that incremental cost for the next transaction is pretty low. The other guys that are trying to solve it are faced with all the chicken & egg start-up costs whereas the incumbents are already at scale.